Professional Documents
Culture Documents
Electric Machines NPTEL PDF
Electric Machines NPTEL PDF
Sasidhara Rao
Transformers
1 Introduction
Transformers can link two or more electric circuits. In its simple form two electric
circuits can be linked by a magnetic circuit, one of the electric coils is used for the creation
of a time varying magnetic filed. The second coil which is made to link this field has an
induced voltage in the same. The magnitude of the induced emf is decided by the number
of turns used in each coil. Thus the voltage level can be increased or decreased by changing
the number of turns. This excitation winding is called a primary and the output winding
is called a secondary. As a magnetic medium forms the link between the primary and the
secondary windings there is no conductive connection between the two electric circuits. The
transformer thus provides an electric isolation between the two circuits. The frequency on
the two sides will be the same. As there is no change in the nature of the power, the re-
sulting machine is called a ‘transformer’ and not a ‘converter’. The electric power at one
voltage/current level is only ‘transformed’ into electric power, at the same frequency, to an-
other voltage/current level.
Even though most of the large-power transformers can be found in the power systems,
the use of the transformers is not limited to the power systems. The use of the principle
of transformers is universal. Transformers can be found operating in the frequency range
starting from a few hertz going up to several mega hertz. Power ratings vary from a few
milliwatts to several hundreds of megawatts. The use of the transformers is so wide spread
that it is virtually impossible to think of a large power system without transformers. Demand
on electric power generation doubles every decade in a developing country. For every MVA
of generation the installed capacity of transformers grows by about 7MVA. These figures
show the indispensable nature of power transformers.
2 Basic Principles
coil of N turns. All these N turns link flux lines of φ Weber resulting in the Nφ flux linkages.
In such a case,
ψ = Nφ (2)
and
dφ
e=N volt (3)
dt
The change in the flux linkage can be brought about in a variety of ways
• coil may be static and unmoving but the flux linking the same may change with time.
• flux lines may be constant and not changing in time but the coil may move in space
linking different value of flux with time.
• both 1 and 2 above may take place. The flux lines may change in time with coil moving
in space.
These three cases are now elaborated in sequence below, with the help of a coil with a simple
geometry.
L
B
- +
B = Bm sin ωt (4)
where Bm is the peak amplitude of the flux density. ω is the angular rate of change with
time. Then, the instantaneous value of the flux linkage is given by,
em = Nφm .ω (7)
Fig. 2(b) shows the same example as above but with a small difference. The flux
density is held constant at B Tesla. The flux linked by the coil at the current position is
φ = B.L.X Weber. The conductor is moved with a velocity v = dx/dt normal to the flux,
cutting the flux lines and changing the flux linkages. The induced emf as per the application
of Faraday’s law of induction is e = N.B.L.dx/dt = B.L.v volt.(Here N=1)
Please note,the actual flux linked by the coil is immaterial. Only the change in the
flux linkages is needed to be known for the calculation of the voltage. The induced emf is
in step with the change in ψ and there is no phase shift. If the flux density B is distributed
sinusoidally over the region in the horizontal direction, the emf induced also becomes sinu-
soidal. This type of induced emf is termed as speed emf or rotational emf, as it arises out of
the motion of the conductor. The polarity of the induced emf is obtained by the application
of the Lenz’s law as before. Here the changes in flux linkages is produced by motion of the
conductor. The current in the conductor, when the coil ends are closed, makes the conductor
experience a force urging the same to the left. This is how the polarity of the emf shown in
fig.2b is arrived at. Also the mmf of the loop aids the field mmf to oppose change in flux
linkages. This principle is used in d.c machines and alternators.
The third case under the application of the Faraday’s law arises when the flux changes
and also the conductor moves. This is shown in Fig. 2(c).
The first term is due to the changing flux and hence is a transformer emf. The second
term is due to moving conductor or is a speed emf. When the terminals are closed such as
to permit a current the conductor experiences a force and also the mmf of the coil opposes
the change in flux linkages. This principle is used in a.c. machines where the field is time
varying and conductors are moving under the same.
The first case where there is a time varying field and a stationary coil resulting in
a transformer emf is the subject matter in the present section. The case two will be re-
visited under the study of the d.c machines and synchronous machines. Case three will be
extensively used under the study of a.c machines such as induction machines and also in a.c.
commutator machines.
Next in the study of the transformers comes the question of creating a time varying
filed. This is easily achieved by passing a time varying current through a coil. The winding
which establishes the field is called the primary. The other winding, which is kept in that
field and has a voltage induced in it, is called a secondary. It should not be forgotten that
the primary also sees the same time varying field set up by it linking its turns and has an
induced emf in the same. These aspects will be examined in the later sections. At first
the common constructional features of a transformer used in electric power supply system
operating at 50 Hz are examined.
3 Constructional features
Transformers used in practice are of extremely large variety depending upon the
end use. In addition to the transformers used in power systems, in power transmission and
distribution, a large number of special transformers are in use in applications like electronic
supplies, rectification, furnaces, traction etc. Here the focus is on power transformers only.
The principle of operation of these transformers also is the same but the user requirements
differ. Power transformers of smaller sizes could be air cooled while the larger ones are
oil cooled. These machines are highly material intensive equipments and are designed to
match the applications for best operating conditions. Hence they are ‘tailor made’ to a
job. This brings in a very large variety in their constructional features. Here more common
constructional aspects alone are discussed. These can be broadly divided into
1. Core construction
2. Winding arrangements
3. Cooling aspects
Transformer core for the power frequency application is made of highly permeable
material. The high value of permeability helps to give a low reluctance for the path of
the flux and the flux lines mostly confine themselves to the iron. Relative permeability µr
well over 1000 are achieved by the present day materials. Silicon steel in the form of thin
laminations is used for the core material. Over the years progressively better magnetic prop-
erties are obtained by going in for Hot rolled non-oriented to Hot rolled grain oriented steel.
Later better laminations in the form of cold Rolled Grain Oriented (CRGO), -High B (HiB)
grades became available. The thickness of the laminations progressively got reduced from
over 0.5mm to the present 0.25mm per lamination. These laminations are coated with a thin
layer of insulating varnish, oxide or phosphate. The magnetic material is required to have
a high permeability µ and a high saturation flux density, a very low remanence Br and a
small area under the B-H loop-to permit high flux density of operation with low magnetizing
current and low hysteresis loss. The resistivity of the iron sheet itself is required to be high
to reduce the eddy current losses. The eddy current itself is highly reduced by making the
laminations very thin. If the lamination is made too thin then the production cost of steel
laminations increases. The steel should not have residual mechanical stresses which reduce
their magnetic properties and hence must be annealed after cutting and stacking.
In the case of very small transformers (from a few volt-amperes to a few kilo volt-
amperes) hot rolled silicon steel laminations in the form of E & I, C & I or O as shown in
Fig. 3 are used and the core cross section would be a square or a rectangle. The percentage
of silicon in the steel is about 3.5. Above this value the steel becomes very brittle and also
very hard to cut. The saturation flux density of the present day steel lamination is about 2
Tesla.
Broadly classifying, the core construction can be separated into core type and
shell type. In a core type construction the winding surrounds the core. A few examples of
single phase and three phase core type constructions are shown in Fig. 4. In a shell type on
the other hand the iron surrounds the winding.
In the case of very small transformers the conductors are very thin and round.
These can be easily wound on a former with rectangular or square cross section. Thus no
special care is needed for the construction of the core. The cross section of the core also
would be square or rectangular. As the rating of the transformer increases the conductor size
(a ) (b)
(c)
10
1.phase
3.phase
HV LV core LV HV LV HV
also increases. Flat conductors are preferred to round ones. To wind such conductor on a
rectangular former is not only difficult but introduces stresses in the conductor, at the bends.
From the short circuit force with stand capability point of view also this is not desirable.
Also, for a given area enclosed the length of the conductor becomes more. Hence it results in
more load losses. In order to avoid all these problems the coils are made cylindrical and are
wound on formers on heavy duty lathes. Thus the core construction is required to be such as
to fill the circular space inside the coil with steel laminations. Stepped core construction thus
becomes mandatory for the core of large transformers. Fig. 5 shows a few typical stepped core
constructions. When the core size increases it becomes extremely difficult to cool the same
(Even though the core losses are relatively very small). Cooling ducts have to be provided
in the core. The steel laminations are grain oriented exploiting the simple geometry of the
transformer to reduce the excitation losses. The iron losses in the lamination, when the flux
11
0.14 0.14
0.16 0.1 0.1 0.07 0.12
0.16 0.12 0.07
0.09 0.09
0.53 0.42 0.3
0.71D
d d d
duct
duct
12
is oriented in the direction of grain orientation, is about 30% of that in the normal direction.
Another important aspect to be carefully checked and monitored is the air gaps in
Path of
flux
HV
LV
Windings
Core
(a) (b)
series in the path of the main flux. As the reluctance of air path is about 1000 times more
than that of the steel, an air path of 1mm will require a mmf needed by a 1 meter path in iron.
Hence butt joints between laminations must be avoided. Lap joints are used to pro-
vide alternate paths for flux lines thus reducing the reluctance of the flux paths. Some typical
constructional details are shown in Fig. 6. In some power transformers the core is built up
by threading a long strip of steel through the coil in the form of a toroid. This construction
is normally followed in instrument transformers to reduce the magnetizing current and hence
the errors.
13
Large cores made up of laminations must be rendered adequately stiff by the provi-
sion of stiffening plates usually called as flitch plates. Punched through holes and bolts are
progressively being avoided to reduce heating and melting of the through bolts. The whole
stack is wrapped up by strong epoxy tapes to give mechanical strength to the core which
can stand in upright position. Channels and angles are used for the frame and they hold the
bottom yoke rigidly.
3.2 Windings
14
HV LV
Core
LV
HV
(a)Concentric coil
LV
HV
Core
15
are common with shell type transformers. In the figure the letters L and H indicate the low
voltage and high voltage windings. In concentric arrangement, in view of the lower insulation
and clearance requirements, the LV winding is placed close to the core which is at ground
potential. The HV winding is placed around the LV winding. Also taps are provided on HV
winding when voltage change is required. This is also facilitated by having the HV winding
as the outer winding.
Three most common types of coils viz. helical, cross over and disc coils are shown in Fig. 8.
Helical coils
Disc coils cross over coils
Helical Windings One very common cylindrical coil arrangement is the helical winding.
This is made up of large cross section rectangular conductor wound on its flat side.
The coil progresses as a helix. This is commonly used for LV windings. The insulation
16
requirement also is not too high. Between layers no insulation (other than conductor
insulation) is needed as the voltage between layers is low. The complexity of this
type of winding rapidly increases as the current to be handled becomes more. The
conductor cross section becomes too large and difficult to handle. The eddy current
losses in the conductor rapidly increases. Hence two or more conductors have to be
wound and connected in parallel. The parallel circuits bring in problems of current
sharing between the circuits. Transpositions of the parallel paths have to be adopted
to reduce unequal current distribution. The modern practice is to use continuously
transposed and bunched conductors.
Cross over coils The second popular winding type is the cross over coil. These are made
of circular conductors not exceeding 5 to 6 sq mm in cross section. These are used for
HV windings of relatively small transformers. These turns are wound in several layers.
The length and thickness of each block is made in line with cooling requirements. A
number of such blocks can be connected in series, leaving cooling ducts in between the
blocks, as required by total voltage requirement.
Disc coils Disc coils consist of flat conductors wound in a spiral form at the same place
spiralling outwards. Alternate discs are made to spiral from outside towards the center.
Sectional discs or continuous discs may be used. These have excellent thermal prop-
erties and the behavior of the winding is highly predictable. Winding of a continuous
disc winding needs specialized skills.
Sandwich coils Sandwich windings are more common with shell type core construction.
They permit easy control over the short circuit impedance of the transformer. By
bringing HV and LV coils close on the same magnetic axis the leakage is reduced
and the mutual flux is increased. By increasing the number of sandwiched coils the
17
3.3 Insulation
Scaling advantages make the design of larger and larger unit sizes of transformers
economically attractive. This can be explained as below. Consider a transformer of certain
rating designed with certain flux density and current density. If now the linear dimensions
are made larger by a factor of K keeping the current and flux densities the same the core and
conductor areas increase by a factor of K 2 . The losses in the machine, which are proportional
to the volume of the materials used, increase by a factor of K 3 .The rating of the machine
increases by a factor of K 4 .
18
The surface area however increases by a factor of K 2 only. Thus the ratio of loss per
surface area goes on increasing by a factor of K. The substantial increase in the output is
the major attraction in going in for larger units. However cooling of the transformer becomes
more and more difficult. As the rating increases better cooling techniques are needed.
Simple air cooling of the transformers is adopted in dry type transformers. The limit
for this is reached by the time the rating is a few kVA. Hence air cooling is used in low
voltage machines. This method of cooling is termed as AN(Air Natural). Air Blast(AB)
method improves on the above by directing the blast of air at the core and windings. This
permits some improvement in the unit sizes.
OB(Oil Blast) method is an improvement over the ON-type and it directs a blast of
air on the cooling surface. In the above two cases the flow of oil is by natural convective
forces. The rate of circulation of oil can be increased with the help of a pump, with the
cooling at the surface remaining natural cooling to air. This is termed as OFN (Oil Forced
Natural). If now a forced blast of air is also employed, the cooling method become OFB(
Oil Forced Blast). A forced circulation of oil through a radiator is done with a blast of air
19
Main tank
Radiator
Tubes
(a)
Conservator
Bushing & Breather
water outlet
Radiator
oil pump
water inlet
(b)
Conservator&
Breather
Bushing
Radiator
(c)
20
Figure 9: Some Typical Cooling Arrangements
over the radiator surface. Substantial amount of heat can be removed by employing a water
cooling. Here the hot oil going into the radiator is cooled by a water circuit. Due to the
high specific heat of water, heat can be evacuated effectively. Next in hierarchy comes OFW
which is similar to OFB except that instead of blast of air a forced circulation of cool water
in the radiator is used in this. Some cooling arrangements are shown in Fig. 9.
In many large sized transformers the cooling method is matched with the amount
of heat that is required to be removed. As the load on the transformer changes the heat
generated within also changes. Suitable cooling method can be pressed into service at that
time. This gives rise to the concept of mixed cooling technique.
ON/OB Works as ON but with increased load additional air blast is adopted. This gives
the ratings to be in the ratio of 1:1.5
The temperature rise permitted in the British standard specification for power transformers
are tabulated below.
Type winding oil core
Class A Class B
◦
C ◦
C ◦
C
AN,AB 55 75 - As
ON,OB,OW 60 - 50 for
OFN,OFB 65 - 50 adjacent
OFW 70 - 50 winding
21
Even though the basic functions of the oil used in transformers are a) heat conduc-
tion and b) electrical insulation, there are many other properties which make a particular oil
eminently suitable. Organic oils of vegetative or animal origin are good insulators but tend
to decompose giving rise to acidic by-products which attack the paper or cloth insulation
around the conductors.
Mineral oils are suitable from the point of electrical properties but tend to form sludge.
The properties that are required to be looked into before selecting an oil for transformer
application are as follows:
Insulting property This is a very important property. However most of the oils naturally
fulfil this. Therefore deterioration in insulating property due to moisture or contami-
nation may be more relevant.
Viscosity It is important as it determines the rate of flow of the fluid. Highly viscous fluids
need much bigger clearances for adequate heat removal.
Purity The oil must not contain impurities which are corrosive. Sulphur or its compounds
as impurities cause formation of sludge and also attack metal parts.
Sludge formation Thickening of oil into a semisolid form is called a sludge. Sludge for-
mation properties have to be considered while choosing the oil as the oil slowly forms
semi-solid hydrocarbons. These impede flows and due to the acidic nature, corrode
metal parts. Heat in the presence of oxygen is seen to accelerate sludge formation. If
the hot oil is prevented from coming into contact with atmospheric air sludge formation
22
Acidity Oxidized oil normally produces CO2 and acids. The cellulose which is in the paper
insulation contains good amount of moisture. These form corrosive vapors. A good
breather can reduce the problems due to the formation of acids.
Flash point And Fire point Flash point of an oil is the temperature at which the oil
ignites spontaneously. This must be as high as possible (not less than 160◦ C from the
point of safety). Fire point is the temperature at which the oil flashes and continuously
burns. This must be very high for the chosen oil (not less than 200◦ C).
Inhibited oils and synthetic oils are therefore used in the transformers. Inhibited oils
contain additives which slow down the deterioration of properties under heat and moisture
and hence the degradation of oil. Synthetic transformer oil like chlorinated diphenyl has
excellent properties like chemical stability, non-oxidizing, good dielectric strength, moisture
repellant, reduced risk due fire and explosion.
It is therefore necessary to check the quality of the oil periodically and take corrective
steps to avoid major break downs in the transformer.
There are several other structural and insulating parts in a large transformer. These
are considered to be outside the scope here.
23
4 Ideal Transformer
Earlier it is seen that a voltage is induced in a coil when the flux linkage associated
with the same changed. If one can generate a time varying magnetic field any coil placed in
the field of influence linking the same experiences an induced emf. A time varying field can
be created by passing an alternating current through an electric coil. This is called mutual
induction. The medium can even be air. Such an arrangement is called air cored transformer.
Indeed such arrangements are used in very high frequency transformers. Even though the
principle of transformer action is not changed, the medium has considerable influence on the
working of such devices. These effects can be summarized as the followings.
1. The magnetizing current required to establish the field is very large, as the reluctance
of the medium is very high.
2. There is linear relationship between the mmf created and the flux produced.
5. It is very hard to direct the flux lines as we desire, as the whole medium is homogeneous.
If the secondary is not loaded the energy stored in the magnetic field finds its way
back to the source as the flux collapses. If the secondary winding is connected to a load then
part of the power from the source is delivered to the load through the magnetic field as a link.
The medium does not absorb and lose any energy. Power is required to create the field and
not to maintain the same. As the winding losses can be made very small by proper choice
of material, the ideal efficiency of a transformer approaches 100%. The large magnetizing
24
Primary
Leakage
x flux
Secondary
Mutual flux
(a)
Leakage flux
X
Primary
Mutual flux
Secondary
Iron core
(b)
25
1. Due to the large value for the permeance ( µr of the order of 1000 as compared to
air) the magnetizing current requirement decreases dramatically. This can also be
visualized as a dramatic increase in the flux produced for a given value of magnetizing
current.
2. The magnetic medium is linear for low values of induction and exhibits saturation type
of non-linearity at higher flux densities.
3. The iron also has hysteresis type of non-linearity due to which certain amount of power
is lost in the iron (in the form of hysteresis loss), as the B-H characteristic is traversed.
4. Most of the flux lines are confined to iron path and hence the mutual flux is increased
very much and leakage flux is greatly reduced.
5. The flux can be easily ‘directed’ as it takes the path through steel which gives great
freedom for the designer in physical arrangement of the excitation and output windings.
6. As the medium is made of a conducting material eddy currents are induced in the
same and produce losses. These are called ‘eddy current losses’. To minimize the
eddy current losses the steel core is required to be in the form of a stack of insulated
laminations.
From the above it is seen that the introduction of magnetic core to carry the flux
introduced two more losses. Fortunately the losses due to hysteresis and eddy current for
the available grades of steel is very small at power frequencies. Also the copper losses in the
26
1. Magnetic circuit is linear and has infinite permeability. The consequence is that a van-
ishingly small current is enough to establish the given flux. Hysteresis loss is negligible.
As all the flux generated confines itself to the iron, there is no leakage flux.
2. Windings do not have resistance. This means that there are no copper losses, nor there
is any ohmic drop in the electric circuit.
In fact the practical transformers are very close to this model and hence no major
departure is made in making these assumptions.
Fig. 11 shows a two winding ideal transformer. The primary winding has T1 turns and is
connected to a voltage source of V1 volts. The secondary has T2 turns. Secondary can be
connected to a load impedance for loading the transformer. The primary and secondary are
shown on the same limb and separately for clarity.
27
v1=V1mcosωt
µ
8
φ
io 0
+ +
~ e1 T1 i1
- + i2
+ +
e2 T2 v1=V1sinωt e1 e2
-
+ -
v1=V1cosωt µ
8
φ
i1
+ +
N e1 T1
-
-
i2
+
e2 T2
ZL
-
(c)Loaded machine
28
mmf I0 T1 I0 T1 Aµ
F lux φ = = l = . (9)
Reluctance µA
l
This current is the result of a sinusoidal voltage V applied to the primary. As the
current through the loop is zero (or vanishingly small), at every instant of time, the sum of
the voltages must be zero inside the same. Writing this in terms of instantaneous values we
have,
v1 − e1 = 0 (10)
where v1 is the instantaneous value of the applied voltage and e1 is the induced emf due to
Faradays principle. The negative sign is due to the application of the Lenz’s law and shows
that it is in the form of a voltage drop. Kirchoff’s law application to the loop will result in
the same thing.
29
Here ψ1peak is the peak value of the flux linkages of the primary. The same mutual
flux links the secondary winding. However the magnitude of the flux linkages will be ψ2peak =
T2 .φm . The induced emf in the secondary can be similarly obtained as ,
I1
I2
+
+
V1 E1 E2 V2
- -
The voltages E1 and E2 are obtained by the same mutual flux and hence they are
in phase. If the winding sense is opposite i.e., if the primary is wound in clockwise sense
and the secondary counter clockwise sense then if the top terminal of the first winding is
at maximum potential the bottom terminal of the second winding would be at the peak
potential. Similar problem arises even when the sense of winding is kept the same, but the
30
two windings are on opposite limbs (due to the change in the direction of flux). Hence in
the circuit representation of transformers a dot convention is adopted to indicate the ter-
minals of the windings that go high (or low) together. (Fig. 12). This can be established
experimentally by means of a polarity test on the transformers. At a particular instant of
time if the current enters the terminal marked with a dot it magnetizes the core. Similarly
a current leaving the terminal with a dot demagnetizes the core.
i1 T1 − i2 T2 = i0 T1 but i0 → 0 (16)
If the reference directions for the two currents are chosen as in the Fig. 12, then the above
equation can be written in phasor form as,
T2
I¯1 T1 = I¯2 T2
or I¯1 = .I¯2 (18)
T1
E1 T1 I2
Also = = E1 I1 = E2 I2 (19)
E2 T2 I1
31
Thus voltage and current transformation ratio are inverse of one another. If an impedance
of ZL is connected across the secondary,
Ē2 Ē2
I¯2 = ¯ or Z¯L = ¯ (20)
ZL I2
Ē1 T1 Ē2 T1
Z̄i = ¯ = ( )2 . ¯ = ( )2 .Z¯L (21)
I1 T2 I2 T2
V1 E1
I2
E2 V2
I1
θ1 θ2
φ φ
Finally, the phasor diagram for the operation of the ideal transformer is shown in
Fig. 13 in which θ1 and θ2 are power factor angles on the primary and secondary sides. As
32
the transformer itself does not absorb any active or reactive power it is easy to see that
θ1 = θ2 .
Thus, from the study of the ideal transformer it is seen that the transformer provides
electrical isolation between two coupled electric circuits while maintaining power invariance
at its two ends. However, grounding of loads and one terminal of the transformer on the
secondary/primary side are followed with the provision of leakage current detection devices
to safe guard the persons working with the devices. Even though the isolation aspect is
a desirable one its utility cannot be over emphasized. It can be used to step up or step
down the voltage/current at constant volt-ampere. Also, the transformer can be used for
impedance matching. In the case of an ideal transformer the efficiency is 100% as there are
no losses inside the device.
33
5 Practical Transformer
Finite permeability of the magnetic circuit necessitates a finite value of the current
to be drawn from the mains to produce the mmf required to establish the necessary flux.
The current and mmf required is proportional to the flux density B that is required to be
established in the core.
φ
B = µH; B= (22)
A
where A is the area of cross section of the iron core m2 . H is the magnetizing force which is
given by,
T1
H = i. (23)
l
where l is the length of the magnetic path, m. or
Aµ(iT1 )
φ = B.A = = permeance ∗ mmf (here that of primary) (24)
l
The magnetizing force and the current vary linearly with the applied voltage as long
as the magnetic circuit is not saturated. Once saturation sets in, the current has to vary in
34
a nonlinear manner to establish the flux of sinusoidal shape. This non-linear current can be
resolved into fundamental and harmonic currents. This is discussed to some extent under
harmonics. At present the effect of this non-linear behavior is neglected as a secondary
effect. Hence the current drawn from the mains is assumed to be purely sinusoidal and
directly proportional to the flux density of operation. This current can be represented by a
current drawn by an inductive reactance in the circuit as the net energy associated with the
same over a cycle is zero. The energy absorbed when the current increases is returned to
the electric circuit when the current collapses to zero. This current is called the magnetizing
current of the transformer. The magnetizing current Im is given by Im = E1 /Xm where Xm
is called the magnetizing reactance. The magnetic circuit being lossy absorbs and dissipates
the power depending upon the flux density of operation. These losses arise out of hysteresis,
eddy current inside the magnetic core. These are given by the following expressions:
Ph ∝ B 1.6 f (25)
Pe ∝ B 2 f 2 t2 (26)
For a constant voltage, constant frequency operation B is constant and so are these
losses. An active power consumption by the no-load current can be represented in the input
circuit as a resistance Rc connected in parallel to the magnetizing reactance Xm . Thus the
no-load current I0 may be made up of Ic (loss component) and Im (magnetizing component
35
as )
I¯0 = I¯c − j I¯m (27)
Ic2 Rc – gives the total core losses (i.e. hysteresis + eddy current loss)
2
Im Xm - Reactive volt amperes consumed for establishing the mutual flux.
Finite µ of the magnetic core makes a few lines of flux take to a path through the air.
Thus these flux lines do not link the secondary winding. It is called as leakage flux. As the
path of the leakage flux is mainly through the air the flux produced varies linearly with the
primary current I1 . Even a large value of the current produces a small value of flux. This
flux produces a voltage drop opposing its cause, which is the current I1 . Thus this effect of
the finite permeability of the magnetic core can be represented as a series inductive element
jxl1 . This is termed as the reactance due to the primary leakage flux. As this leakage flux
varies linearly with I1 , the flux linkages per ampere and the primary leakage inductance
are constant (This is normally represented by ll1 Henry). The primary leakage reactance
therefore becomes
xl1 = 2πf ll1 ohm (28)
A similar effect takes place on the secondary side when the transformer is loaded.
The secondary leakage reactance jxl2 arising out of the secondary leakage inductance ll2 is
given by
Finally, the primary and secondary windings are wound with copper (sometimes alu-
minium in small transformers) conductors; thus the windings have a finite resistance (though
36
r1 jxl1 I’2 φ
I1
+
Io +
V1
~ E1 T1
Rc jXm
-
r2 jxl2 I2
V2 ZL E2 T2
-
(a)Physical arrangement
r1 jXl1
I1 I’2 r2 jXl2
I2
Io
Ic Im
E1 E2 ZL V2
V1 Rc
jXm
(b)Equivalent circuit
37
small). This is represented as a series circuit element, as the power lost and the drop pro-
duced in the primary and secondary are proportional to the respective currents. These are
represented by r1 and r2 respectively on primary and secondary side. A practical transformer
sans these imperfections (taken out and represented explicitly in the electric circuits) is an
′
ideal transformer of turns ratio T1 : T2 (voltage ratio E1 : E2 ). This is seen in Fig. 14. I2
in the circuit represents the primary current component that is required to flow from the
mains in the primary T1 turns to neutralize the demagnetizing secondary current I2 due to
the load in the secondary turns. The total primary current
By solving this circuit for any load impedance ZL one can find out the performance of the
loaded transformer.
The circuit shown in Fig. 14(b). However, it is not very convenient for use due to
the presence of the ideal transformer of turns ratio T1 : T2 . If the turns ratio could be made
unity by some transformation the circuit becomes very simple to use. This is done here by
replacing the secondary by a ‘hypothetical’ secondary having T1 turns which is ‘equivalent ’
to the physical secondary. The equivalence implies that the ampere turns, active and reactive
power associated with both the circuits must be the same. Then there is no change as far
as their effect on the primary is considered. Thus
′ ′ I2 ′ ′ ′
V2 = aV2 , I2 = , r2 = a2 r2 , xl2 = a2 xl2 ZL = a2 ZL .
a
T1
where a -turns ratio T2
38
This equivalent circuit is as shown in Fig. ??(a). As the ideal transformer in this
case has a turns ratio of unity the potentials on either side are the same and hence they
may be conductively connected dispensing away with the ideal transformer. This particular
equivalent circuit is as seen from the primary side. It is also possible to refer all the pri-
mary parameters to secondary by making the hypothetical equivalent primary winding on
the input side having the number of turns to be T2 . Such an equivalent circuit having all
the parameters referred to the secondary side is shown in fig. 15.
The equivalent circuit can be derived, with equal ease, analytically using the Kir-
choff’s equations applied to the primary and secondary. Referring to fig. 14(a), we have (by
neglecting the shunt branch)
Multiply both sides of Eqn.34 by ‘a’ [This makes the turns ratio unity and retains the power
invariance].
aE2 = aV2 + aI2 (r2 + jxl2 ) but aE2 = E1 (36)
39
A similar procedure can be used to refer all parameters to secondary side. (Shown in fig. 15.)
40
6 Phasor diagrams
Io
Ic Im
Rc V’2 Z’L
V1 jXm
(a)
r1 jxl1 r’2 jx’l2 R jX
I1 I’2 I1 I’2
R=r1+r’2
Io
Ic Im
Z’L V’2 x=xl1+x’l2
V1 Rc V1 V’2
jxm
I1=I’2
(b) (c)
41
On similar lines to the ideal transformer the phasor diagram of operation can be
drawn for a practical transformer also. The positions of the current and induced emf phasor
are not known uniquely if we start from the phasor V1 . Hence it is assumed that the phasor
φ is known. The E1 and E2 phasor are then uniquely known. Now, the magnetizing and loss
components of the currents can be easily represented. Once I0 is known, the drop that takes
place in the primary resistance and series reactance can be obtained which when added to
E1 gives uniquely the position of V1 which satisfies all other parameters. This is represented
in Fig. 17(a) as phasor diagram on no-load.
42
V1
IoX l1
E1 Ior1 E2
Io
φ φ
Im Il
(a)No-load
V1
I1X l1
I1r1
E1 I2x2 I r
2 2
E2 I2
V2
I’2 Il
Io
φ φ
(b)On-load
43
7 Testing of Transformers
This is nothing but the resistance measurement of the windings by applying a small
d.c voltage to the winding and measuring the current through the same. The ratio gives
the winding resistance, more commonly feasible with high voltage windings. For low voltage
windings a resistance-bridge method can be used. From the d.c resistance one can get the
a.c. resistance by applying skin effect corrections.
44
V3
S
A2
Vs V1 a2 a2
A2
~ V2
+ +
V
-
A1 a1 a1
A1
(a)A.C.test (b)D.C.test
This is needed for identifying the primary and secondary phasor polarities. It is
a must for poly phase connections. Both a.c. and d.c methods can be used for detecting
the polarities of the induced emfs. The dot method discussed earlier is used to indicate the
polarities. The transformer is connected to a low voltage a.c. source with the connections
made as shown in the fig. 18(a). A supply voltage Vs is applied to the primary and the
readings of the voltmeters V1 , V2 and V3 are noted. V1 : V2 gives the turns ratio. If V3 reads
V1 −V2 then assumed dot locations are correct (for the connection shown). The beginning and
end of the primary and secondary may then be marked by A1 − A2 and a1 − a2 respectively.
If the voltage rises from A1 to A2 in the primary, at any instant it does so from a1 to a2 in
the secondary. If more secondary terminals are present due to taps taken from the windings
they can be labeled as a3 , a4 , a5 , a6 . It is the voltage rising from smaller number towards
larger ones in each winding. The same thing holds good if more secondaries are present.
45
Fig. 18(b) shows the d.c. method of testing the polarity. When the switch S is closed if the
secondary voltage shows a positive reading, with a moving coil meter, the assumed polarity
is correct. If the meter kicks back the assumed polarity is wrong.
A Io
Im Ic
V1 V V2
V1 jXm
Rc
(a)Physical Arrangement
(b)Equivalent Circuit
As the name suggests, the secondary is kept open circuited and nominal value
of the input voltage is applied to the primary winding and the input current and power are
measured. In Fig. 19(a) V, A, W are the voltmeter, ammeter and wattmeter respectively.
Let these meters read V1 , I0 and W0 respectively.Fig. 19(b) shows the equivalent circuit of
the transformer under this test. The no load current at rated voltage is less than 1 percent of
nominal current and hence the loss and drop that take place in primary impedance r1 + jxl1
due to the no load current I0 is negligible. The active component Ic of the no load current I0
46
represents the core losses and reactive current Im is the current needed for the magnetization.
Thus the watt meter reading
W0 = V1 Ic = Pcore (38)
W0
∴ Ic = (39)
V
q1
∴ Im = I02 − Ic2 or (40)
V1 V1
Rc = andXm = (41)
Ic Im
V1
Io
The parameters measured already are in terms of the primary. Sometimes the pri-
mary voltage required may be in kilo-Volts and it may not be feasible to apply nominal
voltage to primary from the point of safety to personnel and equipment. If the secondary
voltage is low, one can perform the test with LV side energized keeping the HV side open
circuited. In this case the parameters that are obtained are in terms of LV . These have to
be referred to HV side if we need the equivalent circuit referred to HV side.
47
Sometimes the nominal value of high voltage itself may not be known, or in doubt,
especially in a rewound transformer. In such cases an open circuit characteristics is first
obtained, which is a graph showing the applied voltage as a function of the no load current.
This is a non linear curve as shown in Fig. 20. This graph is obtained by noting the current
drawn by transformer at different applied voltage, keeping the secondary open circuited. The
usual operating point selected for operation lies at some standard voltage around the knee
point of the characteristic. After this value is chosen as the nominal value the parameters
are calculated as mentioned above.
The purpose of this test is to determine the series branch parameters of the equiv-
alent circuit of Fig. 21(b). As the name suggests, in this test primary applied voltage, the
current and power input are measured keeping the secondary terminals short circuited. Let
these values be Vsc , Isc and Wsc respectively. The supply voltage required to circulate rated
current through the transformer is usually very small and is of the order of a few percent
of the nominal voltage. The excitation current which is only 1 percent or less even at rated
voltage becomes negligibly small during this test and hence is neglected. The shunt branch
′
is thus assumed to be absent. Also I1 = I2 as I0 ≃ 0. Therefore Wsc is the sum of the
copper losses in primary and secondary put together. The reactive power consumed is that
absorbed by the leakage reactance of the two windings.
2 ′
Wsc = Isc (r1 + r2 ) (42)
Vsc
Zsc = (43)
I
qsc
′
(xl1 + xl2 ) = 2 − (r + r ′ )2
Zsc (44)
1 2
48
W
A
Vsc
V
(a)Physical Arrangement
r1 jxl1 r’2 jx’l2
Isc
Vsc
(b)Equivalent Circuit
49
Load Test helps to determine the total loss that takes place, when the transformer
is loaded. Unlike the tests described previously, in the present case nominal voltage is applied
across the primary and rated current is drown from the secondary. Load test is used mainly
1. to determine the rated load of the machine and the temperature rise
Rated load is determined by loading the transformer on a continuous basis and observ-
ing the steady state temperature rise. The losses that are generated inside the transformer
on load appear as heat. This heats the transformer and the temperature of the transformer
increases. The insulation of the transformer is the one to get affected by this rise in the
temperature. Both paper and oil which are used for insulation in the transformer start get-
ting degenerated and get decomposed. If the flash point of the oil is reached the transformer
goes up in flames. Hence to have a reasonable life expectancy the loading of the transformer
must be limited to that value which gives the maximum temperature rise tolerated by the
insulation. This aspect of temperature rise cannot be guessed from the electrical equivalent
circuit. Further, the losses like dielectric losses and stray load losses are not modeled in the
50
equivalent circuit and the actual loss under load condition will be in error to that extent.
Many external means of removal of heat from the transformer in the form of different cooling
methods give rise to different values for temperature rise of insulation. Hence these permit
different levels of loading for the same transformer. Hence the only sure way of ascertaining
the rating is by conducting a load test.
Further, the temperature rise of the transformer is due to the losses that take place
‘inside’ the transformer. The efficiency of the transformer is above 99% even in modest sizes
which means 1 percent of power handled by the transformer actually goes to heat up the
machine. The remaining 99% of the power has to be dissipated in a load impedance external
to the machine. This is very wasteful in terms of energy also. ( If the load is of unity power
factor) Thus the actual loading of the transformer is seldom resorted to. Equivalent loss
methods of loading and ‘Phantom’ loading are commonly used in the case of transformers.
The load is applied and held constant till the temperature rise of transformer reaches a
steady value. If the final steady temperature rise is lower than the maximum permissible
value, then load can be increased else it is decreased. That load current which gives the
maximum permissible temperature rise is declared as the nominal or rated load current and
the volt amperes are computed using the same.
51
In the equivalent loss method a short circuit test is done on the transformer. The
short circuit current is so chosen that the resulting loss taking place inside the transformer
is equivalent to the sum of the iron losses, full load copper losses and assumed stray load
losses. By this method even though one can pump in equivalent loss inside the transformer,
the actual distribution of this loss vastly differs from that taking place in reality. Therefore
this test comes close to a load test but does not replace one.
W 1
A
2Io Io Io
V1
V
I’2 I2 I’2
I2
W 2
A
Vs V
In Phantom loading method two identical transformers are needed. The windings
are connected back to back as shown in Fig. 22. Suitable voltage is injected into the loop
formed by the two secondaries such that full load current passes through them. An equiv-
52
alent current then passes through the primary also. The voltage source V1 supplies the
magnetizing current and core losses for the two transformers. The second source supplies
the load component of the current and losses due to the same. There is no power wasted
in a load ( as a matter of fact there is no real load at all) and hence the name Phantom
or virtual loading. The power absorbed by the second transformer which acts as a load is
pushed back in to the mains. The two sources put together meet the core and copper losses
of the two transformers. The transformers work with full flux drawing full load currents and
hence are closest to the actual loading condition with a physical load.
53
As stated earlier, transformers of various sizes, ratings, voltage ratios can be seen
being used in a power system. The parameters of the equivalent circuits of these machines
also vary over a large range. Also the comparison of these machines are made simple if all
the parameters are normalized. If simple scaling of the parameters is done then one has
to carry forward the scaling factors in the calculations. Expressing in percent basis is one
example of scaling. However if the scaling is done on a logical basis one can have a simple
representation of the parameters without the bother of the scaling factors. Also different
units of measurement are in use in the different countries (FPS, CGS, MKS, etc;). These
units also underwent several revisions over the years. If the transformer parameter can be
freed from the units then the system becomes very simple. The ‘per unit’ system is developed
keeping these aspects in mind. The parameters of the transformer are referred to some base
values and thus get scaled. In the case of power system a common base value is adopted
in view of different ratings of the equipments used. In the case of individual equipments,
its own nominal parameters are used as base values. Some base parameters can be chosen
as independent base values while some others become derived base parameters. Once the
base values are identified the per unit values are calculated for any parameter by dividing
the same by its base value. The units must be the same for both the parameters and their
bases. Thus the per unit value is a unit-less dimensionless number. Let us choose nominal
voltage and nominal current on the primary side of a transformer as the base values Vbase
and Ibase . Other base values like volt ampere Sbase , short circuit impedance Zbase can be
54
Normally Sbase and Vbase are known from name plate details. Other base values can be
derived from them.
V (volt)
Vp.u = ,
Vbase (volt)
I(Amps) I(amps)
Ip.u = = Sbase (48)
Ibase (amps) V base
Many times, when more transformers are involved in a circuit one is required to choose
a common base value for all of them. Parameters of all the machines are expressed on this
common base. This is a common problem encountered in the case of parallel operation of
two or more transformers. The conversion of the base values naturally lead to change in the
per unit values of their parameters. An impedance Zp.u.old on the old base of Sbaseold and
Vbaseold shall get modified on new base Sbasenew ,Vbasenew as
2
Vbase old Sbase new
Zp.u.new = (Zp.u.old. ) 2
(50)
Sbase old Vbase new
The term inside the bracket is nothing but the ohmic value of the impedance and this gets
converted into the new per unit value by the new Sbase and Vbase .
If all the equivalent circuit parameters are referred to the secondary side and per unit
values of the new equivalent circuit parameters are computed with secondary voltage and
55
current as the base values, there is no change in the per unit values. This can be easily seen by,
′
′ ′ S ′ 1
Zp.u. = Zohm . base
′2 but Zohm = .Zohm (51)
Vbase a2
Where a - is the turns ratio of primary to secondary
Z - impedance as seen by primary,
′
Z - impedance as seen by secondary.
′
Sbase = Sbase - as the transformer rating is unaltered.
′
Vbase = Vbase . a1
′
From the above relationships it can be seen that Zp.u. = Zp.u..
This becomes obvious if we realize that the mmf of the core for establishing a given
flux is the same whether it is supplied through primary or the secondary. Also the active
power and reactive power absorbed inside the transformer are not dependant on the winding
connected to supply. This is further illustrated by taking the equivalent circuit of a trans-
former derived earlier and expressing the same in per unit form.
Thus the per unit values help in dispensing away the scaling constants. The veracity
of the parameters can be readily checked. Comparison of the parameters of the machines
with those of similar ones throw in useful information about the machines. Comparing the
efficiencies of two transformers at any load one can say that the transformer with a higher
p.u.resistance has higher copper losses without actually computing the same.
Application of per unit values for the calculation of voltage regulation, efficiency and
load sharing of parallel connected transformers will be discussed later at appropriate places.
56
9 Voltage Regulation
Modern power systems operate at some standard voltages. The equipments work-
ing on these systems are therefore given input voltages at these standard values, within
certain agreed tolerance limits. In many applications this voltage itself may not be good
enough for obtaining the best operating condition for the loads. A transformer is interposed
in between the load and the supply terminals in such cases. There are additional drops
inside the transformer due to the load currents. While input voltage is the responsibility of
the supply provider, the voltage at the load is the one which the user has to worry about.
If undue voltage drop is permitted to occur inside the transformer the load voltage becomes
too low and affects its performance. It is therefore necessary to quantify the drop that takes
place inside a transformer when certain load current, at any power factor, is drawn from its
output leads. This drop is termed as the voltage regulation and is expressed as a ratio of
the terminal voltage (the absolute value per se is not too important).
The voltage regulation can be defined in two ways - Regulation Down and Regulation
up. These two definitions differ only in the reference voltage as can be seen below.
Regulation down: This is defined as ” the change in terminal voltage when a load current
at any power factor is applied, expressed as a fraction of the no-load terminal voltage”.
Expressed in symbolic form we have,
|Vnl | − |Vl |
Regulation = (52)
|Vnl |
Vnl and Vl are no-load and load terminal voltages. This is the definition normally used
in the case of the transformers, the no-load voltage being the one given by the power
57
supply provider on which the user has no say. Hence no-load voltage is taken as the
reference.
Regulation up: Here again the regulation is expressed as the ratio of the change in the
terminal voltage when a load at a given power factor is thrown off, and the on load
voltage. This definition if expressed in symbolic form results in
|Vnl | − |Vl |
Regulation = (53)
|Vl |
This definition is more commonly used in the case of alternators and power systems
as the user-end voltage is guaranteed by the power supply provider. He has to generate
proper no-load voltage at the generating station to provide the user the voltage he has asked
for. In the expressions for the regulation, only the numerical differences of the voltages are
taken and not vector differences.
In the case of transformers both definitions result in more or less the same value for
the regulation as the transformer impedance is very low and the power factor of operation is
quite high. The power factor of the load is defined with respect to the terminal voltage on
load. Hence a convenient starting point is the load voltage. Also the full load output voltage
is taken from the name plate. Hence regulation up has some advantage when it comes to its
application. Fig. 23 shows the phasor diagram of operation of the transformer under loaded
′
condition. The no-load current I0 is neglected in view of the large magnitude of I2 . Then
58
Re jXe
I’2
V1 V’2
A
O
θ B
C E
φ V’2 I2’Xe
I2’Re
I2’
(b)Phasor Diagram
59
′
I1 = I2 .
′ ′
V1 = I2 (Re + jXe ) + V2 (54)
p
OD = V1 = [OA + AB + BC]2 + [CD]2
q
′ ′ ′ ′ ′
= [V2 + I2 Re cos φ + I2 Xe sin φ]2 + [I2 Xe cos φ − I2 Re sin φ]2 (55)
Also,
′ ′
V1 = V2 + I2 .(Re + jXe ) (56)
′ ′
= V2 + I2 (cos φ − j sin φ)(Re + jXe )
′
|V1 | − |V2 |
q
∴ RegulationR = ′ = (1 + v1 )2 + v22 − 1 (57)
|V2 |
′
I2 Xe
ex = V2
′ =per unit reactance drop
as v1 and v2 are small.
v22 v2
∴ R ≃ 1 + v1 + − 1 ≃ v1 + 2 (61)
2(1 + e1 ) 2
(ex sin φ − er cos φ)2
∴ regulation R = er cos φ ± ex sin φ + (62)
2
60
Powers higher than 2 for v1 and v2 are negligible as v1 and v2 are already small. As
v2 is small its second power may be neglected as a further approximation and the expression
for the regulation of the transform boils down to
regulation R = er cos φ ± ex sin φ
The negative sign is applicable when the power factor is leading. It can be seen from
the above expression, the full load regulation becomes zero when the power factor is leading
and er cos φ = ex sin φ or tan φ = er /ex
or the power factor angle φ = tan−1 (er /ex ) = tan−1 (Re /Xe ) leading.
61
V1
D
V2 I2’Xe
θ
O φ
I2’Re C
I2’
A
B
Thus this expression may not be as convenient as the earlier one due to the square root
involved.
Fig. 25 shows the variation of full load regulation of a typical transformer as the
power factor is varied from zero power factor leading, through unity power factor, to zero
power factor lagging.
It is seen from Fig. 25 that the full load regulation at unity power factor is nothing but
the percentage resistance of the transformer. It is therefore very small and negligible. Only
with low power factor loads the drop in the series impedance of the transformer contributes
substantially to the regulation. In small transformers the designer tends to keep the Xe very
low (less than 5%) so that the regulation performance of the transformer is satisfactory.
62
3
%Regulation
2
power factor
1
-2
-3
-4
-5
63
A low value of the short circuit impedance /reactance results in a large short circuit
current in case of a short circuit. This in turn results in large mechanical forces on the
winding. So, in large transformers the short circuit impedance is made high to give better
short circuit protection to the transformer which results in poorer regulation performance.
In the case of transformers provided with taps on windings, so that the turns ratio can be
changed, the voltage regulation is not a serious issue. In other cases care has to be exercised
in the selection of the short circuit impedance as it affects the voltage regulation.
64
10 Efficiency
Transformers which are connected to the power supplies and loads and are in op-
eration are required to handle load current and power as per the requirements of the load.
An unloaded transformer draws only the magnetization current on the primary side, the
secondary current being zero. As the load is increased the primary and secondary currents
increase as per the load requirements. The volt amperes and wattage handled by the trans-
former also increases. Due to the presence of no load losses and I 2 R losses in the windings
certain amount of electrical energy gets dissipated as heat inside the transformer. This gives
rise to the concept of efficiency.
Efficiency of a power equipment is defined at any load as the ratio of the power output to
the power input. Putting in the form of an expression,
output power Input power − losses inside the machine
Ef f iciency η = = (69)
input power Input power
losses inside the machine
= 1− = 1 − def f iciency
input power
output power
=
output + losses inside the machine
output power
More conveniently the efficiency is expressed in percentage. %η = input power
∗ 100
While the efficiency tells us the fraction of the input power delivered to the load, the
deficiency focuses our attention on losses taking place inside transformer. As a matter of
fact the losses heat up machine. The temperature rise decides the rating of the equipment.
The temperature rise of the machine is a function of heat generated the structural configu-
ration, method of cooling and type of loading (or duty cycle of load). The peak temperature
attained directly affects the life of the insulations of the machine for any class of insulation.
These aspects are briefly mentioned under section 7.5 on load test.
65
100
Efficiency% 75
50
25
0 0.5 x 1
A typical curve for the variation of efficiency as a function of output is given in Fig. 26
The losses that take place inside the machine expressed as a fraction of the input is some
times termed as deficiency. Except in the case of an ideal machine, a certain fraction of
the input power gets lost inside the machine while handling the power. Thus the value for
the efficiency is always less than one. In the case of a.c. machines the rating is expressed
in terms of apparent power. It is nothing but the product of the applied voltage and the
current drawn. The actual power delivered is a function of the power factor at which this
current is drawn. As the reactive power shuttles between the source and the load and has a
zero average value over a cycle of the supply wave it does not have any direct effect on the
efficiency. The reactive power however increases the current handled by the machine and
the losses resulting from it. Therefore the losses that take place inside a transformer at any
66
given load play a vital role in determining the efficiency. The losses taking place inside a
transformer can be enumerated as below:
3. Iron loss
4. Dielectric loss
Primary and secondary copper losses take place in the respective winding resistances
due to the flow of the current in them.
′
Pc = I12 r1 + I22 r2 = I22 Re (70)
The primary and secondary resistances differ from their d.c. values due to skin effect and the
temperature rise of the windings. While the average temperature rise can be approximately
used, the skin effect is harder to get analytically. The short circuit test gives the value of Re
taking into account the skin effect.
The iron losses contain two components - Hysteresis loss and Eddy current loss. The
Hysteresis loss is a function of the material used for the core.
Ph = Kh B 1.6 f
67
For constant voltage and constant frequency operation this can be taken to be con-
stant. The eddy current loss in the core arises because of the induced emf in the steel
lamination sheets and the eddies of current formed due to it. This again produces a power
loss Pe in the lamination.
Pe = Ke B 2 f 2 t2
where t is the thickness of the steel lamination used. As the lamination thickness is much
smaller than the depth of penetration of the field, the eddy current loss can be reduced by
reducing the thickness of the lamination. Present day laminations are of 0.25 mm thickness
and are capable of operation at 2 Tesla. These reduce the eddy current losses in the core.
This loss also remains constant due to constant voltage and frequency of operation. The
sum of hysteresis and eddy current losses can be obtained by the open circuit test.
The dielectric losses take place in the insulation of the transformer due to the large
electric stress. In the case of low voltage transformers this can be neglected. For constant
voltage operation this can be assumed to be a constant.
The stray load losses arise out of the leakage fluxes of the transformer. These leakage
fluxes link the metallic structural parts, tank etc. and produce eddy current losses in them.
Thus they take place ’all round’ the transformer instead of a definite place , hence the name
’stray’. Also the leakage flux is directly proportional to the load current unlike the mutual
flux which is proportional to the applied voltage. Hence this loss is called ’stray load’ loss.
This can also be estimated experimentally. It can be modeled by another resistance in the
series branch in the equivalent circuit. The stray load losses are very low in air-cored trans-
formers due to the absence of the metallic tank.
68
Thus, the different losses fall in to two categories Constant losses (mainly voltage
dependant) and Variable losses (current dependant). The expression for the efficiency of the
transformer operating at a fractional load x of its rating, at a load power factor of θ2 , can
be written as
xS cos θ2
η= (71)
xS cos θ2 + Pconst + x2 Pvar
′ ′
Here S in the volt ampere rating of the transformer (V2 I2 at full load), Pconst being constant
losses and Pvar the variable losses at full load.
For a given power factor an expression for η in terms of the variable x is thus obtained.
By differentiating η with respect to x and equating the same to zero, the condition for
maximum efficiency is obtained. In the present case that condition comes out to be
r
2 Pconst
Pconst = x Pvar or x = (72)
Pvar
That is, when constant losses equal the variable losses at any fractional load x the
efficiency reaches a maximum value. The maximum value of that efficiency at any given
power factor is given by,
xS cos θ2 xS cos θ2
ηmax = = (73)
xS cos θ2 + 2Pconst xS cos θ2 + 2x2 Pvar
From the expression for the maximum efficiency it can be easily deduced that this
maximum value increases with increase in power factor and is zero at zero power factor of
the load. It may be considered a good practice to select the operating load point to be at the
maximum efficiency point. Thus if a transformer is on full load, for most part of the time
then the ηmax can be made to occur at full load by proper selection of constant and variable
69
losses. However, in the modern transformers the iron losses are so low that it is practically
impossible to reduce the full load copper losses to that value. Such a design wastes lot of
copper. This point is illustrated with the help of an example below.
Two 100 kVA transformers A nd B are taken. Both transformers have total full load
losses to be 2 kW. The break up of this loss is chosen to be different for the two transformers.
Transformer A: iron loss 1 kW, and copper loss is 1 kW. The maximum efficiency of 98.04%
occurs at full load at unity power factor. Transformer B: Iron loss =0.3 kW and full load
copper loss =1.7 kW. This also has a full load η of 98.04%. Its maximum η occurs at
q
0.3
a fractional load of 1.7
= 0.42. The maximum efficiency at unity power factor being
42
42+0.6
∗ 100 = 98.59%. At the corresponding point the transformer A has an efficiency of
42
42+1.0+0.1764
∗ 100 = 97.28%. Transformer A uses iron of more loss per kg at a given flux
density, but transformer B uses lesser quantity of copper and works at higher current density.
Large capacity transformers used in power systems are classified broadly into Power trans-
formers and Distribution transformers. The former variety is seen in generating stations and
large substations. Distribution transformers are seen at the distribution substations. The
basic difference between the two types arise from the fact that the power transformers are
switched in or out of the circuit depending upon the load to be handled by them. Thus at
50% load on the station only 50% of the transformers need to be connected in the circuit.
On the other hand a distribution transformer is never switched off. It has to remain in the
circuit irrespective of the load connected. In such cases the constant loss of the transformer
continues to be dissipated. Hence the concept of energy based efficiency is defined for such
70
100 100
50 50
Power Loss %
Load % of full load
6 12 18 24
Time,hrs 12 24
71
transformers. It is called ’all day’ efficiency. The all day efficiency is thus the ratio of the
energy output of the transformer over a day to the corresponding energy input. One day
is taken as a duration of time over which the load pattern repeats itself. This assumption,
however, is far from being true. The power output varies from zero to full load depending
on the requirement of the user and the load losses vary as the square of the fractional loads.
The no-load losses or constant losses occur throughout the 24 hours. Thus, the comparison
of loads on different days becomes difficult. Even the load factor, which is given by the
ratio of the average load to rated load, does not give satisfactory results. The calculation
of the all day efficiency is illustrated below with an example. The graph of load on the
transformer, expressed as a fraction of the full load is plotted against time in Fig. 27. In an
actual situation the load on the transformer continuously changes. This has been presented
by a stepped curve for convenience. The average load can be calculated by
Pn
Sn ni=1 xi ti cos θi
P
i=1 Pi
Average load over a day = = (74)
24 24
where Pi is the load during an interval i. n intervals are assumed. xi is the fractional load.
Si = xi Sn where Sn is nominal load. The average loss during the day is given by
Pn
Pc i=1 x2i ti
Average loss = Pi + (75)
24
This is a non-linear function. For the same load factor different average loss can be
there depending upon the values of xi and ti . Hence a better option would be to keep the
constant losses very low to keep the all day efficiency high. Variable losses are related to
load and are associated with revenue earned. The constant losses on the other hand has to
be incurred to make the service available. The concept of all day efficiency may therefore be
more useful for comparing two transformers subjected to the same load cycle.
72
The concept of minimizing the lost energy comes into effect right from the time of
procurement of the transformer. The constant losses and variable losses are capitalized and
added to the material cost of the transformer in order to select the most competitive one,
which gives minimum cost taking initial cost and running cost put together. Obviously the
iron losses are capitalized more in the process to give an effect to the maximization of energy
efficiency. If the load cycle is known at this stage, it can also be incorporated in computation
of the best transformer.
73
11 Auto Transformer
I1
C
T1 I2
B
V1
T2 V2 ZL
I2
A
I1
V1 : V2 = T1 : T2 (76)
74
When the secondary winding delivers a load current of I2 ampere the demagnetizing ampere
turns is I2 T2 . This will be countered by a current I1 flowing from the source through the
T1 turns such that,
I1 T1 = I2 T2 (77)
A current of I1 ampere flows through the winding between B and C . The current
in the winding between A and B is (I2 − I1 ) ampere. The cross section of the wire to be
selected for AB is proportional to this current assuming a constant current density for the
whole winding. Thus some amount of material saving can be achieved compared to a two
winding transformer. The magnetic circuit is assumed to be identical and hence there is
no saving in the same. To quantify the saving the total quantity of copper used in an auto
transformer is expressed as a fraction of that used in a two winding transformer as,
This means that an auto transformer requires the use of lesser quantity of copper
given by the ratio of turns. This ratio therefore denotes the savings in copper. As the
space for the second winding need not be there, the window space can be less for an auto
transformer, giving some saving in the lamination weight also. The larger the ratio of the
voltages, smaller is the savings. As T2 approaches T1 the savings become significant. Thus
auto transformers become ideal choice for close ratio transformations. The savings in mate-
rial is obtained, however, at a price. The electrical isolation between primary and secondary
75
φ
I1+I2 I1 I2
I2 V2
V1
V1+V2 ZL
I1+I2 I1
I2
has to be sacrificed.
If we are not looking at the savings in the material, even then going in for the auto
transformer type of connection can be used with advantage, to obtain higher output. This
can be illustrated as follows. Fig. 29 shows a regular two winding transformer of a voltage
ratio V1 : V2 , the volt ampere rating being V1 I1 = V2 I2 = S. If now the primary is connected
across a supply of V1 volt and the secondary is connected in series addition manner with the
primary winding, the output voltage becomes (V1 + V2 ) volt. The new output of this auto
transformer will now be
V1 V1
I2 (V1 + V2 ) = I2 V2 (1 + ) = S(1 + ) (81)
V2 V2
I2
= V1 (I1 + I2 ) = S(1 + ) (82)
I1
76
V1 V1
I2 (V1 − V2 ) = I2 V2 ( − 1) = S( − 1) (83)
V2 V2
I1
r1,xl1 I1 I2
V1
r2,xl2 V2
(I2 -I1)
I1 I2
As mentioned earlier the magnetizing current can be neglected, for simplicity. Writing
77
Note that the resistance r1 and leakage reactance xl1 refer to that part of the winding where
only the primary current flows. Similarly on the load side we have,
= aV2 + I1 (r1 + jxl1 + r2 + jxl2 + a2 r2 + ja2 xl2 − ar2 − ajxl2 − ar2 − jaxl2 )
Equation (85) yields the equivalent circuit of Fig. 31 where Re = r1 + (a − 1)2 r2 and
Xe = xl1 + (a − 1)2 xl2 .
The magnetization branch can now be hung across the mains for completeness. The
above equivalent circuit can now be compared with the approximate equivalent circuit of
a two winding case Re = r1 + a2 r2 and Xe = xl1 + a2 xl2 . Thus in the case of an auto
transformer total value of the short circuit impedance is lower and so also the percentage
resistance and reactance. Thus the full load regulation is lower. Having a smaller value
of short circuit impedance is sometimes considered to be a disadvantage. That is because
78
Re=r1+(a-1)2r2
Re jXe Xl=xl1+(a-1)2xl2
Io
Ic Im
V1 Rc jXm V’2=aV1
the short circuit currents become very large in those cases. The efficiency is higher in auto
transformers compared to their two winding counter part at the same load. The phasor
diagram of operation for the auto transformer drawing a load current at a lagging power
factor angle of θ2 is shown in Fig. 32. The magnetizing current is omitted here again for
simplicity.
From the foregoing study it is seen that there are several advantages in going in for the
autotransformer type of arrangement. The voltage/current transformation and impedance
conversion aspects of a two winding transformer are retained but with lesser material (and
hence lesser weight) used. The losses are reduced increasing the efficiency. Reactance is
reduced resulting in better regulation characteristics. All these benefits are enhanced as
the voltage ratio approaches unity. The price that is required to be paid is loss of electri-
cal isolation and a larger short circuit current (and larger short circuit forces on the winding).
79
I1x1 (I2-I1)r2
I1r1
(I2-I1)x2
V1
I2
E1
E2 (I2-I1)x2
I1
(I2-I1)r2
V2
θ2
θ1
I2
φ
80
Auto transformers are used in applications where electrical isolation is not a critical
requirement. When the ratio V2 : V1 is 0.3 or more they are used with advantage. The
normal applications are motor starters, boosters or static balancers.
Variable
a.c output
V in
M oving contact
The conductor is selected based on the maximum secondary current that could be
drawn as the output voltage varies in practically continuous manner. These are used in
81
voltage stabilizers, variable d.c. arrangements (with a diode bridge) in laboratories, motor
starters, dimmers etc.
82
12 Harmonics
83
ϕ
e ϕ
ϕ’’ ϕ’’
ϕ’
ϕ’
i’ϕ i’’ϕ
t’ t’’ t iϕ
i’ϕ
’’
iϕ
iϕ
84
magnitude of 3rd harmonic voltage. These effects get even more pronounced for higher
order harmonics. As these harmonic voltages get short circuited through the low impedance
of the supply they produce harmonic currents. These currents produce effects according to
Lenz’s law and tend to neutralize the harmonic flux and bring the flux wave to a sinusoid.
Normally third harmonic is the largest in its magnitude and hence the discussion is based on
it. The same can be told of other harmonics also. In the case of a single phase transformer
the harmonics are confined mostly to the primary side as the source impedance is much
smaller compared to the load impedance. The understanding of the phenomenon becomes
more clear if the transformer is supplied with a sinusoidal current source. In this case current
has to be sinusoidal and the harmonic currents cannot be supplied by the source and hence
the induced emf will be peaky containing harmonic voltages. When the load is connected on
the secondary side the harmonic currents flow through the load and voltage tends to become
sinusoidal. The harmonic voltages induce electric stress on dielectrics and increased electro
static interference. The harmonic currents produce losses and electro magnetic interference
as already noted above.
In the case of single phase transformers connected to form three phase bank, each
transformer is magnetically decoupled from the other. The flow of harmonic currents are
decided by the type of the electrical connection used on the primary and secondary sides.
Also, there are three fundamental voltages in the present case each displaced from the other
by 120 electrical degrees. Because of the symmetry of the a.c. wave about the time axis
only odd harmonics need to be considered. The harmonics which are triplen (multiples of
three) behave in a similar manner as they are co-phasal or in phase in the three phases. The
85
non-triplen harmonics behave in a similar manner to the fundamental and have ±120◦ phase
displacement between them. The harmonic behavior of poly-phase banks can be discussed
now.
Dd connection In three phase banks with mesh connection on both primary side and sec-
ondary side a closed path is available for the triplen harmonics to circulate currents.
Thus the supply current is nearly sinusoidal (but for the non-triplen harmonic cur-
rents). The triplen harmonic currents inside the closed mesh winding correct the flux
density wave to be nearly sinusoidal. The secondary voltages will be nearly sinusoidal.
Third harmonics currents flow both in the primary and the secondary and hence the
magnitudes of these currents, so also the drops due to them will be lower.
Dy and Yd connection (without neutral connection) Behavior of the bank with mesh
connection on one side is similar to the one discussed under Dd connection. The har-
monic currents and drops and the departure of the flux density from sinusoidal are
larger in the present case compared to Dd banks.
Yy connection without neutral wires With both primary and secondary connected in
star no closed path exists. As the triplen harmonics are always in phase, by virtue
of the Y connection they get canceled in the line voltages. Non-triplen harmonics
√
like fundamental, become 3 times phase value and appear in the line voltages. Line
currents remain sinusoidal except for non-triplen harmonic currents. Flux wave in each
transformer will be flat topped and the phase voltages remain peaked. The potential
of the neutral is no longer steady. The star point oscillates due to the third harmonic
voltages. This is termed as ”oscillating neutral”.
Yy connection with neutral wires When a neutral wire is provided the triplen har-
86
monic current can flow and the condition is similar to the single phase case (with a
star connected 4 wire source or with the system earth). The neutral wire carries three
times the triplen harmonic current of one transformer as these currents are co-phasal.
Unloaded secondary neutral will not be operative. Other polyphase connections not
discussed above explicitly will fall under one type or the other of the cases discussed.
In a Yy connection, to obtain third harmonic suppression one may provide a third
winding which is connected in mesh, which can be an unloaded winding. It is called a
tertiary. This winding improves the single phase to earth fault detection also. Further,
this winding can be used to feed some permanent station loads also. Such transform-
ers are designated as Yyd transformers. If the neutral wires are provided and also
mesh connected winding is present, then triplen harmonics are ’shared’ between them
depending upon their impedances.
87
high reluctance. Thus the flux in the core remains nearly sinusoidal, so also the induced emf.
This happens irrespective of the type of connection used. The triplen order flux, sometimes
links the tank and produces loss in the same.
88
ers
89
circuit are all necessary for the same. To facilitate the manufacturer and users, interna-
tional standards are also available. Each winding has two ends designated as 1 and 2. The
HV winding is indicated by capital letters and the LV winding by small letters. If more
terminals are brought out from a winding by way of taps there are numbered in the increas-
ing numbers in accordance to their distance from 1 (eg A1 , A2 , A3 ...). If the induced emf
at an instant is from A1 to A2 on the HV winding it will rise from a1 to a2 on the LV winding.
Out of the different polyphase connections three phase connections are mostly en-
countered due to the wide spread use of three phase systems for generation, transmission
and utilization. Three balanced 3-phase voltages can be connected in star or mesh fashion
to yield a balanced 3-phase 3-wire system. The transformers that work on the 3-phase sup-
ply have star, mesh or zig-zag connected windings on either primary secondary or both. In
addition to giving different voltage ratios, they introduce phase shifts between input and
output sides. These connections are broadly classified into 4 popular vector groups.
1. Group I: zero phase displacement between the primary and the secondary.
3. Group III: 30◦ lag phase displacement of the secondary with respect to the primary.
4. Group IV: 30◦ lead phase displacement of the secondary with respect to the primary.
A few examples of the physical connections and phasor diagrams are shown in Fig. 35
and Fig. 36 corresponding to each group. The capital letters indicates primary and the
small letters the secondary. D/d stand for mesh, Y/y - for star, Z/z for zig-zag. The angu-
lar displacement of secondary with respect to the primary are shown as clock position, 0◦
90
C1 C2 B2
C 2 c2 c1
A2 a2
A1 A 2 a2 a1
B1 B 2 b2 b1
B2
C2
C1 C 2 c2 c1
n A2
A1 A 2 a4 a1
B1 B 2 b4 b1
c1
C2
C1 C 2 c4 B2
(a)
G roup2 180 0 Phase shift
C1 C 2 c1 C2
c1 B2
A2
A1 a2
A 2a1
a1
B1 b2
B 2b1
b1
c2
C1 C 2c1
c1
C2 B2
n A2
A1 A 2 a3
B1 a1
B 2b 3
b1
C1 C 2 c3
c1 C2 B2
(b)
C1 C 2 c2 c1
N C2 A2 B2 a2
A1 A 2 a2
a1
B1 B 2 b2
b1
C1 C 2 c2
c1
C2 B2
N n A2
A1 A 2 a4
a1
B1 B 2 b4
b1
C1 C 2 c4
c1 C2 B2
(a)
G roup4 + 30 0 Phase shift
B1 B 2 b2 b1
C1 c2 c1
C2 B2
A2
N a2
A1
A 2 a1
a1
B1 b2 C2
B 2 b1
b1
c2 B2
C1
C 2 c1 B2
c1
n a3
A2 C2
N
A1
A2 a4
b3 C2
B1 a1
B2 b4
b1
C1 c3 B2
C 2 c4
c1
(b)
referring to 12 o’clock position. These vector groups are especially important when two or
more transformers are to be connected in parallel.
Star connection is normally cheaper as there are fewer turns and lesser cost of insula-
tion. The advantage becomes more with increase in voltage above 11kv. In a star connected
winding with earthed-neutral the maximum voltage to the earth is ( √13 )of the line voltage.
Also star connection permits mixed loading due to the presence of the neutral. Mesh con-
nections are advantageous in low voltage transformers as insulation costs are insignificant
and the conductor size becomes ( √13 ) of that of star connection and permits ease of winding.
The common polyphase connections are briefly discussed now.
Star/star (Yy0, Yy6)connection This is the most economical one for small high voltage
transformers. Insulation cost is highly reduced. Neutral wire can permit mixed load-
ing. Triplen harmonics are absent in the lines. These triplen harmonic currents cannot
flow, unless there is a neutral wire. This connection produces oscillating neutral. Three
phase shell type units have large triplen harmonic phase voltage. However three phase
core type transformers work satisfactorily. A tertiary mesh connected winding may be
required to stabilize the oscillating neutral due to third harmonics in three phase banks.
Mesh/mesh (Dd0, Dd6) This is an economical configuration for large low voltage trans-
formers. Large amount of unbalanced load can be met with ease. Mesh permits a cir-
culating path for triplen harmonics thus attenuates the same. It is possible to operate
with one transformer removed in open delta or Vee connection meeting 58 percent of
the balanced load. Three phase units cannot have this facility. Mixed single phase
loading is not possible due to the absence of neutral.
93
Zig zag/ star (ZY1 or Zy11) Zigzag connection is obtained by inter connection of phases.
4-wire system is possible on both sides. Unbalanced loading is also possible. Oscillat-
ing neutral problem is absent in this connection. This connection requires 15% more
turns for the same voltage on the zigzag side and hence costs more.
Generally speaking a bank of three single phase transformers cost about 15% more
than their 3-phase counter part. Also, they occupy more space. But the spare capac-
ity cost will be less and single phase units are easier to transport.
Mesh connected three phase transformers resemble 3- single phase units but kept in
a common tank. In view of this single tank, the space occupied is less. Other than
that there is no big difference. The 3-phase core type transformer on the other hand
has a simple core arrangement. The three limbs are equal in cross section. Primary
and secondary of each phase are housed on the same limb. The flux setup in any limb
will return through the other two limbs as the mmf of those limbs are in the proper
directions so as to aid the same. Even though magnetically this is not a symmetrical
arrangement, as the reluctance to the flux setup by side limbs is different from that of
the central limb, it does not adversely affect the performance. This is due to the fact
that the magnetizing current itself forms a small fraction of the total phase current
drawn on load. The added advantage of 3-phase core is that it can tolerate substantially
94
large value of 3rd harmonic mmf without affecting the performance. The 3rd harmonic
mmf of the three phases will be in phase and hence rise in all the limbs together.
The 3rd harmonic flux must therefore find its path through the air. Due to the high
reluctance of the air path even a substantially large value of third harmonic mmf
produces negligible value of third harmonic flux. Similarly unbalanced operation of the
transformer with large zero sequence fundamental mmf content also does not affect its
performance. Even with Yy type of poly phase connection without neutral connection
the oscillating neutral does not occur with these cores. Finally, three phase cores
themselves cost less than three single phase units due to compactness.
95
formers
By parallel operation we mean two or more transformers are connected to the same
supply bus bars on the primary side and to a common bus bar/load on the secondary side.
Such requirement is frequently encountered in practice. The reasons that necessitate parallel
operation are as follows.
2. The power demand might have increased over a time necessitating augmentation of the
capacity. More transformers connected in parallel will then be pressed into service.
3. To ensure improved reliability. Even if one of the transformers gets into a fault or is
taken out for maintenance/repair the load can continued to be serviced.
4. To reduce the spare capacity. If many smaller size transformers are used one machine
can be used as spare. If only one large machine is feeding the load, a spare of similar
rating has to be available. The problem of spares becomes more acute with fewer
machines in service at a location.
Fig. 37 shows the physical arrangement of two single phase transformers working in
parallel on the primary side. Transformer A and Transformer B are connected to input
voltage bus bars. After ascertaining the polarities they are connected to output/load bus
96
V1 A IA V2
E1 E2
load
IB
E1 E2
B
supply bus Load bus
97
bars. Certain conditions have to be met before two or more transformers are connected in
parallel and share a common load satisfactorily. They are,
2. The per unit impedance of each machine on its own base must be the same.
3. The polarity must be the same, so that there is no circulating current between the
transformers.
4. The phase sequence must be the same and no phase difference must exist between the
voltages of the two transformers.
These conditions are examined first with reference to single phase transformers and then the
three phase cases are discussed.
Same voltage ratio Generally the turns ratio and voltage ratio are taken to be the same.
If the ratio is large there can be considerable error in the voltages even if the turns ratios
are the same. When the primaries are connected to same bus bars, if the secondaries
do not show the same voltage, paralleling them would result in a circulating current
between the secondaries. Reflected circulating current will be there on the primary
side also. Thus even without connecting a load considerable current can be drawn
by the transformers and they produce copper losses. In two identical transformers
with percentage impedance of 5 percent, a no-load voltage difference of one percent
will result in a circulating current of 10 percent of full load current. This circulating
current gets added to the load current when the load is connected resulting in unequal
sharing of the load. In such cases the combined full load of the two transformers can
never be met without one transformer getting overloaded.
98
Per unit impedance Transformers of different ratings may be required to operate in par-
allel. If they have to share the total load in proportion to their ratings the larger
machine has to draw more current. The voltage drop across each machine has to be
the same by virtue of their connection at the input and the output ends. Thus the
larger machines have smaller impedance and smaller machines must have larger ohmic
impedance. Thus the impedances must be in the inverse ratios of the ratings. As the
voltage drops must be the same the per unit impedance of each transformer on its
own base, must be equal. In addition if active and reactive power are required to be
shared in proportion to the ratings the impedance angles also must be the same. Thus
we have the requirement that per unit resistance and per unit reactance of both the
transformers must be the same for proper load sharing.
Polarity of connection The polarity of connection in the case of single phase transform-
ers can be either same or opposite. Inside the loop formed by the two secondaries
the resulting voltage must be zero. If wrong polarity is chosen the two voltages get
added and short circuit results. In the case of polyphase banks it is possible to have
permanent phase error between the phases with substantial circulating current. Such
transformer banks must not be connected in parallel. The turns ratios in such groups
can be adjusted to give very close voltage ratios but phase errors cannot be compen-
sated. Phase error of 0.6 degree gives rise to one percent difference in voltage. Hence
poly phase transformers belonging to the same vector group alone must be taken for
paralleling.
Transformers having −30◦ angle can be paralleled to that having +30◦ angle by re-
versing the phase sequence of both primary and secondary terminals of one of the
transformers. This way one can overcome the problem of the phase angle error.
99
Phase sequence The phase sequence of operation becomes relevant only in the case of
poly phase systems. The poly phase banks belonging to same vector group can be
connected in parallel. A transformer with +30◦ phase angle however can be paralleled
with the one with −30◦ phase angle, the phase sequence is reversed for one of them
both at primary and secondary terminals. If the phase sequences are not the same
then the two transformers cannot be connected in parallel even if they belong to same
vector group. The phase sequence can be found out by the use of a phase sequence
indicator.
Performance of two or more single phase transformers working in parallel can be com-
puted using their equivalent circuit. In the case of poly phase banks also the approach
is identical and the single phase equivalent circuit of the same can be used. Basically
two cases arise in these problems. Case A: when the voltage ratio of the two trans-
formers is the same and Case B: when the voltage ratios are not the same. These are
discussed now in sequence.
Always two transformers of equal voltage ratios are selected for working in parallel.
This way one can avoid a circulating current between the transformers. Load can be switched
on subsequently to these bus bars. Neglecting the parallel branch of the equivalent circuit the
above connection can be shown as in Fig. 38(a),(b). The equivalent circuit is drawn in terms
of the secondary parameters. This may be further simplified as shown under Fig. 38(c). The
voltage drop across the two transformers must be the same by virtue of common connection
at input as well as output ends. By inspection the voltage equation for the drop can be
100
jX A jX A
RA IA RA
V 2’ IA
V1
ZA
V1 jX B I V 2’
RB IB
RB jX B
IB
ZB
(a) (b)
ZA IA
I
ZB IB
VL
V ’L
V Load
ZL
(c)
Figure 38: Equivalent Circuit for Transformers working in Parallel -Simplified circuit and
Further simplification for identical voltage ratio
101
written as
IA ZA = IB ZB = IZ = v (say) (87)
HereI = IA + IB (88)
ZA ZB
Z= (89)
ZA + ZB
v IZ ZB
Thus IA = = = I. (90)
ZA ZA ZA + ZB
v IZ ZA
and IB = = = I.
ZB ZB ZA + ZB
If the terminal voltage is V = IZL then the active and reactive power supplied by each of
the two transformers is given by
From the above it is seen that the transformer with higher impedance supplies lesser
load current and vice versa. If transformers of dissimilar ratings are paralleled the trans-
former with larger rating shall have smaller impedance as it has to produce the same drop
as the other transformer, at a larger current. Thus the ohmic values of the impedances must
IA ZB
be in the inverse ratio of the ratings of the transformers. IA ZA = IB ZB , therefore IB
= ZA
.
Expressing the voltage drops in p.u basis, we aim at the same per unit drops at any load for
the transformers. The per unit impedances must therefore be the same on their respective
bases.
Fig. 39 shows the phasor diagram of operation for these conditions. The drops are
magnified and shown to improve clarity. It is seen that the total voltage drop inside the
102
IAXA
V2 IX IBXB
θA θB
φA
φB φ
IB
IR IBRB
IL
IA IARA
Figure 39: Phasor Diagram of Operation for two Transformers working in Parallel
transformers is v but the currents IA and IB are forced to have a different phase angle due
to the difference in the internal power factor angles θA and θB . This forces the active and
reactive components of the currents drawn by each transformer to be different ( even in
the case when current in each transformer is the same). If we want them to share the load
current in proportion to their ratings, their percentage ( or p.u) impedances must be the
same. In order to avoid any divergence and to share active and reactive powers also properly,
θA = θB . Thus the condition for satisfactory parallel operation is that the p.u resistances
and p.u reactance must be the same on their respective bases for the two transformers. To
determine the sharing of currents and power either p.u parameters or ohmic values can be
used.
103
RA jX A
IA
RB jX B I
IB
ZL VL
EA EB
One may not be able to get two transformers of identical voltage ratio in
spite of ones best efforts. Due to manufacturing differences, even in transformers built as
per the same design, the voltage ratios may not be the same. In such cases the circuit
representation for parallel operation will be different as shown in Fig. 40. In this case the
two input voltages cannot be merged to one, as they are different. The load brings about a
common connection at the output side. EA and EB are the no-load secondary emf. ZL is
the load impedance at the secondary terminals. By inspection the voltage equation can be
written as below:
EA = IA ZA + (IA + IB )ZL = V + IA ZA ·
Solving the two equations the expression for IA and IB can be obtained as
104
EA ZB + (EA − EB )ZL
IA = and (94)
ZA ZB + ZL (ZA + ZB )
EB ZA + (EB − EA )ZL
IB =
ZA ZB + ZL (ZA + ZB )
ZA and ZB are phasors and hence there can be angular difference also in addition to
the difference in magnitude. When load is not connected there will be a circulating current
between the transformers. The currents in that case can be obtained by putting ZL = ∞
( after dividing the numerator and the denominator by ZL ). Then,
(EA − EB )
IA = −IB = (95)
(ZA + ZB )
If the load impedance becomes zero as in the case of a short circuit, we have,
EA EB
IA = and IB = (96)
ZA ZB
Instead of the value of ZL if the value of V is known , the currents can be easily determined
( from Eqns. 93 ) as
EA − V EB − V
IA = and IB = (97)
ZA ZB
If more than two transformers are connected across a load then the calculation of
load currents following the method suggested above involves considerable amount of compu-
tational labor. A simpler and more elegant method for the case depicted in Fig. 41 is given
below. It is known by the name parallel generator theorem.
IL = IA + IB + IC + ......
EA − V EB − V EC − V
But IA = , IB = , IC =
ZA ZB ZC
V = IL .ZL (98)
105
RA jXA
IA
EA
RB jXB
IB I
EB
RC jXC
IC
EC
ZL V
V EA − V EB − V EC − V
= + + + ... (99)
ZL ZA ZB ZC
1 1 1 1 EA EB EC
V( + + + + ...) = + + + ...
ZL ZA ZB ZC ZA ZB ZC
= ISCA + ISCB + ISCC + .... (100)
1 1 1 1 1
( + + + + ...) = (101)
ZL ZA ZB ZC Z
V = Z(ISCA + ISCB + ISCC + ....) (102)
From this V can be obtained. Substituting V in Eqn. 98, IA , IB etc can be obtained.
Knowing the individual current phasor, the load shared by each transformer can be com-
puted.
106
On a power system the transformers are additionally required to perform the task of
regulation of active and reactive power flows.
Regulation B B
transformer Main
1 1
transformer
2
Reverser tertiary2 Reverser
The voltage control is performed by changing the turns ratio. This is done by provi-
sion of taps in the winding. The volts per turn available in large transformers is quite high
and hence a change of even one turn on the LV side represents a large percentage change
in the voltage. Also the LV currents are normally too large to take out the tapping from
the windings. LV winding being the inner winding in a core type transformer adds to the
107
difficulty of taking out of the taps. Hence irrespective of the end use for which tapping is put
to, taps are provided on the HV winding. Provision of taps to control voltage is called tap
changing. In the case of power systems, voltage levels are some times changed by injecting a
suitable voltage in series with the line. This may be called buck-boost arrangement. In ad-
dition to the magnitude, phase of the injected voltage may be varied in power systems. The
tap changing arrangement and buck boost arrangement with phase shift are shown in Fig. 42.
Tap changing can be effected when a) the transformers is on no- load and b) the load
is still remains connected to the transformer. These are called off load tap changing and
on load tap changing. The Off load tap changing relatively costs less. The tap positions
are changed when the transformer is taken out of the circuit and reconnected. The on-load
tap changer on the other hand tries to change the taps without the interruption of the load
current. In view of this requirement it normally costs more. A few schemes of on-load tap
changing are now discussed.
108
s 2
3
Reactor 4
HV LV
109
Reactor method The diagram of connections is shown in Fig. 43. This method employs
an auxiliary reactor to assist tap changing. The switches for the taps and that across
the reactor(S) are connected as shown. The reactor has a center tapped winding on
a magnetic core. The two ends of the reactor are connected to the two bus bars to
which tapping switches of odd/even numbered taps are connected. When only one tap
is connected to the reactor the shorting switch S is closed minimizing the drop in the
reactor. The reactor can also be worked with both ends connected to two successive
taps. In that case the switch ’S’ must be kept open. The reactor limits the circulating
current between the taps in such a situation. Thus a four step tapped winding can be
used for getting seven step voltage on the secondary(see the table of switching). The
advantage of this type of tap changer are
3. Switches need not interrupt load current as a alternate path is always provided.
110
The major objection to this scheme seems to be that the reactor is in the circuit always
generating extra loss.
Series booster method In this case a separate transformer is used to buck/boost the
voltage of the main transformer. The main transformer need not be having a tapped
arrangement. This arrangement can be added to an existing system also. Fig. 42shows
the booster arrangement for a single phase supply. The reverser switch reverses the
polarity of the injected voltage and hence a boost is converted into a buck and vice
versa. The power rating of this transformer need be a small fraction of the main
transformer as it is required to handle only the power associated with the injected
voltage. One precaution to be taken with this arrangement is that the winding must
111
A B
a2 a1 b1 b2 b3 b4
a4 a3
LV
HV
112
not be open circuited. If it gets open circuited the core (B in fig) gets highly saturated.
In spite of the small ratings and low voltages and flexibility, this method of voltage
control costs more mainly due to the additional floor space it needs. The methods of
voltage regulation discussed so far basically use the principle of tap changing and hence
the voltage change takes place in steps. Applications like a.c. and d.c. motor speed
control, illumination control by dimmers, electro-chemistry and voltage stabilizers need
continuous control of voltage. This can be obtained with the help of moving coil voltage
regulators.
a2
5% a1
V out
V in S
95% b1
b2
moving coil voltage regulators Fig. 45 shows the physical arrangement of one such trans-
former. a, b are the two primary windings wound on a long core, wound in the opposite
113
sense. Thus the flux produced by each winding takes a path through the air to link
the winding. These fluxes link their secondaries a2 and b2 . A short circuited moving
coil s is wound on the same limb and is capable of being held at any desired position.
This moving coil alters the inductances of the two primaries. The sharing of the total
applied voltage thus becomes different and also the induced emf in the secondaries a2
and b2 . The total secondary voltage in the present case varies from 10 percent to 20
percent of the input in a continuous manner. The turns ratios of a1 : a2 and b1 : b2
5 95
are 4.86 and 10.6 respectively. 4.86
+ 10.6
= 10% when s is in the top position. In
95 5
the bottom position it becomes 4.86
+ 10.6
= 20%. By selecting proper ratios for the
secondaries a2 and b2 one can get the desired voltage variation.
Variable secondary
a.c voltage
V1 V1
V2
sliding contact
Sliding contact regulators These have two winding or auto transformer like construction.
The winding from which the output is taken is bared and a sliding contact taps the
voltage. The minimum step size of voltage change obtainable is the voltage across a
single turn. The conductor is chosen on the basis of the maximum load current on the
114
output side. In smaller ratings this is highly cost effective. Two winding arrangements
are also possible. The two winding arrangement provides electrical isolation also.These
are shown in Fig. 46.
115
D.C Machines
1 Introduction
The steam age signalled the beginning of an industrial revolution. The advantages
of machines and gadgets in helping mass production and in improving the services spurred
the industrial research. Thus a search for new sources of energy and novel gadgets received
great attention. By the end of the 18th century the research on electric charges received
a great boost with the invention of storage batteries. This enabled the research work on
moving charges or currents. It was soon discovered ( in 1820 ) that, these electric currents
are also associated with magnetic field like a load stone. This led to the invention of an
electromagnet. Hardly a year later the force exerted on a current carrying conductor placed
in the magnetic field was invented. This can be termed as the birth of a motor. A better
understanding of the inter relationship between electric and magnetic circuits was obtained
with the enumeration of laws of induction by Faraday in 1831. Parallel research was contem-
porarily being done to invent a source of energy to recharge the batteries in the form of a
d.c. source of constant amplitude (or d.c. generator). For about three decades the research
on d.c. motors and d.c. generators proceeded on independent paths. During the second half
of the 19th century these two paths merged. The invention of a commutator paved the way
for the birth of d.c. generators and motors. These inventions generated great interest in the
generation and use of electrical energy. Other useful machines like alternators, transformers
and induction motors came into existence almost contemporarily. The evolution of these
machines was very quick. They rapidly attained the physical configurations that are being
used even today. The d.c. power system was poised for a predominant place as a preferred
system for use, with the availability of batteries for storage, d.c. generators for conversion of
mechanical energy into electrical form and d.c. motors for getting mechanical outputs from
electrical energy.
The limitations of the d.c. system however became more and more apparent
as the power demand increased. In the case of d.c. systems the generating stations and the
load centers have to be near to each other for efficient transmission of energy. The invention
of induction machines in the 1880s tilted the scale in favor of a.c. systems mainly due to
the advantage offered by transformers, which could step up or step down the a.c.voltage
levels at constant power at extremely high efficiency. Thus a.c. system took over as the
preferred system for the generation transmission and utilization of electrical energy. The
d.c. system, however could not be obliterated due to the able support of batteries. Further,
d.c. motors have excellent control characteristics. Even today the d.c. motor remains an
industry standard as far as the control aspects are concerned. In the lower power levels and
also in regenerative systems the d.c. machines still have a major say.
Electric machines can be broadly classified into electrostatic machines and electro-
magnetic machines. The electrostatic principles do not yield practical machines for commer-
cial electric power generation. The present day machines are based on the electro-magnetic
principles. Though one sees a variety of electrical machines in the market, the basic under-
lying principles of all these are the same. To understand, design and use these machines the
following laws must be studied.
Most of the present day machines have one or two electric circuits linking a common
magnetic circuit. In subsequent discussions the knowledge of electric and magnetic circuit
laws is assumed. The attention is focused on the Faraday’s law and Biot Savart’s law in the
present study of the electrical machines.
Faraday proposed this law of Induction in 1831. It states that if the magnetic
flux lines linking a closed electric coil changes, then an emf is induced in the coil. This
emf is proportional to the rate of change of these flux linkages. This can be expressed
mathematically,
dψ
e∝ (1)
dt
3
where ψ is the flux linkages given by the product of flux lines in weber that are linked
and N the number of turns of the coil. This can be expressed as,
dΦ
e∝N (2)
dt
Here N is the number of turns of the coil, and Φ is the flux lines in weber link-
ing all these turns. The direction of the induced emf can be determined by the application of
Lenz’s law. Lenz’s law states that the direction of the induced emf is such as to produce an
effect to oppose this change in flux linkages. It is analogous to the inertia in the mechanical
systems.
The changes in the flux linkages associated with a turn can be brought about by
(i) changing the magnitude of the flux linking a static coil
(ii) moving the turn outside the region of a steady field
(iii) moving the turn and changing the flux simultaneously
These may be termed as Case(i), Case(ii), and Case(iii) respectively.
This is now explained with the help of a simple geometry. Fig. 1 shows a rectan-
gular loop of one turn (or N=1). Conductor 1 is placed over a region with a uniform flux
density of B Tesla. The flux lines, the conductor and the motion are in mutually perpendic-
ular directions. The flux linkages of the loop is BLN weber turns. If the flux is unchanging
and conductor stationary, no emf will be seen at the terminals of the loop. If now the flux
alone changes with time such that B = Bm . cos ωt, as in Case(i), an emf given by
d
e= (Bm .L.N cos ωt) = −(Bm .L.Nω). sin ωt.
dt
= −jBm .L.Nω. cos ωt volt (3)
L
B
- +
dψ d(Bm .L.N) dX
e= = = Bm .L.N volts (4)
dt dt dt
but
dX
=v ∴ e = Bm .L.N.v volts (5)
dt
The emf induced in the loop is directly proportional to the uniform flux density under which
it is moving with a velocity v. This type of voltage is called speed emf (or rotational emf).
The Case(iii) refers to the situation where B is changing with time and so also is X. Then
the change in flux linkage and hence the value of e is given by
In this case both transformer emf and speed emf are present.
The Case(i) has no mechanical energy associated with it. This is the principle
used in transformers. One coil carrying time varying current produces the time varying field
and a second coil kept in the vicinity of the same has an emf induced in it. The induced emf
of this variety is often termed as the transformer emf.
The Case(ii) is the one which is employed in d.c. machines and alternators. A
static magnetic field is produced by a permanent magnet or by a coil carrying a d.c. current.
A coil is moved under this field to produce the change in the flux linkages and induce an emf
in the same. In order to produce the emf on a continuous manner a cylindrical geometry
is chosen for the machines. The direction of the field, the direction of the conductor of the
coil and the direction of movement are mutually perpendicular as mentioned above in the
example taken.
In the example shown above, only one conductor is taken and the flux ’cut’ by
the same in the normal direction is used for the computation of the emf. The second con-
ductor of the turn may be assumed to be far away or unmoving. This greatly simplifies the
computation of the induced voltage as the determination of flux linkages and finding its rate
of change are dispensed with. For a conductor moving at a constant velocity v the induced
emf becomes just proportional to the uniform flux density of the magnetic field where the
conductor is situated. If the conductor, field and motion are not normal to each other then
the mutually normal components are to be taken for the computation of the voltage. The
induced emf of this type is usually referred to as a rotational emf (due to the geometry).
cal energy conversion results in the generation of both transformer and rotational emf to be
present in the coil moving under a changing field. This principle is utilized in the induction
machines and a.c. commutator machines. The direction of the induced emf is
emf and
current
Force Motion F
B
(a) (b)
decided next. This can be obtained by the application of the Lenz’s law and the law of
interaction. This is illustrated in Fig. 3.
In Case(ii), the direction of the induced emf is as shown. Here again one could
derive the same from the application of the Lenz’s law. The changes in the flux linkages is
emf
current
F
Motion,Force
B
(a) (b)
brought about by the sweep or movement of the conductor. The induced emf, if permitted
to drive a current which produces an opposing force, is as shown in the figure. If one looks
closely at the field around the conductor under these conditions it is as shown in Fig. 2(a)and
(b). The flux lines are more on one side of the conductor than the other. These lines seem
to urge the conductor to the left with a force F . As F opposes v and the applied force,
mechanical energy gets absorbed in this case and the machine works as a generator. This
force is due to electro magnetic interaction and is proportional to the current and the flux
swept. Fig. 3(a)and (b) similarly explain the d.c.motor operation. The current carrying con-
ductor reacts with the field to develop a force which urges the conductor to the right. The
induced emf and the current are seen to act in opposite direction resulting in the absorption
of electric energy which gets converted into the mechanical form.
In Case (iii) also the direction of the induced emf can be determined in a
similar manner. However, it is going to be more complex due to the presence of transformer
emf and rotational emf which have phase difference between them.
Putting mathematically, in the present study of d.c.machines,
F = B.L.I Newton
When the generated voltage drives a current, it produces a reaction force on the
mechanical system which absorbs the mechanical energy. This absorbed mechanical energy
is the one which results in the electric current and the appearance of electrical energy in
the electrical circuit. The converse happens in the case of the motor. If we force a current
against an induced emf then the electrical power is absorbed by the same and it appears
as the mechanical torque on the shaft. Thus, it is seen that the motoring and generating
actions are easily changeable with the help of the terminal conditions.
D.C. machines are the electro mechanical energy converters which work from a d.c.
source and generate mechanical power or convert mechanical power into a d.c. power. These
machines can be broadly classified into two types, on the basis of their magnetic structure.
They are,
1. Homopolar machines
2. Heteropolar machines.
Homopolar generators
Even though the magnetic poles occur in pairs, in a homopolar generator the conductors
are arranged in such a manner that they always move under one polarity. Either north pole
or south pole could be used for this purpose. Since the conductor encounters the magnetic
flux of the same polarity every where it is called a homopolar generator. A cylindrically
symmetric geometry is chosen. The conductor can be situated on the surface of the rotor
with one slip-ring at each end of the conductor. A simple structure where there is only
one cylindrical conductor with ring brushes situated at the ends is shown in Fig. 4. The
excitation coil produces a field which enters the inner member from outside all along the
periphery. The conductor thus sees only one pole polarity or the flux directed in one sense.
A steady voltage now appears across the brushes at any given speed of rotation. The polarity
of the induced voltage can be reversed by reversing either the excitation or the direction of
10
+ N - Brush
A B
Flux
S
Field A B
+ N -
coil
11
rotation but not both. The voltage induced would be very low but the currents of very large
amplitudes can be supplied by such machines. Such sources are used in some applications
like pulse-current and MHD generators, liquid metal pumps or plasma rockets. The steady
field can also be produced using a permanent magnet of ring shape which is radially mag-
netized. If higher voltages are required one is forced to connect many conductors in series.
This series connection has to be done externally. Many conductors must be situated on the
rotating structure each connected to a pair of slip rings. However, this modification intro-
duces parasitic air-gaps and makes the mechanical structure very complex. The magnitude
of the induced emf in a conductor 10 cm long kept on a rotor of 10 cm radius rotating at
3000 rpm, with the field flux density being 1 Tesla every where in the air gap, is given by
e = BLv
3000
= 1 ∗ 0.1 ∗ 2π ∗ 0.1 ∗ = 3.14 volt
60
The voltage drops at the brushes become very significant at this level bringing down the
efficiency of power conversion. Even though homopolar machines are d.c. generators in a
strict sense that they ’generate’ steady voltages, they are not quite useful for day to day use.
A more practical converters can be found in the d.c. machine family called ”hetero-polar”
machines.
In the case of a hetero-polar generator the induced emf in a conductor goes through a
cyclic change in voltage as it passes under north and south pole polarity alternately. The
induced emf in the conductor therefore is not a constant but alternates in magnitude. For
12
c
B -
d
+
A
Load
Field coil
Pole
v
N
v
10
11 9
12 8
Commutator S1 7
A+ -
v
B
1 F1
S2
Armature core
v
F2
S3
6
F4
Yoke
F3
2
S4
v
3 5
4
13
a constant velocity of sweep the induced emf is directly proportional to the flux density
under which it is moving. If the flux density variation is sinusoidal in space, then a sine
wave voltage is generated. This principle is used in the a.c generators. In the case of d.c.
generators our aim is to get a steady d.c. voltage at the terminals of the winding and not
the shape of the emf in the conductors. This is achieved by employing an external element,
which is called a commutator, with the winding.
14
Fig. 6 shows a 2-pole, 12 coil, ring wound armature of a machine. The 12 coils
are placed at uniform spacing around the rotor. The junction of each coil with its neighbor
is connected to a commutator segment. Each commutator segment is insulated from its
neighbor by a mica separator. Two brushes A and B are placed on the commutator which
looks like a cylinder. If one traces the connection from brush A to brush B one finds that
there are two paths. In each path a set of voltages get added up. The sum of the emfs is
constant(nearly). The constancy of this magnitude is altered by a small value corresponding
to the coil short circuited by the brush. As we wish to have a maximum value for the output
voltage, the choice of position for the brushes would be at the neutral axis of the field. If
the armature is turned by a distance of one slot pitch the sum of emfs is seen to be constant
even though a different set of coils participate in the addition. The coil which gets short
circuited has nearly zero voltage induced in the same and hence the sum does not change
substantially. This variation in the output voltage is called the ’ripple’. More the number of
coils participating in the sum lesser would be the ’percentage’ ripple.
A rectangular flux density wave form has some advantages in the derivation
of the voltage between the brushes. Due to this form of the flux density curve, the induced
15
emf in each turn of the armature becomes constant and equal to each other. With this back
ground the emf induced between the brushes can be derived. The value of the induced in
one conductor is given by
Ec = Bav .L.v Volt (7)
where
Bav - Average flux density over a pole pitch, Tesla.
L- Length of the ’active’ conductor, m.
v- Velocity of sweep of conductor, m/sec.
If there are Z conductors on the armature and they form b pairs of parallel circuits between
the brushes by virtue of their connections, then number of conductors in a series path is
Z/2b.
The induced emf between the brushes is
Z
E = Ec . (8)
2b
Z
E = Bav .L.v. Volts (9)
2b
But v = (2p).Y.n where p is the pairs of poles Y is the pole pitch, in meters, and n is the
number of revolutions made by the armature per second.
Also Bav can be written in terms of pole pitch Y , core length L, and flux per pole φ as
φ
Bav = Tesla (10)
(L.Y )
φ Z φpZn
E= .L.(2p.Y.n). = volts (11)
(L.Y ) 2b b
The number of pairs of parallel paths is a function of the type of the winding chosen. This
16
When the armature is loaded, the armature conductors carry currents. These current
carrying conductors interact with the field and experience force acting on the same. This
force is in such a direction as to oppose their cause which in the present case is the relative
movement between the conductors and the field. Thus the force directly opposes the motion.
Hence it absorbs mechanical energy. This absorbed mechanical power manifests itself as the
converted electrical power. The electrical power generated by an armature delivering a
current of Ia to the load at an induced emf of E is EIa Watts. Equating the mechanical and
electrical power we have
2πnT = EIa (12)
where T is the torque in Nm. Substituting for E from Eqn. 11, we get
p.φ.Z.n
2πnT = .Ia (13)
b
This shows that the torque generated is not a function of the speed. Also,
it is proportional to ’total flux’ and ’Total ampere conductors’ on the armature, knowing that
Ia /2b is Ic the conductor current on the armature. The expression for the torque generated
can also be derived from the first principles by the application of the law of interaction. The
law of interaction states that the force experienced by a conductor of length L kept in a
17
The total work done by an armature with Z conductors in one revolution is given by,
φ
Wa = Bav .L.Ic .Z.(2p.Y ) Joules = .L.Ic .Z.2p.Y Joules (16)
L.Y
The work done per second or the power converted by the armature is,
Ia
AsIc = (18)
2b
Ia
= φ.p.Z.n. (19)
b
which is nothing but EIa .
The above principles can easily be extended to the case of motoring mode
of operation also. This will be discussed next in the section on motoring operation of d.c.
machines.
In the motoring operation the d.c. machine is made to work from a d.c. source and
absorb electrical power. This power is converted into the mechanical form. This is briefly
discussed here. If the armature of the d.c. machine which is at rest is connected to a d.c.
source then, a current flows into the armature conductors. If the field is already excited then
18
these current carrying conductors experience a force as per the law of interaction discussed
above and the armature experiences a torque. If the restraining torque could be neglected the
armature starts rotating in the direction of the force. The conductors now move under the
field and cut the magnetic flux and hence an induced emf appears in them. The polarity of
the induced emf is such as to oppose the cause of the current which in the present case is the
applied voltage. Thus a ’back emf’ appears and tries to reduce the current. As the induced
emf and the current act in opposing sense the machine acts like a sink to the electrical power
which the source supplies. This absorbed electrical power gets converted into mechanical
form. Thus the same electrical machine works as a generator of electrical power or the
absorber of electrical power depending upon the operating condition. The absorbed power
gets converted into electrical or mechanical power. This is briefly explained earlier with the
help of Figure 3(a) and 3(b). These aspects would be discussed in detail at a later stage.
19
As mentioned earlier the d.c. machines were invented during the second half of the 19th
century. The initial pace of development work was phenomenal. The best configurations
stood all the competition and the test of time and were adopted. Less effective options were
discarded. The present day d.c. generator contains most, if not all, of the features of the
machine developed over a century earlier. To appreciate the working and the characteristics
of these machines, it is necessary to know about the different parts of the machine - both
electrical and non-electrical. The description would also aid the understanding of the reason
for selecting one form of construction or the other. An exploded view of a small d.c.
1. Body
2. Poles
20
3. Armature
5. Commutating poles
6. Compensating winding
The constructional aspects relating to these parts are now discussed briefly in sequence.
Body The body constitutes the outer shell within which all the other parts are housed.
This will be closed at both the ends by two end covers which also support the bearings
required to facilitate the rotation of the rotor and the shaft. Even though for the
generation of an emf in a conductor a relative movement between the field and the
conductor would be enough, due to practical considerations of commutation, a rotating
conductor configuration is selected for d.c. machines. Hence the shell or frame supports
the poles and yoke of the magnetic system. In many cases the shell forms part of the
magnetic circuit itself. Cast steel is used as a material for the frame and yoke as the
flux does not vary in these parts. In large machines these are fabricated by suitably
welding the different parts. Those are called as fabricated frames. Fabrication as
against casting avoids expensive patterns. In small special machines these could be
made of stack of laminations suitably fastened together to form a solid structure.
Main poles Solid poles of fabricated steel with seperate/integral pole shoes are fastened
to the frame by means of bolts. Pole shoes are generally laminated. Sometimes pole
body and pole shoe are formed from the same laminations. Stiffeners are used on both
21
sides of the laminations. Riveted through bolts hold the assembly together. The pole
shoes are shaped so as to have a slightly increased air gap at the tips.
Inter-poles These are small additional poles located in between the main poles. These can
be solid, or laminated just as the main poles. These are also fastened to the yoke by
bolts. Sometimes the yoke may be slotted to receive these poles. The inter poles could
be of tapered section or of uniform cross section. These are also called as commutating
poles or compoles. The width of the tip of the compole can be about a rotor slot pitch.
Armature The armature is where the moving conductors are located. The armature is
constructed by stacking laminated sheets of silicon steel. Thickness of these lamination
is kept low to reduce eddy current losses. As the laminations carry alternating flux
the choice of suitable material, insulation coating on the laminations, stacking it etc
are to be done more carefully. The core is divided into packets to facilitate ventilation.
The winding cannot be placed on the surface of the rotor due to the mechanical forces
coming on the same. Open parallel sided equally spaced slots are normally punched in
the rotor laminations. These slots house the armature winding. Large sized machines
employ a spider on which the laminations are stacked in segments. End plates are
suitably shaped so as to serve as ’Winding supporters’. Armature construction process
must ensure provision of sufficient axial and radial ducts to facilitate easy removal of
heat from the armature winding.
Field windings In the case of wound field machines (as against permanent magnet excited
machines) the field winding takes the form of a concentric coil wound around the main
poles. These carry the excitation current and produce the main field in the machine.
Thus the poles are created electromagnetically. Two types of windings are generally
employed. In shunt winding large number of turns of small section copper conductor is
22
used. The resistance of such winding would be an order of magnitude larger than the
armature winding resistance. In the case of series winding a few turns of heavy cross
section conductor is used. The resistance of such windings is low and is comparable
to armature resistance. Some machines may have both the windings on the poles.
The total ampere turns required to establish the necessary flux under the poles is
calculated from the magnetic circuit calculations. The total mmf required is divided
equally between north and south poles as the poles are produced in pairs. The mmf
required to be shared between shunt and series windings are apportioned as per the
design requirements. As these work on the same magnetic system they are in the form
of concentric coils. Mmf ’per pole’ is normally used in these calculations.
Armature winding As mentioned earlier, if the armature coils are wound on the surface of
the armature, such construction becomes mechanically weak. The conductors may fly
away when the armature starts rotating. Hence the armature windings are in general
pre-formed, taped and lowered into the open slots on the armature. In the case of
small machines, they can be hand wound. The coils are prevented from flying out due
to the centrifugal forces by means of bands of steel wire on the surface of the rotor in
small groves cut into it. In the case of large machines slot wedges are additionally used
to restrain the coils from flying away. The end portion of the windings are taped at
the free end and bound to the winding carrier ring of the armature at the commutator
end. The armature must be dynamically balanced to reduce the centrifugal forces at
the operating speeds.
Compensating winding One may find a bar winding housed in the slots on the pole
shoes. This is mostly found in d.c. machines of very large rating. Such winding is
called compensating winding. In smaller machines, they may be absent. The function
23
1.Clamping cone
4 2.Insulating cups
3.Commutator bar
3 4.Riser
2
2 5.Insulating gasket
1
5
Commutator Commutator is the key element which made the d.c. machine of the present
day possible. It consists of copper segments tightly fastened together with mica/micanite
insulating separators on an insulated base. The whole commutator forms a rigid and
solid assembly of insulated copper strips and can rotate at high speeds. Each com-
mutator segment is provided with a ’riser’ where the ends of the armature coils get
connected. The surface of the commutator is machined and surface is made concentric
with the shaft and the current collecting brushes rest on the same. Under-cutting the
mica insulators that are between these commutator segments has to be done periodi-
cally to avoid fouling of the surface of the commutator by mica when the commutator
gets worn out. Some details of the construction of the commutator are seen in Fig. 8.
Brush and brush holders Brushes rest on the surface of the commutator. Normally
electro-graphite is used as brush material. The actual composition of the brush depends
on the peripheral speed of the commutator and the working voltage. The hardness of
the graphite brush is selected to be lower than that of the commutator. When the
24
brush wears out the graphite works as a solid lubricant reducing frictional coefficient.
More number of relatively smaller width brushes are preferred in place of large broad
brushes. The brush holders provide slots for the brushes to be placed. The connection
Pigtail
Pressure
spring
Brush
(a)
Radial Trailing
Reaction
Motion of commutator
(b)
Figure 9: Brush holder with a Brush and Positioning of the brush on the commutator
from the brush is taken out by means of flexible pigtail. The brushes are kept pressed
on the commutator with the help of springs. This is to ensure proper contact between
25
the brushes and the commutator even under high speeds of operation. Jumping of
brushes must be avoided to ensure arc free current collection and to keep the brush
contact drop low. Fig. 9 shows a brush holder arrangement. Radial positioning of the
brushes helps in providing similar current collection conditions for both direction of
rotation. For unidirectional drives trailing brush arrangement or reaction arrangement
may be used in Fig. 9-(b) Reaction arrangement is preferred as it results in zero side
thrust on brush box and the brush can slide down or up freely. Also staggering of the
brushes along the length of the commutator is adopted to avoid formation of ’tracks’
on the commutator. This is especially true if the machine is operating in a dusty
environment like the one found in cement plants.
Other mechanical parts End covers, fan and shaft bearings form other important me-
chanical parts. End covers are completely solid or have opening for ventilation. They
support the bearings which are on the shaft. Proper machining is to be ensured for
easy assembly. Fans can be external or internal. In most machines the fan is on the
non-commutator end sucking the air from the commutator end and throwing the same
out. Adequate quantity of hot air removal has to be ensured.
Bearings Small machines employ ball bearings at both ends. For larger machines roller
bearings are used especially at the driving end. The bearings are mounted press-fit
on the shaft. They are housed inside the end shield in such a manner that it is not
necessary to remove the bearings from the shaft for dismantling. The bearings must be
kept in closed housing with suitable lubricant keeping dust and other foreign materials
away. Thrust bearings, roller bearings, pedestal bearings etc are used under special
cases. Care must be taken to see that there are no bearing currents or axial forces on
the shaft both of which destroy the bearings.
26
4 Armature Windings
Main field
N
Commutator Compole field
& Brush X
X
x
x x
x
x
X
x x
Shaft Compensating
x x S
S winding
x x
v
x x
X
x
Armature
x
winding x
x
X X
Yoke
N X
Fig. 10 gives the cross sectional view of a modern d.c. machine show-
ing all the salient parts. Armature windings, along with the commutators, form the heart
of the d.c. machine. This is where the emf is induced and hence its effective deployment
enhances the output of the machine. Fig. 11(a) shows one coil of an armature of Gramme
ring arrangement and Fig. 11(b) shows one coil as per drum winding arrangement. Earlier,
a simple form of this winding in the form of Gramme ring winding was presented for easy
understanding. The Gramme ring winding is now obsolete as a better armature winding has
27
X Ν X Ν
X X
φ φ
Α
Α Α’
28
been invented in the form of a drum winding. The ring winding has only one conductor in
a turn working as an active conductor. The second conductor is used simply to complete
the electrical connections. Thus the effectiveness of the electric circuit is only 50 percent.
Looking at it differently, half of the magnetic flux per pole links with each coil. Also, the
return conductor has to be wound inside the bore of the rotor, and hence the rotor diameter
is larger and mounting of the rotor on the shaft is made difficult.
In a drum winding both forward and return conductors are housed in slots cut
on the armature (or drum). Both the conductors have emf induced in them. Looking at it
differently the total flux of a pole is linked with a turn inducing much larger voltage induced
in the same. The rotor is mechanically robust with more area being available for carrying
the flux. There is no necessity for a rotor bore. The rotor diameters are smaller. Mechanical
problems that existed in ring winding are no longer there with drum windings. The coils
could be made of single conductors (single turn coils) or more number of conductors in series
(multi turn coils). These coils are in turn connected to form a closed winding. The two sides
of the coil lie under two poles one north and the other south, so that the induced emf in
them are always additive by virtue of the end connection. Even though the total winding
is a closed one the sum of the emfs would be zero at all times. Thus there is no circulating
current when the armature is not loaded. The two sides of the coil, if left on the surface, will
fly away due to centrifugal forces. Hence slots are made on the surface and the conductors
are placed in these slots and fastened by steel wires to keep them in position. Each armature
slot is partitioned into two layers, a top layer and a bottom layer. The winding is called as
a double layer winding. This is a direct consequence of the symmetry consideration. The
distance, measured along the periphery of the armature from any point under a pole to a
similar point under the neighboring pole is termed as a pole pitch. The forward conductor
is housed in the top layer of a slot and the return conductor is housed in the bottom layer
29
B
S A
N
C
D
Lower
coil side
A’
N
Active
S A N A’ S
Inactive
Armature
C D
30
of a slot which is displaced by about one pole pitch. The junction of two coils is terminated
on a commutator segment. Thus there are as many commutator segments as the number of
coils. In a double layer winding in S slots there are 2S layers. Two layers are occupied by a
coil and hence totally there are S coils. The S junctions of these S coils are terminated on S
commutator segments. The brushes are placed in such a manner that a maximum voltage
appears across them. While the number of parallel circuits in the case of ring winding is
equal to the number of poles, in the case of drum winding a wide variety of windings are
possible. The number of brushes and parallel paths thus vary considerably. The physical
arrangement of a single coil is shown in Fig. 12 to illustrate its location and connection to
the commutators.
Fig. 13 shows the axial side view while Fig. 13-(b) shows the cut and spread view
of the machine. The number of turns in a coil can be one (single turn coils) or more (multi
turn coils ). As seen earlier the sum of the instantaneous emfs appears across the brushes.
This sum gets altered by the voltage of a coil that is being switched from one circuit to the
other or which is being commutated. As this coil in general lies in the magnetic neutral
axis it has a small value of voltage induced in it. This change in the sum expressed as the
fraction of the total induced voltage is called as the ripple. In order to reduce the ripple,
one can increase the number of coils coming in series between the brushes. As the number
of coils is the same as the number of slots in an armature with two coil sides per slot one is
forced to increase the number of slots. However increasing the slot number makes the tooth
width too narrow and makes them mechanically weak.
To solve this problem the slots are partitioned vertically to increase the number
of coil sides. This is shown in Fig. 14. In the figure, the conductors a, b and c belong to a
coil. Such 2/3 coils occupy the 2/3 top coil sides of the slot. In the present case the number
of coils in the armature is 2S/3S.
31
(a)End view
11 2’
10
3’
1’
12
1 2 3 4 5 6 7 8 9 10 11 12
N S N S
12
1’
2’
11
12 1 2 3 4 5 6 7 8 9 10 11
+ - + Motion
-
(b)Developed view
Press board
Copper
Mica Tape
Press board
33
As mentioned earlier, in a drum winding, the coils span a pole pitch where
ever possible. Such coils are called ’full pitched’ coils. The emf induced in the two active
conductors of such coils have identical emfs with opposite signs at all instants of time. If the
span is more than or less than the full pitch then the coil is said to be ’chorded’. In chorded
coils the induced emfs of the two conductor may be of the same sign and hence oppose each
other( for brief intervals of time). Slight short chording of the coil reduces overhang length
and saves copper and also improves commutation. Hence when the pole pitch becomes frac-
tional number, the smaller whole number may be selected discarding the fractional part.
Similar to the pitch of a coil one can define the winding pitch and commutator
pitch. In a d.c. winding the end of one coil is connected to the beginning of another coil
(not necessarily the next), this being symmetrically followed to include all the coils on the
armature. Winding pitch provides a means of indicating this. Similarly the commutator
pitch provides the information regarding the commutators to which the beginning and the
end of a coil are connected. Commutator pitch is the number of ’micas’ between the ends of
a coil. For all these information to be simple and useful the numbering scheme of the coils
and commutator segments becomes important. One simple method is to number only the
top coil side of the coils in sequence. The return conductor need not be numbered. As a
double layer is being used the bottom coil side is placed in a slot displaced by one coil span
′ ′
from the top coil side. Some times the coils are numbered as 1 − 1 , 2 − 2 etc. indicating
′ ′
the second sides by 1 , 2 etc. The numbering of commutators segments are done similarly.
The commutator segment connected to top coil side of coil 1 is numbered 1. This method
of numbering is simple and easy to follow. It should be noted that changing of the pitch
34
of a coil slightly changes the induced emf in the same. The pitch of the winding however
substantially alters the nature of the winding.
The armature windings are classified into two families based on this. They are
called lap winding and wave winding. They can be simply stated in terms of the commutator
pitch used for the winding.
35
Progressive Retrogressive
yc =+1 yc = -1
s1 s2 F1 F2 F2
s2 F3 s3
1 2 3 1 2 3 4
Coil span
s1 F1
1 2 _1
c+
p
36
1 2
S
N S N S
13 14 1 2 3 4 5 6 7 8 9 10 11 12 Motion
A1 + B1 - A2 + B2 -
37
equally.
The increase in the number of parallel paths in the armature winding brings
about a problem of circulating current. The induced emfs in the different paths tend to
differ slightly due to the non-uniformities in the magnetic circuit. This will be more with the
increase in the number of poles in the machine. If this is left uncorrected, circulating currents
appear in these closed parallel paths. This circulating current wastes power, produces heat
and over loads the brushes under loaded conditions. One method commonly adopted in d.c.
machines to reduce this problem is to provide equalizer connections. As the name suggests
these connections identify similar potential points of the different parallel paths and connect
them together to equalize the potentials. Any difference in the potential generates a local
circulating current and the voltages get equalized. Also, the circulating current does not
flow through the brushes loading them. The number of such equalizer connections, the
cross section for the conductor used for the equalizer etc are decided by the designer. An
example of equalizer connection is discussed now with the help of a 6-pole armature having
150 commutator segments. The coil numbers 1, 51 and 101 are identically placed under the
poles of same polarity as they are one pole-pair apart. There are 50 groups like that. In
order to limit the number of links to 5(say), the following connections are chosen. Then
1,11,21,31, and 41 are the coils under the first pair of poles. These are connected to their
counter parts displaced by 50 and 100 to yield 5 equalizer connections. There are 10 coils
connected in series between any two successive links. The wave windings shall be examined
next.
38
S N S N
20 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Motion
+ A1 - B1 + A2 - B2
(a)Winding layout
_ 5
Full pitch: 21/4=5.25 ~
Span : 1 to 6
Yc=
_1
C+
2
= 21
_1
+
2
=
11
10 }
Commutator pitch 1-11 for retrogressive winding
1-11-21-10-20-9-19-8-18-7-17-
v
A2
A1
6-16-5-15-4-14-3-13-2-12-1
v
B2
B1
(b)Parallel paths
39
In wave windings the coils carrying emf in the same direction at a time are all
grouped together and connected in series. Hence in a simple wave winding there are only
two paths between the brushes, the number of conductors in each path being 50 percent of
the total conductors. To implement a wave winding one should select the commutator pitch
as
C±1
yc = (21)
p
where C is the total segments on the commutator. yc should be an integer number; C and
p should satisfy this relation correctly. Here also the positive sign refers to the progressive
winding and the negative sign yields a retrogressive winding. yc = (C ± m)/p yields a multi-
plex wave winding of order m. A simple wave winding for 4 poles in 21 slots is illustrated in
Fig. 17. As could be seen from the figure, the connection to the next (or previous) adjacent
coil is reached after p coils are connected in series. The winding closes on itself after all the
coils are connected in series. The position for the brushes is indicated in the diagram.
It is seen from the formula for the commutator pitch, the choice of commutator
segments for wave winding is restricted. The number of commutator segments can only be
one more or one less than some multiple of pole pairs. As the number of parallel circuits is
2 for a simple wave winding irrespective of the pole numbers it is preferred in multi polar
machine of lower power levels.
As mentioned earlier the simple wave winding forms two parallel paths, duplex
wave winding has 2*2=4 etc. The coils under all the north poles are grouped together in
40
one circuit and the other circuit collects all the coils that are under all the south poles. Two
brush sets are therefore adequate. Occasionally people employ brush sets equal to the num-
ber of poles. This arrangement does not increase the number of parallel circuits but reduces
the current to be collected by each brush set. This can be illustrated by an example. A
4-pole wave connected winding with 21 commutator segments is taken. yc = (21 − 1)/2 = 10
. A retrogressive wave winding results. The total string of connection can be laid out as
shown below. If coil number 1 is assumed to be in the neutral axis then other neutral axis
coils are a pole pitch apart i.e. coils 6, 11, 16.
If the brushes are kept at commutator segment 1 and 6, nearly half the num-
ber of coils come under each circuit. The polarity of the brushes are positive and negative
alternately. Or, one could have two brushes at 11 and 16 or any two adjacent poles. By
having four brushes at 1, 6, 11 and 16 and connecting 1,11 and 6,16 still only two parallel
circuits are obtained. The brush currents however are halved. This method permits the use
of commutator of shorter length as lesser current is to be collected by each brush and thus
saving on the cost of the commutator. Fig. 17(b) illustrates this brush arrangement with
respect to a 21 slot 4 pole machine. Similarly proceeding, in a 6-pole winding 2,4 or 6 brush
sets may be used.
41
Due to the restrictions posed by lap and wave windings on the choice of number
of slots and commutator segments a practical difficulty arises. Each machine with a certain
pole number, voltage and power ratings may require a particular number of slots and com-
mutator segments for a proper design. Thus each machine may be tailor made for a given
specification. This will require stocking and handling many sizes of armature and commu-
tator.
42
5 Armature reaction
Earlier, an expression was derived for the induced emf at the terminals of the
armature winding under the influence of motion of the conductors under the field established
by field poles. But if the generator is to be of some use it should deliver electrical output to a
load. In such a case the armature conductors also carry currents and produce a field of their
own. The interaction between the fields must therefore must be properly understood in order
to understand the behavior of the loaded machine. As the magnetic structure is complex
and as we are interested in the flux cut by the conductors, we primarily focus our attention
on the surface of the armature. A sign convention is required for mmf as the armature and
field mmf are on two different members of the machine. The convention used here is that
the mmf acting across the air gap and the flux density in the air gap are shown as positive
when they act in a direction from the field system to the armature. A flux line is taken
and the value of the current enclosed is determined. As the magnetic circuit is non-linear,
the field mmf and armature mmf are separately computed and added at each point on the
surface of the armature. The actual flux produced is proportional to the total mmf and the
permeance. The flux produced by field and that produced by armature could be added to
get the total flux only in the case of a linear magnetic circuit. The mmf distribution due to
the poles and armature are discussed now in sequence.
Fig. 18 shows the distribution of mmf due to field coils over two pole pitches. It
is a step curve with the width being equal to the pole arc. The permeance variation at the
surface is given by Fig. 18 assuming the air gap under the pole to be uniform and neglecting
43
N S
mmf
Permeance
Practical
Flux density
Ideal flux density
44
the slotting of the armature. The no-load flux density curve can be obtained by multiplying
mmf and permeance. Allowing for the fringing of the flux, the actual flux density curve
would be as shown under Fig. 18.
N S
N-Pole
S-Pole
Generator Flux
mmf
Figure 19: Mmf and flux distribution under the action of armature alone carrying current
The armature has a distributed winding, as against the field coils which
are concentrated and concentric. The mmf of each coil is shifted in space by the number of
slots. For a full pitched coil, each coil produces a rectangular mmf distribution. The sum
of the mmf due to all coils would result in a stepped triangular wave form. If we neglect
slotting and have uniformly spaced coils on the surface, then the mmf distribution due to the
armature working alone would be a triangular distribution in space since all the conductors
carry equal currents. MMF distribution is the integral of the ampere conductor distribution.
45
This is depicted in Fig. 19. This armature mmf per pole is given by
1 Ic .Z
Fa = .
2 2p
where Ic is the conductor current and Z is total number of conductors on the armature. This
peak value of the mmf occurs at the inter polar area, shifted from the main pole axis by half
the pole pitch when the brushes are kept in the magnetic neutral axis of the main poles.
Brush axis
N S
A Generator B
c
D
C
Field a
flux
B
o’
A o
b
Armature flux
Total flux
The mmf of field coils and armature coils are added up and the re-
sultant mmf distribution is obtained as shown in Fig. 20.
46
This shows the decrease in the mmf at one tip of a pole and a substantial rise
at the other tip. If the machine has a pole arc to pole pitch ratio of 0.7 then 70% of the
armature reaction mmf gets added at this tip leading to considerable amount of saturation
under full load conditions. The flux distribution also is shown in Fig. 20. This is obtained
by multiplying mmf and permeance waves point by point in space. Actual flux distribution
differs from this slightly due to fringing. As seen from the figure, the flux in the inter polar
region is substantially lower due to the high reluctance of the medium. The air gaps under
the pole tips are also increased in practice to reduce excessive saturation of this part. The
advantage of the salient pole field construction is thus obvious. It greatly mitigates the
effect of the armature reaction. Also, the coils under going commutation have very little
emf induced in them and hence better commutation is achieved. Even though the armature
reaction produced a cross magnetizing effect, the net flux per pole gets slightly reduced,
on load, due to the saturation under one tip of the pole. This is more so in modern d.c.
machines where the normal excitation of the field makes the machine work under some level
of saturation.
In some small d.c. machines the brushes are shifted from the position of the mag-
netic neutral axis in order to improve the commutation. This is especially true of machines
with unidirectional operation and uni-modal (either as a generator or as a motor) operation.
Such a shift in the direction of rotation is termed ‘lead’ (or forward lead). Shift of brushes
in the opposite to the direction of rotation is called ‘backward lead’. This lead is expressed
in terms of the number of commutator segments or in terms of the electrical angle. A pole
pitch corresponds to an electrical angle of 180 degrees. Fig. 21 shows the effect of a forward
47
Rotation
a
Field c
flux
Armature flux
Total flux
Rotation
b’
a’ a
θ
b
S
(b)Calculation of demagnetizing mmf per pole
48
brush lead on the armature reaction. The magnetization action due to the armature is no
longer entirely cross magnetizing. Some component of the same goes to demagnetize the
main field and the net useful flux gets reduced. This may be seen as the price we pay for
improving the commutation. Knowing the pole arc to pole pitch ratio one can determine
the total mmf at the leading and trailing edges of a pole without shift in the brushes.
Fmax = Ff + α.Fa
where Ff is the field mmf, Fa is armature reaction mmf per pole, and α is the pole arc to
pole pitch ratio.
1 Z.Ic
Fa = . (23)
2 2p
The net flux per pole decreases due to saturation at the trailing edge and
hence additional ampere turns are needed on the pole to compensate this effect. This may
be to the tune of 20 percent in the modern d.c. machines.
The brush shift gives rise to a shift in the axis of the mmf of the armature
reaction. This can be resolved into two components, one in the quadrature axis and sec-
ond along the pole axis as shown in Fig. 21.(b) The demagnetizing and cross magnetizing
component of the armature ampere turn per pole can be written as
2θ
Fd = .Fa (24)
π
2θ
Fq = (1 − ).Fa (25)
π
where θ is the angle of lead . In terms of the number of commutator segments they are
Cl Ic Z Cl
Fd = .
C 4p
or .Ic .Z (26)
4p
C
49
As discussed earlier, for a given polarity of the field and sense of rotation, the
motoring and generating modes differ only in the direction of the armature current. Alter-
natively, for a given sense of armature current, the direction of rotation would be opposite
for the two modes. The leading and trailing edges of the poles change positions if direction
of rotation is made opposite. Similarly when the brush leads are considered, a forward lead
given to a generator gives rise to weakening of the generator field but strengthens the motor
field and vice-versa. Hence it is highly desirable, even in the case of non-reversing drives,
to keep the brush position at the geometrical neutral axis if the machine goes through both
motoring and generating modes.
The second effect of the armature reaction in the case of motors as well as
generators is that the induced emf in the coils under the pole tips get increased when a
pole tip has higher flux density. This increases the stress on the ‘mica’ (micanite) insulation
used for the commutator, thus resulting in increased chance of breakdown of these insulating
sheets. To avoid this effect the flux density distribution under the poles must be prevented
from getting distorted and peaky.
The third effect of the armature reaction mmf distorting the flux density is
that the armature teeth experience a heavy degree of saturation in this region. This increases
the iron losses occurring in the armature in that region. The saturation of the teeth may
be too great as to have some flux lines to link the thick end plates used for strengthening
50
the armature. The increase in iron loss could be as high as 50 percent more at full load
compared to its no-load value.
The above two effects can be reduced by providing a ’compensating’ mmf at
Commutating pole
s
Main pole
N S
N N
S N
Compensating
winding
s
the same spatial rate as the armature mmf. This is provided by having a compensating
winding housed on the pole shoe which carries currents that are directly proportional to the
armature current. The ampere conductors per unit length is maintained identical to that of
51
N S
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Rotation
mmf of
compensating
winding
Resultant
mmf
compole mmf
Armature
mmf
Main field
mmf
52
the armature. The sign of the ampere conductors is made opposite to the armature. This is
illustrated in Fig. 22 and Fig. 23 . Since the compensating winding is connected in series with
the armature, the relationship between armature mmf and the mmf due to compensating
winding remains proper for all modes of working of the machine. The mmf required to be
setup by the compensating winding can be found out to be
Ic .Z polearc
Fc = . (27)
4p polepitch
Under these circumstances the flux density curve remains unaltered under the poles between
no-load and full load.
The axis of the mmf due to armature and the compensating winding being
the same and the signs of mmf being opposite to each other the flux density in the region
of geometric neutral axis gets reduced thus improving the conditions for commutation. One
can design the compensating winding to completely neutralize the armature reaction mmf.
Such a design results in overcompensation under the poles. Improvement in commutation
condition may be achieved simply by providing a commutating pole which sets up a local
field of proper polarity. It is better not to depend on the compensating winding for improv-
ing commutation.
From the analysis of the phenomenon of armature reaction that takes place
in a d.c. machine it can be inferred that the equivalent circuit of the machine need not be
modified to include the armature reaction. The machine can simply be modelled as a voltage
53
source of internal resistance equal to the armature circuit resistance and a series voltage drop
equal to the brush contact drop, under steady state. With this circuit model one can arrive
at the external characteristics of the d.c. machine under different modes of operation.
5.1 Commutation
54
4 3 2 1 Ia Ia
2Ia
4 3 2 1
Motion
(a) 2Ia
tb
4 3 2 Ia 1 Ia
I2 i I1
4 3 2 1
x 2Ia
Thickness (b)
tb
dth 4 3 2 Ia Ia 1
Entering Edge Wi
Length
Leaving Edge
2Ia
4 3 2 1
(c) 2Ia
tb
55
5.1.1 Brushes
If the current density under the brush is assumed to be constant through out the
commutation interval, a simple model for commutation is obtained. For simplicity, the brush
thickness is made equal to thickness of one commutator segment. In Fig. 24(b), the brush
is initially solely resting on segment number 1. The total current of 2Ia is collected by
the brush as shown. As the commutator moves relative to the brush position, the brush
position starts to overlap with that of segment 2. As the current density is assumed to be
constant, the current from each side of the winding is proportional to the area shared on the
56
two segments. Segment 1 current uniformly comes down with segment 2 current increasing
uniformly keeping the total current in the brush constant. The currents I1 and I2 in brush
segments 1 and 2 are given by
x x
I1 = 2Ia (1 − ) and I2 = 2Ia (28)
tb tb
giving I1 + I2 to be 2 Ia .
Here ‘x’ is the width of the brush overlapping on segment 2. The process of commutation
would be over when the current through segment number 1 becomes zero. The current in
the coil undergoing commutation is
(I1 − I2 ) 2x
i = I1 − Ia = Ia − I2 = = Ia (1 − ) (29)
2 tb
tb
Tc = (30)
vc
where vc is the velocity of the commutator. This type of linear commutation is very close
to the ideal method of commutation. The time variation of current in the coil undergoing
commutation is shown in Fig. 25.(a). Fig. 25.(b) also shows the timing diagram for the
currents I1 and I2 and the current densities in entering edge αe , leaving edge αl and also the
mean current density αm in the brush. Machines having very low coil inductances, operating
at low load currents, and low speeds, come close to this method of linear commutation.
In general commutation will not be linear due to the presence of emf of self
induction and induced rotational emf in the coil. These result in retarded and accelerated
commutation and are discussed in sequence.
57
Ia 2Ia
I1 I2
i
αm = α’ = α"
0 Tc Time
Time of Time of Tc
communication -Ia 0 commutation
(a) (b)
Retarded commutation is mainly due to emf of self induction in the coil. Here
the current transfer from 1 to 2 gets retarded as the name suggests. This is best explained
with the help of time diagrams as shown in Fig. 26.(a). The variation of i is the change in
′
the current of the coil undergoing commutation, while i is that during linear commutation.
Fig. 26(b) shows the variation of I1 and current density in the brush at the leaving edge and
Fig. 26.(c) shows the same phenomenon with respect to I2 at entering edge. The value of
current in the coil is given by i undergoing commutation. αm is the mean current density in
the brush given by total current divided by brush area of cross section. αl and αe are the
current density under leaving and entering edges of the brush. As before,
I1 = Ia + i and I2 = Ia − i (31)
58
2Ia P α’=AB/AC
+Ia B
+Ia I1=Ia+i
i C αm
0 t Tc
i’
Q
-Ia 0 A t Tc
0 t Tc t
D
59
AB
αl = .αm (32)
AC
DF
αe = .αm (33)
DE
At the very end of commutation, the current density
′
di di
αe = αm . / (34)
dt dt
di 2Ia
= αm . /
dt Tc
If at this point di/dt = 0 the possibility of sudden breaking of the current and
hence the creation of an arc is removed .
60
α" α’ =AB/AC
α" =DB/DC
Ia D αm
C
α’
Ia B
0 i’ Tc
i
0 i’ Tc
i
B
-Ia A -Ia
(a1 ) (a2 )
α’ α"
Ia C S
A
αm B
Ia R α’ =PR/PQ
Q α" =SR/SQ
i Leaving edge
Entering
edge
0 Tc time
0 Tc time
i’
i
-Ia P -Ia
(b1 ) (b2 )
61
lead given to the brushes or by ii) having the field of suitable polarity at the position of the
brush with the help of a small pole called a commutating pole. In a non-inter pole machine
the brush shift must be changed from forward lead to backward lead depending upon gener-
ating or motoring operation. As the disadvantages of this brush shifts are to be avoided, it
is preferable to leave the brushes at geometric neutral axis and provide commutating poles
of suitable polarity (for a generator the polarity of the pole is the one towards which the
conductors are moving). The condition of commutation will be worse if commutating poles
are provided and not excited or they are excited but wrongly.
The action of the commutating pole is local to the coil undergoing commu-
tation. It does not disturb the main field distribution. The commutating pole winding
overpowers the armature mmf locally and establishes the flux of suitable polarity. The com-
mutating pole windings are connected in series with the armature of a d.c. machine to get
a load dependent compensation of armature reaction mmf.
The commutating pole are also known as compole or inter pole. The air gap
under compole is made large and the width of compole small. The mmf required to be
produced by compole is obtained by adding to the armature reaction mmf per pole Fa the
mmf to establish a flux density of required polarity in the air gap under the compole Fcp
.This would ensure straight line commutation. If sinusoidal commutation is required then
the second component Fcp is increased by 30 to 50 percent of the value required for straight
line commutation.
62
will be reduced by Fa * pole arc/pole pitch. This could have been predicted as the axis of
the compensating winding and armature winding is one and the same. Further, the mmf of
compensating winding opposes that of the armature reaction.
It is seen already that the equivalent circuit model of a d.c. machine becomes very
simple in view of the fact that the armature reaction is cross magnetizing. Also, the axis
of compensating mmf and mmf of commutating poles act in quadrature to the main field.
Thus flux under the pole shoe gets distorted but not diminished (in case the field is not
saturated). The relative connections of armature, compole and compensating winding are
unaltered whether the machine is working as a generator or as a motor; whether the load
is on the machine or not. Hence all these are connected permanently inside the machine.
The terminals reflect only the additional ohmic drops due to the compole and compensating
windings. Thus commutating pole winding, and compensating winding add to the resistance
of the armature circuit and can be considered a part of the same. The armature circuit
can be simply modelled by a voltage source of internal resistance equal to the armature
resistance + compole resistance + compensating winding resistance. The brushes behave
like non-linear resistance; and their effect may be shown separately as an additional constant
voltage drop equal to the brush drop.
The excitation for establishing the required field can be of two types a) Permanent
magnet excitation(PM) b) Electro magnetic excitation. Permanent magnet excitation is
63
Yoke
ly
lg lt lt lg lp
lp
Pole
Field
coil la
Armature
da
64
employed only in extremely small machines where providing a field coil becomes infeasible.
Also, permanent magnet excited fields cannot be varied for control purposes. Permanent
magnets for large machines are either not available or expensive. However, an advantage
of permanent magnet is that there are no losses associated with the establishment of the field.
The required ampere turns for establishing the desired flux per pole may be
computed by doing the magnetic circuit calculations. MMF required for the poles, air gap,
armature teeth, armature core and stator yoke are computed and added. Fig. 28 shows two
poles of a 4-pole machine with the flux paths marked on it. Considering one complete flux
loop, the permeance of the different segments can be computed as,
P = A.µ/l
Where P- permeance
A- Area of cross section of the part
mu- permeability of the medium
l- Length of the part
A flux loop traverses a stator yoke, armature yoke, and two numbers each
of poles, air gap, armature teeth in its path. For an assumed flux density Bg in the pole
region the flux crossing each of the above regions is calculated. The mmf requirement for
65
= B.A
From these expressions the mmf required for each and every part in the path of
the flux is computed and added. This value of mmf is required to establish two poles. It
is convenient to think of mmf per pole which is nothing but the ampere turns required to
be produced by a coil wound around one pole. In the case of small machines all this mmf
is produced by a coil wound around one pole. The second pole is obtained by induction.
This procedure saves cost as only one coil need be wound for getting a pair of poles. This
produces an unsymmetrical flux distribution in the machine and hence is not used in larger
machines. In large machines, half of total mmf is assigned to each pole as the mmf per pole.
The total mmf required can be produced by a coil having large number of turns but taking a
small current. Such winding has a high value of resistance and hence a large ohmic drop. It
can be connected across a voltage source and hence called a shunt winding. Such method of
excitation is termed as shunt excitation. On the other hand, one could have a few turns of
large cross section wire carrying heavy current to produce the required ampere turns. These
windings have extremely small resistance and can be connected in series with a large current
path such as an armature. Such a winding is called a series winding and the method of
excitation, series excitation. A d.c. machine can have either of these or both these types of
excitation.
These are shown in Fig. 29. When both shunt winding and series winding are
present, it is called compound excitation. The mmf of the two windings could be arranged to
aid each other or oppose each other. Accordingly they are called cumulative compounding
and differential compounding. If the shunt winding is excited by a separate voltage source
66
F1
A1
A1
F1
A1 F1 A1
67
then it is called separate excitation. If the excitation power comes from the same machine,
then it is called self excitation. Series generators can also be separately excited or self excited.
The characteristics of these generators are discussed now in sequence.
Ia=0
A2
Prime
mover E
n=const
Induced e.m.f
A1
Decreasing
Magnetisation
Increasing
magnetisation
+
F2 If
Vdc
e.m.f. due to Residual Magnetism
F1
- Exciting Current
(a) (b)
68
separate excitation is normally used for testing of d.c. generators to determine their open
circuit or magnetization characteristic. The excitation current is increased monotonically
to a maximum value and then decreased in the same manner, while noting the terminal
voltage of the armature. The load current is kept zero. The speed of the generator is held
at a constant value. The graph showing the nature of variation of the induced emf as a
function of the excitation current is called as open circuit characteristic (occ), or no-load
magnetization curve or no-load saturation characteristic. Fig. 30(b). shows an example. The
magnetization characteristic exhibits saturation at large values of excitation current. Due
to the hysteresis exhibited by the iron in the magnetic structure, the induced emf does not
become zero when the excitation current is reduced to zero. This is because of the remnant
field in the iron. This residual voltage is about 2 to 5 percent in modern machines. Separate
excitation is advantageous as the exciting current is independent of the terminal voltage
and load current and satisfactory operation is possible over the entire voltage range of the
machine starting from zero.
In a self excited machine, there is no external source for providing excitation current.
The shunt field is connected across the armature. For series machines there is no change in
connection. The series field continues to be in series with the armature.
Self excitation is now discussed with the help of Fig. 31.(a) The pro-
cess of self excitation in a shunt generator takes place in the following manner. When the
armature is rotated a feeble induced emf of 2 to 5 percent appears across the brushes de-
pending upon the speed of rotation and the residual magnetism that is present. This voltage
69
ms
210
oh
1500 rev/min
0
25
s
m
180
oh
0
17
1000 rev/min
s
ms
ohm
150 s
m
oh
oh
280
0
5
14
37
120
F2 A2 500 rev/min
90
Prime
mover 60
n
30
0 1.0
F1 A1
Exciting current,Amperes
Critical Resistance
160 180
s
hm
s
5o
hm
150
12
0o
25
120
120
ms
oh
60
80 90
40 60
30
200 400
0 1500
Total field circuit resistance, ohms speed in rev/min
70
D C Open circuit
characteristic
Q’
R’
Voltage
Q QL
R RL
P’
P PL Armature drop
P1 Q1 characteristic
O’ A’
0 P" Q" A
Excitation current If
Armature current Ia
gets applied across the shunt field winding and produces a small mmf. If this mmf is such
as to aid the residual field then it gets strengthened and produces larger voltage across the
brushes. It is like a positive feed back. The induced emf gradually increases till the voltage
induced in the armature is just enough to meet the ohmic drop inside the field circuit. Under
such situation there is no further increase in the field mmf and the build up of emf also stops.
If the voltage build up is ‘substantial’, then the machine is said to have ‘self excited’.
Fig. 31(b) shows the magnetization curve of a shunt generator. The field resistance
line is also shown by a straight line OC. The point of intersection of the open circuit charac-
71
teristic (OCC) with the field resistance line, in this case C, represents the voltage build up
on self excitation. If the field resistance is increased, at one point the resistance line becomes
a tangent to the OCC. This value of the resistance is called the critical resistance. At this
value of the field circuit resistance the self excitation suddenly collapses. See Fig. 31(c). In-
stead of increasing the field resistance if the speed of the machine is reduced then the same
resistance line becomes a critical resistance at a new speed and the self excitation collapses
at that speed. In this case, as the speed is taken as the variable, the speed is called the
critical speed. In the linear portion of the OCC the ordinates are proportional to the speed
of operation, hence the critical resistance increases as a function of speed Fig. 31.(b) and (d).
4. The speed of operation of the machine must be above the critical speed.
Remedial measures to be taken if the machine fails to self excite are briefly
discussed below.
1. The residual field will be absent in a brand new, unexcited, machine. The field may
be connected to a battery in such cases for a few seconds to create a residual field.
72
2. The polarity of connections have to be set right. The polarity may become wrong
either by reversed connections or reversed direction of rotation. If the generator had
been working with armature rotating in clockwise direction before stopping and if one
tries to self excite the same with counter clockwise direction then the induced emf
opposes residual field, changing the polarity of connections of the field with respect to
armature is normally sufficient for this problem.
3. Field circuit resistance implies all the resistances coming in series with the field winding
like regulating resistance, contact resistance, drop at the brushes, and the armature
resistance. Brush contact resistance is normally high at small currents. The dirt on
the commutator due to dust or worn out mica insulator can increase the total circuit
resistance enormously. The speed itself might be too low so that the normal field
resistance itself is very much more than the critical value. So ensuring good speed,
clean commutator and good connections should normally be sufficient to overcome this
problem.
4. Speed must be increased sufficiently to a high value to be above the critical speed.
5. The load switch must be opened or the load resistance is made very high.
The conditions for self excitation of a series generator remain similar to that of
a shunt machine. In this case the field circuit resistance is the same as the load circuit
resistance and hence it must be made very low to help self excitation. To control the field
mmf a small resistance called diverter is normally connected across the series field. To help
in the creation of maximum mmf during self excitation any field diverter if present must be
73
Terminal voltage
Open circuit PS=PQ-PR
characteristic
Q
Armature
S characteristic
A R B
External
characteristic
P
0
Load Current
open circuited.In a series generator load current being the field current of the machine the
self excitation characteristic or one and the same. This is shown in Fig. 33
Most of the compound machines are basically shunt machines with the series wind-
ing doing the act of strengthening/weakening the field on load, depending up on the con-
nections. In cumulatively compounded machines the mmf of the two fields aid each other
and in a differentially compounded machine they oppose each other. Due to the presence of
the shunt winding, the self excitation can proceed as in a shunt machine. A small difference
exists however depending up on the way the shunt winding is connected to the armature. It
can be a short shunt connection or a long shunt connection. In long shunt connection the
shunt field current passes through the series winding also. But it does not affect the process
of self excitation as the mmf contribution from the series field is negligible.
74
Both series field winding and shunt field winding are wound around the main
poles. If there is any need, for some control purposes, to have more excitation windings
of one type or the other they will also find their place on the main poles. The designed
field windings must cater to the full range of operation of the machine at nominal armature
current. As the armature current is cross magnetizing the demagnetization mmf due to pole
tip saturation alone need be compensated by producing additional mmf by the field.
The d.c. machines give rise to a variety of external characteristics with consid-
erable ease. The external characteristics are of great importance in meeting the requirements
of different types of loads and in parallel operation. The external characteristics, also known
as load characteristics, of these machines are discussed next.
Load characteristics are also known as the external characteristics. External char-
acteristics expresses the manner in which the output voltage of the generator varies as a
function of the load current, when the speed and excitation current are held constant. If
they are not held constant then there is further change in the terminal voltage. The terminal
voltage V can be expressed in terms of the induced voltage E and armature circuit drop as
V = E − Ia Ra − Vb (36)
75
Terminal voltage V B
Volts
0 A
Load current,Ia
As seen from the equation E being function of speed and flux per pole it will
also change when these are not held constant. Experimentally the external characteristics
can be determined by conducting a load test. If the external characteristic is obtained by
subtracting the armature drop from the no-load terminal voltage, it is found to depart from
the one obtained from the load test. This departure is due to the armature reaction which
causes a saturation at one tip of each pole. Modern machines are operated under certain
degree of saturation of the magnetic path. Hence the reduction in the flux per pole with
load is obvious. The armature drop is an electrical drop and can be found out even when
the machine is stationary and the field poles are unexcited. Thus there is some slight droop
in the external characteristics, which is good for parallel operation of the generators.
76
One could easily guess that the self excited machines have slightly higher droop
in the external characteristic as the induced emf E drops also due to the reduction in the
applied voltage to the field. If output voltage has to be held constant then the excitation
current or the speed can be increased. The former is preferred due to the ease with which it
can be implemented. As seen earlier, a brush lead gives rise to a load current dependent mmf
along the pole axis. The value of this mmf magnetizes/demagnetizes the field depending on
whether the lead is backward or forward.
For a given no-load voltage a self excited machine will have more voltage drop at
the terminals than a separately excited machine, as the load is increased. This is due to
the dependence of the excitation current also on the terminal voltage. After certain load
current the terminal voltage decreases rapidly along with the terminal current, even when
load impedance is reduced. The terminal voltage reaches an unstable condition. Also, in a
self excited generator the no-load terminal voltage itself is very sensitive to the point of inter-
section of the magnetizing characteristics and field resistance line. The determination of the
external characteristics of a shunt generator forms an interesting study. If one determines
the load magnetization curves at different load currents then the external characteristics
can be easily determined. Load magnetization curve is a plot showing the variation of the
terminal voltage as a function of the excitation current keeping the speed and armature cur-
rent constant. If such curves are determined for different load currents then by determining
the intersection points of these curves with field resistance line one can get the external
characteristics of a shunt generator. Load saturation curve can be generated from no-load
saturation curve /OCC by subtracting the armature drop at each excitation point. Thus
77
it is seen that these family of curves are nothing but OCC shifted downwards by armature
drop. Determining their intercepts with the field resistance line gives us the requisite result.
Instead of shifting the OCC downwards, the x axis and the field resistance line is shifted ‘up-
wards’ corresponding to the drops at the different currents, and their intercepts with OCC
are found. These ordinates are then plotted on the original plot. This is shown clearly in
Fig. 32. The same procedure can be repeated with different field circuit resistance to yield
external characteristics with different values of field resistance. The points of operation up to
the maximum current represent a stable region of operation. The second region is unstable.
The decrease in the load resistance decreases the terminal voltage in this region.
In the case of series generators also, the procedure for the determination of the
external characteristic is the same. From the occ obtained by running the machine as a sep-
arately excited one, the armature drops are deducted to yield external /load characteristics.
The armature drop characteristics can be obtained by a short circuit test as before.
Fig. 33 shows the load characteristics of a series generator. The first half of
the curve is unstable for constant resistance load. The second half is the region where series
generator connected to a constant resistance load could work stably. The load characteristics
in the first half however is useful for operating the series generator as a booster. In a booster
the current through the machine is decided by the external circuit and the voltage injected
into that circuit is decided by the series generator. This is shown in Fig. 35
78
- +
S1 S2
E2
A1 A2
F2 A2
Booster
Generator
E=E1 +- E2
E1
F1 A1
Main generator
79
IL
S2
If
F2 S1
A2
Load
Prime
Vf # mover
A1
F1
(a)-Connection
Over compounded
Level compounded
Under compounded
Shunt machine
Terminal voltage
Differential
compounding
Load current
(b)-Characteristics
80
D.C. generators are required to operate in parallel supplying a common load when
the load is larger than the capacity of any one machine. In situations where the load is small
but becomes high occasionally, it may be a good idea to press a second machine into operation
only as the demand increases. This approach reduces the spare capacity requirement and its
cost. In cases where one machine is taken out for repair or maintenance, the other machine
can operate with reduced load. In all these cases two or more machines are connected to
operate in parallel.
Parallel operation of two shunt generators is similar to the operation of two storage
batteries in parallel. In the case of generators we can alter the external characteristics easily
while it is not possible with batteries. Before connecting the two machines the voltages of
the two machines are made equal and opposing inside the loop formed by the two machines.
This avoids a circulating current between the machines. The circulating current produces
power loss even when the load is not connected. In the case of the loaded machine the
difference in the induced emf makes the load sharing unequal.
Fig. 37 shows two generators connected in parallel. The no load emfs are made
equal to E1 = E2 = E on no load; the current delivered by each machine is zero. As the load
is gradually applied a total load current of I ampere is drawn by the load. The load voltage
81
s1
v s2
A2 A2
Prime
mover Load
G1 G2
A1 A1
F2 F2
Vf1 Vf2
F1 F1
E1
V2 k V0 j
Total char
E2 a cteristic
V A B C
Terminal
V2
Voltage
I1 V1
I2
I=I1+I2
O D
Load current
82
under these conditions is V volt. Each machines will share this total current by delivering
currents of I1 and I2 ampere such that I1 + I2 = I.
Also terminal voltage of the two machines must also be V volt. This is dictated
by the internal drop in each machine given by equations
where Ra1 and Ra2 are the armature circuit resistances. If load resistance RL is known these
equations can be solved analytically to determine I1 and I2 and hence the manner in which
to total output power is shared. If RL is not known then an iterative procedure has to be
adopted. A graphical method can be used with advantage when only the total load current is
known and not the value of RL or V . This is based on the fact that the two machines have a
common terminal voltage when connected in parallel. In Fig. 38 the external characteristics
of the two machines are first drawn as I and II . For any common voltage the intercepts OA
and OB are measured and added and plotted as point at C. Here OC = OA + OB . Thus
a third characteristics where terminal voltage is function of the load current is obtained.
This can be called as the resultant or total external characteristics of the two machines put
together. With this, it is easy to determine the current shared by each machine at any total
load current I.
The above procedure can be used even when the two voltages of the machines
at no load are different. At no load the total current I is zero ie I1 + I2 = 0 or I1 = −I2 .
Machine I gives out electrical power and machine II receives the same. Looking at the voltage
equations, the no load terminal equation Vo becomes
83
As can be seen larger the values of Ra1 and Ra2 larger is the tolerance for the error
between the voltages E1 and E2 . The converse is also true. When Ra1 and Ra2 are nearly zero
implying an almost flat external characteristic, the parallel operation is extremely difficult.
Series generators are rarely used in industry for supplying loads. Some applications
like electric braking may employ them and operate two or more series generates in parallel.
Fig. 39 shows two series generators connected in parallel supplying load current of I1 and I2 .
If now due to some disturbance E1 becomes E1 + ∆E1 then the excitation of the machine I
increases, increasing the load current delivered. As the total current is I the current supplied
by machine II reduces, so also its excitation and induced emf. Thus machine I takes greater
and greater fraction of the load current with machine II shedding its load. Ultimately the
current of machine II becomes negative and it also loads the first machine. Virtually there is
a short circuit of the two sources, the whole process is thus highly unstable. One remedy is
for a problem as this is to make the two fields immune to the circulating current between the
machines. This is done by connecting an equalizer between the fields as shown in Fig. 39-a
. With the equalizer present, a momentary disturbance does not put the two machines out
of action. A better solution for such problems is to cross connect the two fields as shown in
Fig. 39-b. A tendency to supply a larger current by a machine strengthens the field of the
next machine and increases its induced emf . This brings in stable conditions for operation
rapidly.
84
I1 I1+I2
A2 + A2 +
I1 I2
A1 -
I1+I2 V
A1 -
S2 Equaliser S2
F1 F2
S1 S1 I1+I2
(a)Equalizer connection
I1+I2
I1 I2
A2 + A2 +
G1 G2
A1 - A1 - Load V
I1
S2 S2
I2
S1 I2 - S1
- I1+I2
85
F2 + A2 F2 + A2
Load V
A1 A1
- -
F1 F1
Equalizer
S2 S2
S1 S1
(a)Equalizer connection
F2 + A2 F2 + A2
Load V
A1 A1
- -
F1 F1
S2 S2
S1 S1
86
D.C. motors have a place of pride as far as electrical drives are considered. The
simplicity, and linearity of the control method makes them highly preferred machines in
precision drives. In spite of the great advancements in a.c. drives these machines are still
sought after by the industries. Apart from high precision application they are preferred in
stand alone systems working on batteries and high speed drives off constant voltage mains.
After the field is excited if we pass a current through the armature the rotor experiences
a torque and starts rotating. The direction of the torque can be readily obtained from the
law of interaction. These moving conductors cut the field and induce emf, usually called the
’back emf’ according to Lenz’s law and act as a sink of electrical power from the electrical
source. This absorbed power appears as mechanical power. The converted mechanical power
should overcome the frictional and iron losses before useful work could be done by the same.
The connections to the supply of a d.c. shunt motor are given in Fig. 41.
87
+ +
F2
A2 F2 A2
A1 F1 A1
F1
- -
(a)Separate excitation (b) Shunt excitation
s1
F2 +
s2
A2
DC
Supply
A1
F1
-
(c)Practical arrangement
88
Commonly used connection is where in both the field and the armature are
energized simultaneously Fig. 41(b). As the field has higher inductance and time constant
torque takes some time to reach the full value corresponding to a given armature current.
In Fig. 41.(c), the switch S1 is closed a few seconds prior to switch S2 . By then the field
current would have reached the steady value. So the torque per ampere is high in this case.
The only difference in the second connection Fig. 41.(a) is that the shunt field
winding is connected to a separate source. This connection is used when the armature and
field voltage are different as is common in high voltage d.c. machines. The field voltage is
kept low in such cases for the sake of control purposes. Here again the field circuit must
be energized prior to the armature. Suitable interlock should be provided to prevent the
armature switch being closed prior to / without closing of field circuit as the armature
currents reach very large values still not producing any torque or rotation. The relevant
equations for the motoring operation can be written as below
V − E − Ia Ra − Vb = 0 or E = V − Ia Ra − Vb (39)
p.φ.Z.n pZ
E= = Ke φ.n where Ke = (40)
b b
1 p.φ.ZIa 1 pZ
TM = . = Kt φIa where Kt = . (41)
2π b 2π b
dw
and TM − TL = J (42)
dt
where
TL - Load torque
TM - Motor torque
J - polar moment of inertia.
w - angular velocity = 2π.n
89
The first one is an electrical equation, the second and the third are electro
mechanical in nature and the last equation is the mechanical equation of motion. Ke and
Kt are normally termed as back emf constant and torque constant respectively. Under
steady speed of operation the fourth equation is not required. Using these equations one
can determine the torque speed characteristics of the machine for a given applied voltage.
These characteristics are similar to the external characteristics for a generator. Here the
torque on the machine is assumed to be varying and the corresponding speed of operation
is determined. This is termed as the torque speed characteristic of the motor.
Knowing the value of E and flux one can determine the value of the speed.
Also knowing the armature current and the flux, the value of the torque is found out. This
procedure is repeated for different values of the assumed armature currents and the values
are plotted as in Fig. 42-(a). From these graphs, a graph indicating speed as a function of
torque or the torque-speed characteristics is plotted Fig. 42-(b)(i).
As seen from the figure the fall in the flux due to load increases the speed due
to the fact that the induced emf depends on the product of speed and flux. Thus the speed
90
Line voltage
A No load speed B
Speed
E
F
Torque Flux
G
0 Armature current
(a)Load characteristics
(ii)
(i)
Speed
0 Torque
(b)Torque speed curve
91
of the machine remains more or less constant with load. With highly saturated machines
the on-load speed may even slightly increase at over load conditions. This effects gets more
pronounced if the machine is designed to have its normal field ampere turns much less
than the armature ampere turns. This type of external characteristics introduces instability
during operation Fig. 42(b)(ii) and hence must be avoided. This may be simply achieved by
providing a series stability winding which aids the shunt field mmf.
Following the procedure described earlier under shunt motor, the torque speed
characteristics of a series motor can also be determined. The armature current also happens
to be the excitation current of the series field and hence the flux variation resembles the
magnetization curve of the machine. At large value of the armature currents the useful flux
would be less than the no-load magnetization curve for the machine. Similarly for small
values of the load currents the torque varies as a square of the armature currents as the flux
is proportional to armature current in this region. As the magnetic circuit becomes more
and more saturated the torque becomes proportional to Ia as flux variation becomes small.
Fig. 43(a) shows the variation of E1 , flux , torque and speed following the above procedure
from which the torque-speed characteristics of the series motor for a given applied voltage
V can be plotted as shown in Fig. 43.(b) The initial portion of this torque-speed curve is
seen to be a rectangular hyperbola and the final portion is nearly a straight line. The speed
under light load conditions is many times more than the rated speed of the motor. Such
high speeds are unsafe, as the centrifugal forces acting on the armature and commutator
can destroy them giving rise to a catastrophic break down. Hence series motors are not
recommended for use where there is a possibility of the load becoming zero. In order to
92
Terminal voltage
Back emf
No load Useful
Magnetisation Flux
Torque, Flux and Speed curve
Developed Useful
Torque Torque
Speed
Load current
(a)Load characteristics
Speed
0 Torque
93
safeguard the motor and personnel, in the modern machines, a ‘weak’ shunt field is provided
on series motors to ensure a definite, though small, value of flux even when the armature
current is nearly zero. This way the no-load speed is limited to a safe maximum speed. It is
needless to say, this field should be connected so as to aid the series field.
Two situations arise in the case of compound motors. The mmf of the shunt field
and series field may oppose each other or they may aid each other. The first configuration
is called differential compounding and is rarely used. They lead to unstable operation of
the machine unless the armature mmf is small and there is no magnetic saturation. This
mode may sometimes result due to the motoring operation of a level-compounded generator,
say by the failure of the prime mover. Also, differential compounding may result in large
negative mmf under overload/starting condition and the machine may start in the reverse
direction. In motors intended for constant speed operation the level of compounding is very
low as not to cause any problem.
Cumulatively compounded motors are very widely used for industrial drives.
High degree of compounding will make the machine approach a series machine like charac-
teristics but with a safe no-load speed. The major benefit of the compounding is that the
field is strengthened on load. Thus the torque per ampere of the armature current is made
high. This feature makes a cumulatively compounded machine well suited for intermittent
peak loads. Due to the large speed variation between light load and peak load conditions, a
fly wheel can be used with such motors with advantage. Due to the reasons provided under
shunt and series motors for the provision of an additional series/shunt winding, it can be
94
seen that all modern machines are compound machines. The difference between them is only
in the level of compounding.
95
96
Speed
A B C D
Motors I and II
I II in parallel
0 Torque
97
In the case of series operation the motors shafts of the two machines are connected
to the same load and also the two armatures are series connected. This forces a common
armature current through both the machines and the torques developed by the machines
are proportional to the flux in each machine. Series operation of series motors is adopted
during starting to improve the energy efficiency. This method is ideally suited for shunt
and compound machines with nearly flat torque speed characteristics. Such machines can go
through high amount of dynamics without the fear of becoming unstable. This configuration
is used in steel mills. Having two smaller machines connected to the shaft is preferred over
there in place of one large machine as the moment of inertia of the motors is much reduced,
thus improving the dynamics.
98
Some elementary principles of application alone are dealt with here. The focus is
on the mechanical equation of dynamics which is reproduced here once again.
dw
TM − TL = J (43)
dt
Here TM and TL are the motor torque and the load torques respectively which are expressed
as functions of ω. Under steady state operation dω/dt will be zero. The application of
motors mainly looks at three aspects of operation.
1. Starting
2. Speed control
3. Braking
The speed of the machine has to be increased from zero and brought to the op-
erating speed. This is called starting of the motor. The operating speed itself should be
varied as per the requirements of the load. This is called speed control. Finally, the running
machine has to be brought to rest, by decelerating the same. This is called braking. The
torque speed characteristics of the machine is modified to achieve these as it is assumed
that the variation in the characteristics of the load is either not feasible or desirable. Hence
the methods that are available for modifying the torque speed characteristics and the actual
variations in the performance that these methods bring about are of great importance. When
more than one method is available for achieving the same objective then other criteria like,
initial cost, running cost, efficiency and ease operation are also applied for the evaluation of
the methods. Due to the absence of equipment like transformer, d.c. machine operation in
99
For the machine to start, the torque developed by the motor at zero speed must
exceed that demanded by the load. Then TM − TL will be positive so also is dω/dt, and the
machine accelerates. The induced emf at starting point is zero as the ω = 0 The armature
current with rated applied voltage is given by V /Ra where Ra is armature circuit resistance.
Normally the armature resistance of a d.c. machine is such as to cause 1 to 5 percent drop
at full load current. Hence the starting current tends to rise to several times the full load
current. The same can be told of the torque if full flux is already established. The machine
instantly picks up the speed. As the speed increases the induced emf appears across the
terminals opposing the applied voltage. The current drawn from the mains thus decreases,
so also the torque. This continues till the load torque and the motor torque are equal to
each other. Machine tends to run continuously at this speed as the acceleration is zero at
this point of operation.
The starting is now discussed with respect to specific machines.
100
If armature and field of d.c. shunt motor are energized together, large current is
drawn at start but the torque builds up gradually as the field flux increases gradually. To
improve the torque per ampere of line current drawn it is advisable to energize the field
first. The starting current is given by V /Ra and hence to reduce the starting current to a
safe value, the voltage V can be reduced or armature circuit resistance Ra can be increased.
Variable voltage V can be obtained from a motor generator set. This arrangement is called
Ward-Leonard arrangement. A schematic diagram of Ward-Leonard arrangement is shown
in Fig. 45. By controlling the field of the Ward-Leonard generator one can get a variable
voltage at its terminals which is used for starting the motor.
The second method of starting with increased armature circuit resistance can
be obtained by adding additional resistances in series with the armature, at start. The
current and the torque get reduced. The torque speed curve under these conditions is shown
in Fig. 46(a) . It can be readily seen from this graph that the unloaded machine reaches its
final speed but a loaded machine may crawl at a speed much below the normal speed. Also,
the starting resistance wastes large amount of power. Hence the starting resistance must
be reduced to zero at the end of the starting process. This has to be done progressively,
making sure that the current does not jump up to large values. Starting of series motor and
compound motors are similar to the shunt motor. Better starting torques are obtained for
compound motors as the torque per ampere is more. Characteristics for series motors are
given in fig. 47.
101
+ -
A2 A2 A2
Load variable
M G M
voltage
A1 A1 A1
+ + +
F2 F2 F2
constant
voltage
mains
F1
- F1
-
F1 -
(a)
+
A2
Variable Load
voltage
dc A1
Auto
Diode -
transformer
Constant bridge
voltage
ac mains
+
F2
Static Ward
Leonard system
F1
-
(b)
102
+ - Rext = 0
Rext
F2 A2 Rext increasing
E1
Speed
F1 A1
Constant voltage
source 0 Torque
(a)
v + -
If
F2 > A2
Rext If2 < If rated If2
Vf E1 If rated
Speed
F1 A1
Constant voltage
source
0
Torque
(b)
+ -
F2 A2 V1
Vf E1 V
V2
Speed
F1 A1
V3
V3 < V2 < V1
Variable voltage
source 103
0 Torque
(c)
v
S2 + -
Rext
S1
A2
Speed
E1
A1
Rext = 0
Rext > 0
Constant voltage
0 Torque
sources
(a)
+ -
S2
v
Rd
S1
Speed
A2
Rd =
8
E1
A1 Constant Rd reducing
voltage
0 Torque
sources
(b)
S2 + -
S1
A2
Variable
Speed
V Vrated
voltage
M
A1
V reducing
104
0 Torque
(c)
A2 F2
ra
Rn+1
A1
n+1
rn
R3 n F1
n-1 rn-1
R2
R1 r2
3
2
r1
1
(a)Physical connection
graphical method
Rn+1
Rn
Starting current with time
Rn-1 Imax
R3
Volts
R2 Imin
0 Ia R1
0
Imin Imax Time
105
If the starting resistor is reduced in uniform steps then the current peaks reached
as we cut down the resistances progressively increase. To ascertain that at no step does
the current jump to a large value non-uniform reduction of resistances must be assorted to.
This use of a non-uniform resistance step is called ‘grading’ of the resistors. The calculations
for a starter resistance of a shunt motor are shown below with the help of Fig. 48. In
the figure an n element or n+1 step starter is shown. The armature resistance when all
the external resistances are cut off is ra . The total armature circuit resistance at step 1 is
R1 = (r1 + r2 + ... + rn ) + ra . The field winding is connected across the supply. The starting
current reaches a maximum value Im ax when we move on to a step. One resistance element
is cut from the circuit when the current falls down to Im in . During the instant when the
element is cut the speed and hence the induced emf does not change but the current jumps
back to Im ax . Thus during the starting the current changes between two limits Im ax and
Im in. Writing the expression for the current before and after the resistance is changed on
step Ri and Ri+1 , we have
V −E V −E Im ax Ri
Im in = Im ax = or = (47)
Ri Ri+1 Im in Ri+1
Im ax R1 R2 Rn−1 Rn
= = = ... = = = k(say) (48)
Im in R2 R3 Rn Rn+1
r
n R1 R2 Rn R1 R1 R1
k = ∗ ∗ ... ∗ = = k= n (49)
R2 R3 Rn+1 Rn+1 ra ra
106
Sometimes the ratio k may be required to be fixed. Then the number of steps required can
be calculated as
R1 log Rra1
n log k = log ,n = (50)
ra log k
log R1 − log Rn
log k
Also, r r r
V V V
R= n
= n
= n+1
(51)
I1 ra RI2 ra I2 ra
From these expressions it is seen that to have the ratio k to be unity, the number
of steps should be infinity. Smaller the number of steps larger is the ratio of maximum to
minimum current. Also, it is not possible to choose n and k independently. Im ax is set
by the maximum possible starting current from the point of view of commutation. Im in
is found from the minimum torque against which the starting is required to be performed.
Similar method exists in the case of series motors and compound motors. In these cases the
ratio of currents and the ratio of fluxes are needed. The equation becomes non-linear and a
graphical method is normally adopted for the design of the resistances in those cases.
Resistance method of starting is cheaper and simple and hence is used univer-
sally. But it wastes energy in the starting resistor. Hence this method is not advised when
frequent starting of the motor is required. Ward-Leonard method gives a energy efficient
method of starting. With the help of a auto transformer and rectifier set one can get variable
voltage d.c. supply from a constant voltage a.c power source. This is some times called a
static Ward-Leonard arrangement. This method is becoming more popular over the rotating
machine counter part.
107
In the case of speed control, armature voltage control and flux control methods
are available. The voltage control can be from a variable voltage source like Ward-Leonard
arrangement or by the use of series armature resistance. Unlike the starting conditions the
series resistance has to be in the circuit throughout in the case of speed control. That means
considerable energy is lost in these resistors. Further these resistors must be adequately
cooled for continuous operation. The variable voltage source on the other hand gives the
motor the voltage just needed by it and the losses in the control gear is a minimum. This
method is commonly used when the speed ratio required is large, as also the power rating.
Field control or flux control is also used for speed control purposes. Normally
field weakening is used. This causes operation at higher speeds than the nominal speed.
Strengthening the field has little scope for speed control as the machines are already in a
state of saturation and large field mmf is needed for small increase in the flux. Even though
flux weakening gives higher speeds of operation it reduces the torque produced by the ma-
chine for a given armature current and hence the power delivered does not increase at any
armature current. The machine is said to be in constant power mode under field weakening
mode of control. Above the nominal speed of operation, constant flux mode with increased
applied voltage can be used; but this is never done as the stress on the commutator insulation
increases.
Thus operation below nominal speed is done by voltage control. Above the
nominal speed field weakening is adopted. For weakening the field, series resistances are used
for shunt as well as compound motors. In the case of series motors however field weakening
108
is done by the use of ’diverters’ . Diverters are resistances that are connected in parallel to
the series winding to reduce the field current without affecting the armature current.
When a motor is switched off it ‘coasts’ to rest under the action of frictional forces.
Braking is employed when rapid stopping is required. In many cases mechanical braking
is adopted. The electric braking may be done for various reasons such as those mentioned
below:
5. To improve the through put in many production process by reducing the stopping time.
In many cases electric braking makes more brake power available to the braking
process where mechanical brakes are applied. This reduces the wear and tear of the me-
chanical brakes and reduces the frequency of the replacement of these parts. By recovering
the mechanical energy stored in the rotating parts and pumping it into the supply lines
the overall energy efficiency is improved. This is called regeneration. Where the safety of
the personnel or the equipment is at stake the machine may be required to stop instantly.
Extremely large brake power is needed under those conditions. Electric braking can help
in these situations also. In processes where frequent starting and stopping is involved the
109
process time requirement can be reduced if braking time is reduced. The reduction of the
process time improves the throughput.
Basically the electric braking involved is fairly simple. The electric motor can
be made to work as a generator by suitable terminal conditions and absorb mechanical energy.
This converted mechanical power is dissipated/used on the electrical network suitably.
Braking can be broadly classified into:
1. Dynamic
2. Regenerative
These are now explained briefly with reference to shunt ,series and compound motors.
• Shunt machine
In dynamic braking the motor is disconnected from the supply and connected to a
dynamic braking resistance RDB . In and Fig. 49 this is done by changing the switch
from position 1 to 2 . The supply to the field should not be removed. Due to the
rotation of the armature during motoring mode and due to the inertia, the armature
continues to rotate. An emf is induced due to the presence of the field and the rotation.
This voltage drives a current through the braking resistance. The direction of this
current is opposite to the one which was flowing before change in the connection.
Therefore, torque developed also gets reversed. The machine acts like a brake. The
110
torque speed characteristics separate by excited shunt of the machine under dynamic
braking mode is as shown in Fig. 49(b) for a particular value of RDB . The positive
torque corresponds to the motoring operation. Fig. 50 shows the dynamic braking of
a shunt excited motor and the corresponding torque-speed curve. Here the machine
behaves as a self excited generator.
Below a certain speed the self-excitation collapses and the braking action becomes
Zero.
• Series machine
In the case of a series machine the excitation current becomes zero as soon as the
armature is disconnected from the mains and hence the induced emf also vanishes. In
order to achieve dynamic braking the series field must be isolated and connected to
a low voltage high current source to provide the field. Rather, the motor is made to
work like a separately excited machine. When several machines are available at any
spot, as in railway locomotives, dynamic braking is feasible. Series connection of all
the series fields with parallel connection of all the armatures connected across a single
dynamic braking resistor is used in that case.
• Compound generators
In the case of compound machine, the situation is like in a shunt machine. A separately
excited shunt field and the armature connected across the braking resistance are used.
A cumulatively connected motor becomes differentially compounded generator and the
braking torque generated comes down. It is therefore necessary to reverse the series
field if large braking torques are desired.
111
1 +
F2 + A2 RDB
Vf E
F1
- A1 2
1
-
(a)Connections
Torque 0
(b)Characteristics
112
1 +
F2 + A2 RDB
Vf E
F1
- A1 2
1
-
(a)Connections
Torque 0
(b)Characteristics
113
In regenerative braking as the name suggests the energy recovered from the rotating
masses is fed back into the d.c. power source. Thus this type of braking improves the energy
efficiency of the machine. The armature current can be made to reverse for a constant voltage
operation by increase in speed/excitation only. Increase in speed does not result in braking
and the increase in excitation is feasible only over a small range, which may be of the order of
10 to 15%. Hence the best method for obtaining the regenerative braking is to operate, the
machine on a variable voltage supply. As the voltage is continuously pulled below the value of
the induced emf the speed steadily comes down. The field current is held constant by means
of separate excitation. The variable d.c. supply voltage can be obtained by Ward-Leonard
arrangement, shown schematically in Fig. 51. Braking torque can be obtained right up to
zero speed. In modern times static Ward-Leonard scheme is used for getting the variable
d.c. voltage. This has many advantages over its rotating machine counter part. Static set
is compact, has higher efficiency, requires lesser space, and silent in operation; however it
suffers from drawbacks like large ripple at low voltage levels, unidirectional power flow and
low over load capacity. Bidirectional power flow capacity is a must if regenerative braking is
required. Series motors cannot be regeneratively braked as the characteristics do not extend
to the second quadrant.
8.3.3 Plugging
The third method for braking is by plugging.Fig. 52 shows the method of connection
for the plugging of a shunt motor. Initially the machine is connected to the supply with the
switch S in position number 1. If now the switch is moved to position 2, then a reverse
voltage is applied across the armature. The induced armature voltage E and supply voltage
114
+ -
If
F2 A2 Variable
votage
Vf E V source
F1 A1
(a)Physical connection
Speed
A
B V1
C
V2
V1 > V2
0 Torque
(b)Characteristics
115
2 RB
+ -
1
+
F2 A2
V
Vf
E
F1 2
- A1
1
(a)Physical connection
A
B
Speed
C 0 Torque
(b)Characteristics
116
V aid each other and a large reverse current flows through the armature. This produces a
large negative torque or braking torque. Hence plugging is also termed as reverse voltage
braking. The machine instantly comes to rest. If the motor is not switched off at this instant
the direction of rotation reverses and the motor starts rotating the reverse direction. This
type of braking therefore has two modes viz. 1) plug to reverse and 2) plug to stop. If we
need the plugging only for bringing the speed to zero, then we have to open the switch S
at zero speed. If nothing is done it is plug to reverse mode. Plugging is a convenient mode
for quick reversal of direction of rotation in reversible drives. Just as in starting, during
plugging also it is necessary to limit the current and thus the torque, to reduce the stress on
the mechanical system and the commutator. This is done by adding additional resistance in
series with the armature during plugging.
• Series motors
In the case of series motors plugging cannot be employed as the field current too gets
reversed when reverse voltage is applied across the machine. This keeps the direction
of the torque produced unchanged. This fact is used with advantage, in operating a
d.c. series motor on d.c. or a.c. supply. Series motors thus qualify to be called as
‘Universal motors’.
• Compound motors
Plugging of compound motors proceeds on similar lines as the shunt motors. However
some precautions have to be observed due to the presence of series field winding. A
cumulatively compounded motor becomes differentially compounded on plugging. The
mmf due to the series field can ’over power’ the shunt field forcing the flux to low values
or even reverse the net field. This decreases the braking torque, and increases the
duration of the large braking current. To avoid this it may be advisable to deactivate
117
the series field at the time of braking by short circuiting the same. In such cases the
braking proceeds just as in a shunt motor. If plugging is done to operate the motor
in the negative direction of rotation as well, then the series field has to be reversed
and connected for getting the proper mmf. Unlike dynamic braking and regenerative
braking where the motor is made to work as a generator during braking period, plugging
makes the motor work on reverse motoring mode.
It is seen from the earlier sections that the d.c.machine is capable of having variety of
torque-speed characteristics depending on the circuit conditions. The need for generating
these characteristics will be clear only when they are seen along with the characteristics of
the loads that they operate with. Even though a detailed treatment of motor load systems is
outside the scope here, it may be useful to look into the typical torque-speed characteristics
of some of the common loads.
Loads are broadly divided into,
Passive loads absorb the mechanical energy developed by the motors while
active loads are capable of working as both sinks and sources for mechanical energy. The
direction of rotation may be taken to be clockwise/counter clockwise rotation. Normally the
118
direction in which the load operates most of the time, is taken as the positive direction of
rotation. Any torque which accelerates the motor load system in the positive direction of
rotation is termed as a positive toque. With this rotation torques of motors, generators or
loads can be represented graphically on a four quadrantal diagram. The torque being taken
as an independent variable, is represented along the x-axis. Y-axis represents the speed.
Quadrants. I and III in Fig. 53(a) represent ‘forward motoring’ and ‘reverse motoring’ op-
eration respectively. Quadrants II and IV similarly represent generating/braking quadrants
as they absorb mechanical power and cause braking action.
Fig. 53(b) shows a few typical load characteristics on a four quadrantal dia-
gram.
The characteristics a, b,and c correspond to frictional torque, cutting torque and fan torque
respectively. While the frictional torque is not a function of speed, the cutting toque is pro-
portional to the speed and the fan torque varies as the square of the speed. These can only
absorb mechanical power and hence are represented in quadrantal II for positive direction
of rotation. Similar loads produce characteristics in quadrant IV for negative direction of
rotation.
Fig. 54 shows a typical behaviour of an active load. Here an elevator is taken as an example.
Here the counter weight is assumed to be heavier than the cage and similarly the loaded
cage in assumed to be heavier than the counter weight. As seen from the Fig. 54 the torque
is constant and depends on the difference in the weight of the case and the counter weight,
and the radius of the drum. The characteristics of the load exists in all the four quadrants
and is capable of delivering as well as absorbing mechanical power. Hence it is called as an
active load. The governing equation when the motor and a load are connected together is
dw
TM (w) − TL (w) = J (52)
dt
119
Speed
II I
Torque
III IV
(a)
b a Speed
Torque
a b
(b)
120
W
T W T
speed
Hoisting an
empty cage Hoisting a
loaded cage
Torque
o
T
W
T W
Lowering a
Lowering an loaded cage
empty cage
121
where TM (w) and TM (w) are motor and load torques respectively. J is the polar moment of
dw
inertia of the motor and load put together at the motor shaft. dt
is made positive when the
speed has to be increased in the positive direction and negative when reducing the speed.
Under steady operation TM (w) − TL (W ) = 0. Both motor and load torques are expressed as
functions of the speed. The speed at which motor and load torques are equal and opposite is
the steady state operating speed. By varying the characteristics of the motor (or the load),
this speed can be changed to suit our requirements. Normally the torque speed characteris-
tics of a load cannot be changed easily. Thus most speed control methods adopt, varying the
motor characteristics to achieve speed control. Some typical loads and the motors commonly
used to drive the same are tabulated in Table.
d.c. shunt motor lathes,fans,pumps disc and band saw drive requiring moderate torques.
d.c. series motor Electric traction, high speed tools
d.c. compound motor Rolling mills and other loads requiring large momentary toques.
122
A d.c. machine has to be tested for proper fabrication and trouble free operation.
From the tests one can determine the external characteristics needed for application of these
machines. Also, one can find the efficiency, rating and temperature rise of the machine.
Some of the tests are discussed in sequence now.
Even though any value of applied voltage can be used, the highest permissible
voltage/current is chosen during the test to minimize the errors. The armature circuit
consists of two resistances in series. They are armature winding resistance and resistance
due to the brushes and the brush drop. The brush contact drop behaves like a non-linear
resistance. To separate this from the armature circuit resistance and brush resistance a
number of v-i readings are taken. An equation of V = Vb + IRa form is fitted through
these test points shown graphically in Fig. 55. For large values of I the equivalent armature
resistance is taken to be V /I ohm. If the value of brush drop Vb can be neglected then the
armature resistance Ra = V /I ohm.
The OCC is of great value as it shows the mmf and hence the field current required
to generate a given voltage at any speed, on no load. It is a graph showing the variation
123
v
+
A
A2
DC Supply
A1 -
(a)Physical connection
I.V. Characteristic
dv
Ra =
di
dv
V di
Vb
0 I
(b)Characteristics
124
of the induced emf as a function of excitation current, when the speed is held constant,
with the load current being zero. It is also called the no-load saturation curve or no load
magnetization characteristic. This is experimentally determined by running the machine
as a separately excited generator on no-load at a constant speed and noting the terminal
voltage as a function of the excitation current. This curve can be used to find the OCC at
other speeds and also the self excited voltage when the machine works as a shunt generator.
In the case of short circuit test the armature is kept short circuited through an
ammeter. The machine is demagnetized and an extremely small field current is passed
through the field. The variation of the short circuit current as a function of excitation
current is plotted as the SCC. The speed is to be held constant during this test also. The
short circuit test gives an idea of the armature drop at any load current.
To assess the rating of a machine a load test has to be conducted. When the
machine is loaded, certain fraction of the input is lost inside the machine and appears as
heat, increasing the temperature of the machine. If the temperature rise is excessive then
it affects the insulations, ultimately leading to the breakdown of the insulation and the
machine. The load test gives the information about the efficiency of a given machine at any
load condition. Also, it gives the temperature rise of the machine. If the temperature rise
is below the permissible value for the insulation then the machine can be safely operated
at that load, else the load has to be reduced. The maximum continuous load that can be
125
delivered by the machine without exceeding the temperature rise for the insulation used, is
termed as the continuous rating of the machine. Thus the load test alone can give us the
proper information of the rating and also can help in the direct measurement of the efficiency.
The moment of inertia value is very important for the selection of a proper motor
for drives involving many starts and stops or requiring very good speed control characteris-
tics. The inertia can be determined by a retardation test.
The test works on the principle that when a motor is switched off from the
mains it decelerates and comes to rest. The angular retardation at any speed is proportional
to the retarding torque and is inversely proportional to the inertia. The torque lost at
any speed is calculated by running the motor at that speed steadily on no load and noting
the power input.From this power the losses that takes place in the armature and field are
deducted to get the power converted into mechanical form. All this power is spent in over
coming the mechanical losses at that speed. This can be repeated at any defined speed to
get the lost power (PL ) and torque lost (Tlost ) due to mechanical losses. In a retardation
test the motor speed is taken to some high value and the power to the motor is switched off.
The torque required by the losses is supplied by the energy stored in the motor inertia. The
lost torque at any speed can be written as
PL = Tlost .ω (53)
dw
Tlost = PL /w = J
dt
dw
Here the dt
is the slope of the retardation curve and the (Tlost ) is the torque required to be
126
met at the given speed. From these values the moment of inertia can be computed as
Tlost PL
J= dw
= kgm2 (54)
dt
w. dw
dt
A machine when loaded yields an output. The input to the machine is measured
at that operating point. The the efficiency in per unit is given as the ratio of output power
to input power.
output power
η = (55)
input power
Input power − power lost inside the machine
=
input power
output power
=
output power + power lost inside the machine
The first definition is used in the direct estimation of the efficiency . The other
two definitions are known as determination of efficiency using the loss segregation. For the
segregation of losses one must know the losses that take place inside a d.c. machine. The
losses that take place inside a d.c. machine can be listed as below.
127
7. Mechanical losses
8. Iron losses
Out of these items 1,2,7,8 and 9 will be present in all the d.c. machines. Out of
the remaining one or more may be present depending on which winding is present. These
losses change with temperature of operation. Mechanical losses vary with variation in speed.
Iron losses change with the degree of saturation and distortion of the shape of the field flux
distribution under the poles.
When a d.c. machine is loaded using a suitable load the output delivered by
the machine increases. The input requirement also increases along with the output. The
difference between the input and output powers is the power lost inside the machine as loss.
The efficiency of power conversion is given by the ratio of output power to input power.
Putting in mathematical form for a motor,
V I − losses
η= (56)
VI
for constant speed operation, the speed dependant losses remain constant. The load depen-
dant losses form the variable losses. While the loss that takes place in the brush drop in the
brushes is proportional to the load current, the loss that takes place in the resistance of the
armature is proportional to the square of the load current. Even though the loss that takes
place in a field winding is proportional to the square of the current through that winding, it
is classified under constant losses as the excitation current is held constant during loading.
128
Thus the total losses in a d.c. motor can be expressed in the form
PL = a + bI + cI 2 (57)
V I − PL A
η= = 1 − ( + B + CI) (58)
VI I
a b
When A = V
,B = V
and C = cV .
The term inside the brackets is sometimes referred to as the deficiency. For a
typical d.c.motor these are plotted in Fig. 56(a) as a function of the load current. The
curves a,b,c in the figure represent the efficiency curve taking one component of the loss
at a time. The curve d is the efficiency curve with all three components taken together.
The resultant curve exhibits a maximum. This can be easily seen from the graph that
this maximum occurs when constant losses equal the variable losses. AI = CI or A = CI 2 .
Fig. 56(b) depicts a typical output vs η curve of a d.c.machine.
129
b
c
a
Efficiency
d
current
Output
(b)Output Vs Efficiency
130
Synchronous Machines
1 Introduction
With the development of the technology and the way in which human labour is get-
ting minimized and the comforts increasing tremendously the use of electrical energy is ever
increasing. Basically electric power is the main source of energy for carrying out many func-
tions, as it is a clean and efficient energy source, which can be easily transmitted over long
distances. With the availability of Transformer for changing the voltage levels to a very high
value (of say 132kV to 400kV) the use of AC power has increased rapidly and the DC power
is used only at remote places where AC power cannot be supplied through power lines or
cables or for a few social purposes.
In 1831 Faraday discovered that an emf can be induced (or generated) due to relative
motion between a magnetic field and a conductor of electricity. This voltage was termed
as the induced emf since the emf is produced only due to motion between the conductor
and the magnetic field without actual physical contact between them. The principle of
electromagnetic induction is best understood by referring to Fig. 1. The magnetic field is
produced by the two fixed poles one being the north pole from which the magnetic flux
lines emerge and enter into the other pole known as the south pole. It was found that the
magnitude of the voltage induced in the conductor is proportional to the rate of change of
flux lines linking the conductor.
Mathematically it is given as
dφ φ
e= ≈ volts (1)
dt t
Force on conductor
l producing V
- Conductor
B
Induced N
EMF e
+
where
φ = flux in Webers
t = time in seconds
e = average induced emf in volts.
The above Eqn. 1 holds good only when the magnetic circuit is physically the same at
the end as at the beginning and also during the period of change of flux linkages as well. In
practical rotating machinery, however the change of flux linking each individual conductor
during rotation (of either the conductors or the poles) is not clearly defined or cannot be
easily measured. It is therefore more convenient to express this rate of change of flux in terms
of an average flux density (assumed constant) and the relative velocity between this field
and a single conductor moving through it. For the conductor of active length l moving with
a velocity of v in a magnetic field of flux density B, as shown in Fig. 1, the instantaneous
induced emf is expressed as,
e = Blv V olts (2)
where
B= flux density in Tesla (Wb/m2 )
l = active conductor length (m)
v = relative linear velocity between the conductor and the field (m/s).
This animation would help to understand the concept for a coil rotating in a magnetic field.
Thus the instantaneous voltage e and the average value E of the induced emf are
e = Emsinωt = Emsinθ
e
Em
θ = ωt π 2π
the same if the flux density B and the relative velocity v are both uniform and constant. In
an alternator we want the instantaneous emf to be varying in a sinusoidal manner as shown
in Fig. 2. Hence we should have a field system which will produce a sinusoidal distribution
of flux density in the plane perpendicular to the plane of motion of the conductor.Then,
B
V
V V
B
V V V V
S N
V
S N
V V
V
(a) Conductor moving at right angles (b) Conductor moving parallel to mag-
to magnetic field netic field
B B
V VVVV V V
δ
S δ N S 1800-δ N
(c) Conductor moving at any angle (d) Conductor moving at any angle
across magnetic field across magnetic field
magnetic field as shown in Fig. 1. Eqn. 1 or Eqn. 2 are valid only for this mutually orthogonal
condition for B and v. The other possible cases of motion of conductor with respect to B
are shown in Fig.3 in addition to the mutually orthogonal condition of Fig. 1. When the
conductor moves parallel to B, the induced emf will be zero because the rate of change of
flux linkage is zero as the conductor does not link any new flux line/lines. To account for this
condition of operation, Eqn. 2 must be multiplied by some factor, that takes into account
the direction of motion of conductor so as to make ‘e’ zero for this condition of operation
although B, l and v are finite quantities. Intuitively we may infer that this factor must be
a sine function as it has a zero value at 0◦ and also at 180◦ and a maximum value at 90◦ .
Indeed the emf equation for the general case of a conductor moving in any direction with
respect to the field as shown in Fig. 3 is given by
where δ is the angle formed between B and v always taking B as the reference. All other
quantities are the same as in Eqn. 2.
Motion
S
Field
Induced EMF
(a) Right-hand rule
Motion
B
V V V V
S Induced EMF N
V
The direction of the induced emf is given by Fleming’s Right Hand Rule which
states: If the thuMb, First finger and the seCond finger of the right hand are stretched out
and held in three mutually perpendicular directions such that the First finger is held pointing
in the direction of the magnetic field and the thuMb pointing in the direction of motion,
then the seCond finger will be pointing in the direction of the induced emf such that the
current flows in that direction. As shown in Fig. 4 the induced emf is in a direction so as to
circulate current in the direction shown by the middle finger. Schematically we indicate the
direction of the emf by a dot as shown in Fig. 5(a) to represent an emf so as to send current
in a direction perpendicular to the plane of the paper and out of it. A cross will indicate the
emf of opposite polarity, see Fig. 5(b). Although the Right Hand Rule assumes the magnetic
filed to be stationary, we can also apply this rule to the case of a stationary conductor and
moving magnetic field, by assuming that the conductor is moving in the opposite direction.
For example, as shown in Fig. 4 the direction of the induced emf will be the same if the poles
producing the field had been moved upwards.
The motion of the conductor in a magnetic field can be imparted by the applica-
tion of an external mechanical force to the conductor. In such a case the mechanical work
done in moving the conductor is converted to an electric energy in agreement with the law
of conservation of energy. The electric energy is not produced by the magnetic field since
the field is not changed or destroyed in the process. The name electro mechanical energy
conversion is given to the process of converting energy from mechanical form obtained from
a prime mover, such as an IC engine, water/steam turbine etc, into electric energy.
The emf induced in the conductor will circulate a current through it if a closed circuit
is formed by an external connection. The direction of the current flowing in the conductor
will be such as to oppose the cause of it as stated by Lenz’s Law. A current carrying
conductor located in a magnetic field will experience a force given by Biot-savart’s law:
f = Bli (5)
In other words, whenever a change in flux linkages occur, an emf is induced that
tends to set up a current in such a direction as to produce a magnetic flux that opposes the
cause of it. Thus if a current carrying conductor is placed in a magnetic field as shown in
Fig. 5 the current tends to produce a magnetic field in the direction shown by the dotted
circles.
Generator
Direction
of force
Motor
Motor
Direction
of force
Generator
The direction of the flux lines around the current carrying conductor can be easily
determined by Corkscrew Rule - which states that the flux lines will be in the same direction
as the rotation of a right threaded screw so as to advance in the direction of flow of current.
As a result the magnetic field, for the case shown in Fig. 5(a), is strengthened at the top and
weakened at the bottom of the conductor, thereby setting up a force to move the conductor
downwards. For the case of a Generator, the conductor must be moved up against this
counter force or the opposing force. Similarly the current is to be supplied to the conductor
against the emf generated (known as the counter emf or back emf) in the conductor as it
moves due to the motor action. Thus, the same machine can be operated as a generator
or a motor, depending on whether we supply mechanical power or electrical power to it,
respectively.
Fiel
dw
in d
in g
V
Rotor
V
-a
a N -t
ur
V
arm n
at
w in u re
din
g
V
S tator
V
Flux paths
The generators shown in Fig. 1 and Fig. 4 and discussed in the earlier sections
are clearly impractical for a number of reasons. The main reason is that such generators
require a prime mover that imparts linear or reciprocating motion to the conductor. Most
of the commercial prime movers provide rotary motion in the commercial generators. The
conductors of most commercial generators are rotated about a central axis of a shaft. The
conductors are housed in slots cut in a cylindrical structure (made of magnetic material)
known as the armature. The armature is supported at both ends by means of bearings
attached to the shaft that goes through the center of the armature. The armature is rotated
inside the field structure by providing a small gap between these two members. This gap is
known as the air gap and is usually of the order of 1 to 1.5 cms. If the air gap is maintained
constant throughout the spread of the pole arc, we have a fairly constant flux density under
it in a plane perpendicular to the plane of the conductor’s motion. i.e. in a radial direction
with respect to the field and armature structure. Since the emf is also proportional to B,
the flux density in the air gap of AC generators is arranged to be distributed as closely to
a sine wave as possible by suitable shaping (chamfering as it is technically known) of the
pole shoe. Since the relative motion between the conductors and the magnetic flux lines is
responsible for the production of emf, it is immaterial whether the conductors are rotated
or the magnetic flux producing poles are rotated. In most of the alternators it is the field
that is rotated rather than the conductors. In an alternator the external connection to the
load can be taken directly from the conductors since there is no need for any rectification
as in a DC generator. In a DC generator the rectification of the emf is achieved through a
mechanical rectifier—- the commutator and brush arrangement. Moreover the load current
supplied by the alternator can be easily supplied from stationary coils without any difficulty
as there will be no sparking and wear and tear of the brushes and slip rings. Where as the
low values of D.C excitation current to the field coils can be easily sent through the slip
rings and brush arrangement. Thus the usual arrangement in an elementary synchronous
generator is as shown in Fig. 6. The conductors are housed in slots cut in the armature
structure. Only a single coil of N turns, indicated in its cross-section by the two coil sides
a and -a placed in diametrically opposite slots on the inner periphery of the stator (i.e. the
armature, as it is a stationary member here) is shown in Fig. 6.
The conductors forming these coil sides are parallel to the shaft of the machine and
are connected in series by end connections (not shown in the figure ). The coils are actually
formed by taking a continuous copper wire of suitable cross section round a diamond shaped
bobbin. The completed coil is shown in Fig. 7. The copper wire is usually of fine linen
covered, cotton covered or enamel covered so as to have sufficient insulation between the
conductors of the same coil. The actual layout and interconnection of various coils so as to
obtain the required voltage from the synchronous machine (alternator) is presented in the
following section.
(a) (b)
(c) (d)
10
b b
_ + _
+
SB
S N S N
V
+ _ + _ FB
FA FC
SB
SB SC SA FB FC FA SA SC
(a) (b)
Figure 8: Concentrated three-phase,half-coil wave winding with one slot per phase(one coil
side per slot and instantaneous polarity and phase relation of coils)
11
As the distance between two adjacent corresponding points on the poles is 180 elec-
trical degrees, we can see that the distance between the coil side at the start of A and that
at the start of C must be 120 electrical degrees. Thus, the leading pole tip of a unit north
pole moving to the left in Fig. 8 will induce identical voltages in corresponding coil sides
A, C, and B, respectively, 120 electrical degrees apart. Note that phase B lags phase A by
240 electrical degrees or leads phase A by 120 electrical degrees.Fig. 8(b) is a representation
that is frequently used to depict the windings of the three phases and the phase relationship
between them.
The winding depicted in Fig. 8 is an open winding since both ends of the windings
have been brought out for suitable connections. It is a wave winding since it progresses from
pole to pole. It is a concentrated winding because all the coils of one phase are concentrated
in the same slot under one pole. It is a half-coil winding because there is only one-half of
a coil (one coil side) in each slot. It is a full-pitch winding because the coil sides of one
coil are 180◦ electrical degrees apart i.e., they lie under identical magnetic conditions, but of
opposite polarity under adjacent poles.
Fig. 9, on the other hand shows the coils of a single phase,(A, in this case) distributed
winding distributed over two slots under each pole.
Half-coil (also called single-layer) windings are sometimes used in small induction
motor stators and in the rotors of small wound-rotor induction motors. A cross section
of a half-coil, single-layer winding is shown in Fig. 9(c)(i). Like the dc dynamo armature
windings, most commercial armatures for ac synchronous generators are of the full or whole-
coil two-layer type, shown in cross section at the right in Fig. 9(c)(ii). The whole-coil,
two-layer winding gets its name from the fact that there are two coil sides (one coil) per slot.
Fig. 9(a) shows a single-layer, half-coil lap windings;Fig. 9(b) shows a double-layer, full-coil
lap winding. A cross section of a single layer (half-coil) winding is shown in Fig. 9(c)(i).
Whereas most single-layer windings are full-pitch windings, the two-layer, whole-coil
windings are generally designed on an armature as a chorded or fractional-pitch windings.
This common practice stems from the fact that the primary advantage of the whole-coil
windings is that it permits the use of fractional-pitch coils in order to save copper. As will
12
N S N S
SA FA
(a)
S N S N S
SA FA
(b)
N N
Coil em f Ec
Ε1
β
m f
id ee
il s
co
β/2 coil emf Ε2
Εc
Ε1cosβ/2 Ε1cosβ/2
Figure 11: Fractional-pitch coil - Coil EMF in terms of coil side EMFs for fractional-pitch
coil
In the case of the two-layer winding shown in Fig. 9(b), note that the coil span of
single coil is less than the pole span of 180 electrical degrees. The EMF induced in each coil
side is not in phase, and the resultant coil voltage Ec would be less than the arithmetic sum
14
of the EMF of each coil side, or less than 2E1 . It is obvious that 2E1 must be multiplied by a
factor,kp , that is less than unity, to get the proper value for coil voltage Ec (or Ec = 2E1 kp ).
The pitch factor kp is given by
Ec phasor sum of the EMF of the two coil sides
kp = = (6)
2E1 arithmetic sum of the EMF ′ s of the two coil sides
The pitch factor may be quantified in terms of angles as follows. If we assume that
the induced EMFs of two coils, E1 and E2 , are out of phase with respect to each other by
some angle β as shown in Fig. 11, then the angle between E1 and the resultant coil voltage
Ec is β2 .The resultant coil voltage Ec is from Eqn. 6 and Fig. 11.
β
Ec = 2E1 cos = 2E1 kp . (7)
2
and, therefore,
β
kp = cos (8)
2
The angle β is 1800 minus the number of electrical degrees spanned by the coil, for a short-
pitched coil. For a full pitched coil, therefore, kp = 1 as β = 0.
Since β is the supplementary of the coil span, the pitch factor kp may also be expressed
as
p0
kp = sin (9)
2
As stated earlier there are 180 electrical degrees between the centres of two adjacent
north and south poles. Since 360 electrical degrees represents a full cycle of sinusoidal EMF,
15
we are interested in determining how many sinusoidal cycles are generated in one complete
mechanical rotation, i.e., 360 mechanical degrees for a machine having P poles. The number
of electrical degrees as a function of degrees of mechanical rotation is
Pθ
α= = pθ. (10)
2
where P is the number of poles (always an even integer), p is the number of pole-pairs, and
θ is the number of mechanical degrees of rotation.
Thus, a two-pole machine generates one cycle of sinusoid; a four-pole machine gener-
ates two cycles and so on, in one full revolution of the armature.
The windings shown in Fig. 8 and Fig. 9(b) are called concentrated windings because
all the coil sides of a given phase are concentrated in a single slot under a given pole. For
Fig. 8., in determining the induced ac voltage per phase, it would be necessary only to mul-
tiply the voltage induced in any given coil by the number of series-connected coils in each
phase. This is true for the winding shown in Fig. 8 because the conductors of each coil,
respectively, lie in the same position with respect to the N and S poles as other series coils
in the same phase. Since these individual coil voltages are induced in phase with each other,
they may be added arithmetically. In otherwords, the induced emf per phase is the product
of the emf in one coil and the number of series connected coils in that phase.
Concentrated windings in which all conductors of a given phase per pole are concen-
trated in a single slot, are not commercially used because they have the following disadvan-
tages,
1. They fail to use the entire inner periphery of the stator iron efficiently.
2. They make it necessary to use extremely deep slots where the windings are concen-
trated. This causes an increase in the mmf required to setup the airgap flux.
3. The effect of the second disadvantage is to also increase the armature leakage flux and
the armature reactance.
4. They result in low copper-to-iron ratios by not using the armature iron completely.
16
For the five reasons just given, it is more advantageous to distribute the ar-
mature winding, using more slots and a uniform spacing between slots, than to concentrate
the windings in a few deep slots.
1 2 3 4 5 6 7 8 9 10 11 12
V
V
AS AF
When the slots are distributed around the armature uniformly, the winding
that is inserted is called a distributed winding. A distributed lap winding is shown in Fig. 12.
Note that two coils in phase belt A are displaced by one slot angle (the angular displacement
between two successive slots) with respect to each other. The induced voltages of each of
these coils will be displaced by the same degree to which the slots have been distributed, and
the total voltage induced in any phase will be the phasor sum of the individual coil voltages.
For an armature winding having four coils distributed over say, 2/3 rd of a pole-pitch, in
four slots, the four individual coil side voltages are represented by phasors in Fig. 13 as
displaced by some angle α, the number of electrical degrees between adjacent slots, known
as slot angle. It is 300 for the case of 4 slots per phase belt. Voltages Ec1 , Ec2 , etc., are the
individual coil voltages, and n is the number of coils in a given phase belt, in general.
For a machine using n slots for a phase belt, the belt or distribution factor kd by
which the arithmetic sum of the individual coil voltages must be multiplied in order to yield
the phasor sum is determined by the following method,
Eφ
kd = (11)
nEc
17
As in the case of Eqn. 12., the computation of kd in terms of voltages (either theo-
α
D E
C
B
M
A N
α
O
nα
where
n is the number of slots per pole per phase (s.p.p)
α is the number of electrical degrees between adjacent slots i.e. slot angle
It should be noted from Eqn. 12. that the distribution factor kd for any fixed or given
number of phases is a sole function of the number of distributed slots under a given pole.
As the distribution of coils (slots/pole) increases, the distribution factor kd decreases. It is
not affected by the type of winding, lap or wave, or by the number of turns per coil, etc.
18
It is now possible to derive the computed or expected EMF per phase generated
in a synchronous generator. Let us assume that this generator has an armature winding
consisting of a total number of full pitched concentrated coils C, each coil having a given
number of turns Nc . Then the total number of turns in any given phase of an m-phase
generator armature is
CNc
Np = (12)
m
But Faraday’s law Sec. ?? states that the average voltage induced in a single turn of
two coil sides is
φ
Eav = (13)
t
The voltage induced in one conductor is 2φ/(1/s) = 2φs, where s=speed of rotation
in r.p.s, for a 2 pole generator. Furthermore, when a coil consisting of Nc turns rotates in a
uniform magnetic field, at a uniform speed, the average voltage induced in an armature coil
is
av = 4φN s
E coil c V olts (14)
where φ is the number of lines of flux (in Webers) per pole, Nc is number of turns per coil, s is
the relative speed in revolutions/second (rps) between the coil of Nc turns and the magnetic
field φ.
19
Commercial ac synchronous generators have many poles and may rotate at various
speeds, either as alternators or as synchronous or induction motors.Eqn. 13 was derived for
a two-pole device in which the generated EMF in the stationary armature winding changes
direction every half-revolution of the two-pole rotor. One complete revolution will produce
one complete positive and negative pulse each cycle. The frequency in cycles per second
(Hz) will, as stated previously, depend directly on the speed or number of revolutions per
second (rpm/60) of the rotating field.
If the ac synchronous generator has multiple poles (having, say, two, four, six, or
eight poles...), then for a speed of one revolution per second (1 rpm/60), the frequency
per revolution will be one, two, three, or four ..., cycles per revolution, respectively. The
frequency per revolution, is therefore, equal to the number of pairs of poles. Since the
frequency depends directly on the speed (rpm/60) and also on the number of pairs of poles
(P/2), we may combine these into a single equation in which
P rpm PN P ωm ∗ 60 P ωm ωe
f= ∗ = = ∗ = ∗ = (18)
2 60 120 120 2π 2 2π 2π
where
P is the number of poles
N is the speed in rpm (rev/min)
f is. the frequency in hertz
ωm is the speed in radians per second (rad/s)
ωe is the speed electrical radians per second.
As stated earlier the field windings are provided in the rotor or the rotating member
of the synchronous machine. Basically there are two general classifications for large 3 phase
synchronous generators ——cylindrical rotor and salient-pole rotor - .
20
the smoothness of the rotor contour makes for reduced windage losses and for quiet operation.
Salient-pole rotors are used in low-speed synchronous generators such as those driven
by water wheels. They are also used in synchronous motors. Because of their low speeds
salient-pole generators require a large number of poles as, for example, 60 poles for a 100-rpm
50 Hz generator.
Fig. 14 illustrates two and four pole cylindrical rotors along with a developed
view of the field winding for one pair of poles. One pole and its associated field coil of
a salient-pole rotor is shown in fig. 14.The stator slots in which the armature winding is
embedded are not shown for reasons of simplicity. The approximate path taken by the field
flux, not including leakage flux, is indicated by the dashed lines in Fig. 14. The field coils in
Fig. 14 are represented by filaments but actually (except for the insulation between turns and
between the coil sides and the slot) practically fill the slot more nearly in keeping with fig. 15.
The stepped curve in fig. 15. represents the waveform of the mmf produced by the
distributed field winding if the slots are assumed to be completely filled by the copper in the
coil sides instead of containing current filaments. The sinusoid indicated by the dashed line
in fig. 15 represents approximately the fundamental component of the mmf wave.
The air gap in cylindrical-rotor machines is practically of uniform length except for
the slots in the rotor and in the stator, and when the effect of the slots and the tangential
component of H, which is quite small for the low ratio of air-gap length to the arc subtended
by one pole in conventional machines, are neglected, the stepped mmf wave in fig. 15 produces
a flux-density space wave in which the corners of the steps are rounded due to fringing. The
flux density wave form is therefore more nearly sinusoidal than the mmf waveform when the
effect of the slots is neglected. However, saturation of the iron in the region of maximum
mmf tends to flatten the top of the flux-density wave.
21
d ax is
v v
v
v V
V
v N
i1 V V
q ax is
v
v
V
f’ f
V
v v
V
S S
V
V
S
V V
V
N V
(a)
(a) (b)
V
d-axis
V
V
q-axis
V
(c) (d)
Figure 14: Synchronous machines with stator slots and armature windings omitted (a)Two-
pole cylindrical rotor, (b) Four-pole cylindrical rotor, (c) Developed view of two pole cylin-
drical rotor field structure, (d) Salient pole and field coil
22
d axis
mmf wave
fundamental
q axis component
Airgap
Stator iron
Rotor
iron
Figure 15: Cylindrical rotor mmf wave and its fundamental of a synchronous machine
The brushless excitation system eliminates the usual commutator, collector rings, and
brushes. One arrangement in which a permanent magnet pilot exciter, an ac main exciter,
and a rotating rectifier are mounted on the same shaft as the field of the ac turbogenerator is
shown in Fig. 17. The permanent magnet pilot excitor has a stationary armature and a ro-
tating permanent magnetic field. It feeds 400 Hz, three-phase power to a regulator, which in
turn supplies regulated dc power to the stationary field of a rotating-armature ac exciter, The
23
(a)
Synchronous machine
Exciter-field
rheostat
Rotor
v slip
rings
v
v
v
v
Exciter Three-phase
field armature
Stator
Rotor
Brushes
Exciter(dc generator
(b)
Pilot Synchronous
Main
exciter Prime mover
exciter generator
(c)
Synchronous generator
Pilot exciter Main exciter
v v
Sh.f Ser.f
24
(d)
ac exciter
field ac turbine
Pilot exciter generator stator
armature
Stationary
components
Regulator
output of the ac exciter is rectified by diodes and delivered to the field of the turbo generator.
Brush less excitation systems have been also used extensively in the much smaller
generators employed in aircraft applications where reduced atmospheric pressure intensifies
problems of brush deterioration. Because of their mechanical simplicity, such systems lend
themselves to military and other applications that involve moderate amounts of power.
25
Flux produced by
armature current
Flux
produced Direction
by of rotation
main field
N S
δ
I
(b)Phasor diagram
26
Flux produced by
armature current
Flux
produced
by Direction
main field of rotation
N S
(a)The effect of armature current when the machine operates as a motor at u.p.f
Eo
E
(b)Phasor diagram
27
Flux produced by
armature current
Flux
produced
by
main field
Direction
of rotation
N S N
Eo
I
(b)Phasor diagram
28
Flux produced by
armature current
Flux
produced
by
main field
S S Direction
N of rotation
Eo
(b)Phasor diagram
29
rying armature windings, and a rotor carrying field windings and capable of rotation within
it. The directions of the currents and the flux distribution are as shown in Fig. 18(a), when
the emf induced in the stator coils is the maximum. The coil links no resultant flux but is
in the position of greatest rate of change of flux. The coil position shown is also that for
maximum current when the current is in phase with the voltage: i.e for a pure resistive load.
The current in the coil has no effect on the total flux per pole, but causes a strengthening
on one side and a weakening on the other side of the pole shoes. Thus the armature con-
ductors find themselves in the circumstances illustrated in Fig. 19, and a torque is produced
by the interaction of the main flux φm with the current in the conductors. The torque thus
produced is seen to be opposed to the direction of motion of the rotor - the force on the
conductors is such as to push them to the left and by reaction to push the rotor to the right
(as the armature coils are stationary). The rotor is rotated by a prime mover against this
reaction, so that the electrical power, the product EI, is produced by virtue of the supply
of a corresponding mechanical power. Thus it is evident from the distortion of the main
flux distribution that electrical energy is converted from mechanical energy and the machine
operates as a generator. An unidirectional torque is maintained as the stator conductors
cut N-Pole and S-Pole fluxes alternately resulting in alternating emfs at a frequency equal
to the number of pole-pairs passed per second and the currents also alternate with the emf.
The assumption that the conditions shown in Fig. 18(a) represent co-phasal emf and current
is not quite true. The strengthening of the resultant flux on the right of the poles and an
equivalent amount of weakening on the left effectively shift the main field flux axis against
the direction of rotation, so that the actual e.m.f. E induced in the armature winding is an
angle δ behind the position E0 that it would occupy if the flux were undistorted as shown
in the adjacent phasor diagram Fig. 18(b) pertaining to this condition of operation. Thus
the effect of a resistive (unit power factor (upf)) load connected to a synchronous generator
is to shift the main field flux axis due to what is known as cross-magnetization.
30
Next let us consider this generator to be connected to a purely inductive load so that
the current I in the coils lags behind the e.m.f. E by 90 electrical degrees i.e. corresponding
to a quarter-period, in time scale. Since the coil-position in Fig. 18(a) or Fig. 19(a) represents
that for maximum e.m.f., the poles would have moved through half a pole-pitch before the
current in the coil has reached a maximum as shown in Fig. 20(a). As seen from this figure it
is obvious that the ampere-turns of the stator coils are now in direct opposition to those on
the pole, thereby reducing the total flux and e.m.f. Since the stator and rotor ampere-turns
act in the same direction, there is no flux-distortion, no torque, and hence no additional
mechanical power. This circumstance is in accordance with the fact that there is also no
electrical power output as E and I are in phase quadrature, as shown in Fig. 20(b). The
phasor Eo represents the e,m.f. with no demagnetizing armature current, emphasizing the
reduction in e.m.f. due to the reduced flux.
Likewise, when this generator is connected to a purely capacitive load i.e the current I
in the coil leads the emf E by 90 electrical degrees, the conditions are such that the armature
AT and the field AT will be assisting each other as shown in Fig. 21.
When the generator supplies a load at any other power factor intermediate between
unity and zero, a combination of cross- and direct-magnetization is produced on the magnetic
circuit by the armature current. The cross-magnetization is distorting and torque-producing
as in Fig. 18; the direct-magnetization decreases (for lagging currents) or increases (for lead-
ing currents) the ampere-turns acting on the magnetic circuit as in Fig. 20 and Fig. 21,
affecting the main flux and the e.m.f. accordingly.
For a motor the torque is reversed on account of the current reversal, and the direct-
magnetizing effect is assisting the field ampere-turns for lagging currents. The action of
the armature ampere-turns as described above is called armature-reaction. The effect of
the armature reaction has a far-reaching influence on the performance of the synchronous
motor, particularly as regards the power factor at which it operates and the amount of field
excitation that it requires.
31
output. As the stator currents change direction in the same time as they come from one
magnetic polarity to the next, the torque is unidirectional. The torque of individual phases
is pulsating just like in a single-phase induction machine - but the torque of a three-phase
machine is constant for balanced loads.
For the cylindrical rotor machine the fundamental armature reaction can be more
Ft
C’ A
po
B B’
le
ax
is
Fa
A’ C
32
mmf of
main field Fa
Pole axis
Fad
Faq
Pole axis
mmf of A-A’
Figure 24: Elementary synchronous motor action - Attraction of the unlike poles keep the
rotor locked to the rotating field produced in the stator
33
It is to be noted that the magnetic field shown in Fig. 24 is only diagrammatic and
for better understanding of the action of the synchronous machine - the flux lines may be
considered as elastic bands which will be stretched by application of the mechanical load
on the shaft. Actually the flux lines will enter or leave the stator and rotor surfaces nearly
normally, on account of the high permeability of these members. In a salient-pole machine
the torque is developed chiefly on the sides of the poles and on the sides of the teeth in a
non-salient-pole machine.
34
1. The permeability of all parts of the magnetic circuit of the synchronous machine is
constant - in other words the field and armature fluxes can be treated separately as
proportional to their respective currents so that their effects can be superposed.
2. The air gap is uniform, so that the armature flux is not affected by its position relative
to the poles - in other words we assume the rotor to be cylindrical
4. The armature winding is uniformly distributed and carries balanced sinusoidal currents.
In other words, the harmonics are neglected so that the armature flux is directly
proportional to the fundamental component of the armature reaction mmf implying
that the armature reaction mmf is distributed sinusoidally and rotates at synchronous
speed with constant magnitude.
Assumption (1) is roughly fulfilled when the machine works at low saturation; (2)
and (3) are obviously inaccurate with salient-pole machines and assumption (4) is com-
monly made and introduces negligible error in most cases. The behaviour of an “ideal”
synchronous machine can be indicated qualitatively when the above assumptions (1) to (4)
are made.
The phasor diagrams Fig. 25 for the several conditions contain the phasors of two
emfs viz. Eo and E . The latter is the e.m.f actually existing, while the former is that which
would be induced under no-load conditions, i.e. with no armature current (or armature
reaction).
35
Thus Eo is the e.m.f. corresponding to the flux produced by the field winding only,
while E is that actually produced by the resultant flux due to the combined effect of stator
and rotor ampere-turns. The actual e.m.f. E can be considered as Eo plus a fictitious e.m.f.
proportional to the armature current.
Fig. 25 is drawn in this manner with Er such that the following phasor rela-
tionship is satisfied:
E = Eo + Er (19)
It can be seen from Fig. 25, that Er , is always in phase-quadrature with armature current
and proportional to it (as per the four assumptions (1) to (4) above). The emf Er is thus
similar to an emf induced in an inductive reactance, so that the effect of armature reaction
is exactly the same as if the armature windings had a reactance xa = Er /Ia . This fictitious
reactance xa can added to the armature leakage reactance xl and the combined reactance
( xa +xl ) is known as the synchronous reactance xs . The armature winding apart from these
reactance effects, presents a resistive behaviour also. Synchronous impedance is a tern used
to denote the net impedance presented by each phase of the alternator winding, consisting
of both resistive and reactive components. The behavior of a synchronous machine can be
easily predicted from the equivalent circuit developed using this synchronous reactance xs ,
as explained in the following section.
Economical size requires the magnetic circuit to be somewhat saturated under normal
operating conditions. However, the machine is unsaturated in the short-circuit test, and
the synchronous reactance based on short-circuits and open-circuit test data is only an
approximation at best. Nevertheless, there are many studies in which a value based on
rated open-circuit voltage and the short circuit current suffices. Hence, in Fig. 29, if oc is
rated voltage, ob is the required no-load field current, which also produces the armature
current o′ e on short circuit. The synchronous impedance assuming the armature winding is
star-connected is, accordingly,
oc
Zs = √ (20)
3 ∗ o′ e
Except in very small machines, the synchronous reactance is much greater than the
resistance (ra ) of the armature and the saturated value as well as the unsaturated value of the
synchronous reactance and therefore is considered equal to the magnitude of the synchronous
impedance
1
Xd = (Zs2 − ra2 ) 2 ≈ Zs (21)
36
Er
E Eo
Er
Eo E
I Er
Er
V
Eo Er Er E
E Eo
I Er
Eo
I
Er
(c) Generator (d)Generator zero power factor
37
The line of in Fig. 29 is more nearly representative of the saturated machine than is
the air-gap line. On the basis of this line, an estimate of the field current can be obtained
for a given terminal voltage, load current, and power factor. This is done by calculating Eaf
and making use of the saturated synchronous reactance as follows.
Eaf = V + Zs I (22)
The field current is that required to produce Eaf on the line of.
In the case of a constant voltage source having constant impedance, the impedance
can be found by dividing the open-circuit terminal voltage by the short circuit current.
However, when the impedance is a function of the open-circuit voltage, as it is when the
machine is saturated, the open-circuit characteristic or magnetization curve in addition to
the short-circuit characteristic is required.
Zs Zs
+ + +
Eaf ~ ~ Isc
Eoc=Eaf Eaf
38
of the unsaturated iron is negligible. The equivalent circuit of one phase of a polyphase
synchronous machine is shown in Fig. 26 for the open-circuit condition and for the short
circuit condition. Now Eaf is the same in both cases when the impedance Zs . Where Eaf is
the open-circuit volts per phase and Isc is the short-circuit current per phase.
1.0 1.0
Short-circuit current.A
Field current A
Field current,A (b)
(a) (b)
Figure 27: (a) Open circuit characteristic and (b) Short-circuit characteristic
39
if there were no saturation. On the basis of this convention, the per-unit representation is
such as to make the air-gap lines of all synchronous machines identical.
Ia
dc source Ib
If Ic
The three terminals of the armature are short -circuited each through a current-
measuring circuit, which except for small machines is an instrument current transformer with
an ammeter in its secondary. A diagram of connections in which the current transformers
are omitted is shown in Fig. 28.
The machine is driven at approximately synchronous (rated) speed and measure-
ments of armature short-circuit current are made for various values of field current, usually
up to and somewhat above rated armature current. The short-circuit characteristic (i.e.
armature short circuit current versus field current) is shown in Fig. 27. In conventional
synchronous machines the short-circuit characteristic is practically linear because the iron
is unsaturated up to rated armature current and somewhat beyond, because the magnetic
axes of the armature and the field practically coincide (if the armature had zero resistance
the magnetic axes would be in exact alignment), and the field and armature mmfs oppose
each other.
The open circuit and short-circuit characteristics are represented on the same graph
in Fig. 29. The field current oa produces a line-to line voltage oc on the air- gap line, which
40
41
would be the open-circuit voltage if there were no saturation. The same value of field current
produces the armature current o’d and the unsaturated synchronous reactance is given by:
oc
Xd = √ Ω phase, for a star connected armature (23)
3 o′ d
When the open-circuit characteristic, air-gap line, and the short-circuit characteristic
are plotted in per-unit, then the per unit value of unsaturated synchronous reactance equals
the per-unit voltage on the air-gap line which results from the same value of field current
as that which produces rated short-circuit (one-per unit) armature current. In Fig. 29 this
would be the per-unit value on the air gap line corresponding to the field current og.
42
Ixa
xa xl ra Ixl
I Et E
Ira
V
V Load A
Et E
I
σ
xs ra I Et
IZs IXs
Zs V
Et V Load
φ Ι
43
(Xl + Xa ) = Xs all per phase. This is shown in Fig. 30. The principal characteristics of the
synchronous generator will be obtained qualitatively from this circuit.
Consider a synchronous generator driven at constant speed and with constant exci-
tation. On open circuit the terminal voltage V is the same as the open circuit e.m.f. Et .
Suppose a unity-power-factor load be connected to the machine. The flow of load current
produces a voltage drop IZs in the synchronous impedance, and terminal voltage V is re-
duced. Fig. 31 shows the complexor diagram for three types of load. It will be seen that
the angle σ between Et and V increases with load, indicating a shift of the flux across the
pole faces due to cross- magnetization. The terminal voltage is obtained from the complex
summation
V + Zs = Et
or V = Et − IZs (24)
so that the V/I curve, Fig. 32, is nearly an ellipse with semi-axes Et and Isc . The
current Isc is that which flows when the load resistance is reduced to zero. The voltage V
falls to zero also and the machine is on short-circuit with V = 0 and
For a lagging load of zero power-factor, diagram is given in Fig. 31 The voltage
is given as before and since the resistance in normal machines is small compared with the
synchronous reactance, the voltage is given approximately by
V ≈ Et − IXs (28)
44
IXs
Ixs
Et
Ir Ir v
Et V
Et V
I σ
I σ
σ
I
V1
Ir
Ixs v
Et
Et
Ea
v
Ea
I I
(c) (d)
45
in g
ad
Le ing
0.0 Lead
0.8
0.9 Leading
100 1.0
0.9
0. Lagg
Isc
0
100
Perfect of full -load current
which is the straight line marked for cos φ = 0 lagging in Fig. 32. A leading load of
zero power factor Fig. 31. will have the voltage
V ≈ Et + IXs (29)
another straight line for which, by reason of the direct magnetizing effect of leading
currents, the voltage increases with load.
46
Comparing the voltages on full load (1.0 per unit normal current) in Fig. 32, it will
be seen that much depends on the power factor of the load. For unity and lagging power
factors there is always a voltage drop with increase of load, but for a certain leading power
factor the full-load regulation is zero, i.e. the terminal voltage is the same for both full and
no-load conditions. At lower leading power factors the voltage rises with increase of load,
and the regulation is negative. From Fig. 30, the regulation for a load current I at power
factor cos φ is obtained from the equality
from which the regulation is calculated, when both Et and V are known or found.
200
g
in
gg g
La i n
0 g
0. La
g
0.8 ing
gg
Percent of no -load field excitation
La f
0. 9 up
in g
ad
100 Le
0.9
0.8 Leading
0.0
Le
ad
in g
0
0 100
Percent of full -load current per phase
47
Low-leading-power-factor loads, on the other hand, will require the excitation to be reduced
on account of the direct magnetizing effect of the zero- power-factor component. Fig. 33
shows typical e.m.f./current curves for a constant output voltage. The ordinates of Fig. 33
are marked in percentage of no-load field excitation, to which the e.m.f Et exactly corresponds
when saturation is neglected.
For any load conditions as represented by Fig. 30, the output per phase is
P = V I cos φ. The electrical power converted from mechanical power input is per phase
P1 = Et I cos(φ + σ) (32)
Resolving Et along I
The electrical input is thus the output plus the I 2 R loss, as might be expected. The
prime mover must naturally supply also the friction, windage and core losses, which do not
appear in the phasor diagram.
In large machines the resistance is small compared with the synchronous reactance
so that θ = arc tan(xs /r) ≈ 90◦ , it can be shown that
V Zs
= (34)
sin(90 − θ + σ)2 sin σ
and hence,
P = P1 = Et I cos(φ + σ) ≈ (Et /Xs ).V sin σ (35)
An obvious deduction from the above Eqn. 35is that the greater the field excitation
(corresponding to Et ) the greater is the output per unit angle σ: that is, the more stable
will be the operation.
48
As discussed earlier in Sec. 3.1 the behaviour of a synchronous machine on load can
be determined by the use of synchronous reactance xs which is nothing but the sum of xa
and xl , where xa is a fictitious reactance representing the effect of armature reaction while
xl is the leakage reactance. It was also mentioned that this method of representing the
effect of armature reaction by a fictitious reactance xa was applicable more aptly only for a
cylindrical rotor (non-salient pole) machine. This was so as the procedure followed therein
was valid only when both the armature and main field m.m.f.’s act upon the same magnetic
circuit and saturation effects are absent.
It was shown in Sec. ?? that the effect of armature reaction in the case of a salient
pole synchronous machine can be taken as two components - one acting along the direct
axis (coinciding with the main field pole axis) and the other acting along the quadrature
axis (inter-polar region or magnetic neutral axis) - and as such the mmf components of
armature-reaction in a salient-pole machine cannot be considered as acting on the same
magnetic circuit. Hence the effect of the armature reaction cannot be taken into account by
considering only the synchronous reactance, in the case of a salient pole synchronous machine.
In fact, the direct-axis component Fad acts over a magnetic circuit identical with
that of the main field system and produces a comparable effect while the quadrature-axis
component Faq acts along the interpolar space, resulting in an altogether smaller effect and,
in addition, a flux distribution totally different from that of Fad or the main field m.m.f.
This explains why the application of cylindrical-rotor theory to salient-pole machines for
predicting the performance gives results not conforming to the performance obtained from
an actual test.
Blondel’s two-reaction theory considers the effects of the quadrature and direct-axis
components of the armature reaction separately. Neglecting saturation, their different effects
are considered by assigning to each an appropriate value of armature-reaction “reactance,”
respectively xad and xaq . The effects of armature resistance and true leakage reactance
( xl ) may be treated separately, or may be added to the armature reaction coefficients on the
assumption that they are the same, for either the direct-axis or quadrature-axis components
of the armature current (which is almost true). Thus the combined reactance values can be
expressed as :
49
It is essential to clearly note the difference between the quadrature- and direct-axis
components Iaq , and Iad of the armature current Ia , and the reactive and active components
Iaa and Iar . Although both pairs are represented by phasors in phase quadrature, the former
are related to the induced emf Et while the latter are referred to the terminal voltage V .
These phasors are clearly indicated with reference to the phasor diagram of a (salient pole)
synchronous generator supplying a lagging power factor (pf) load, shown in Fig. ??(a). We
have
q
Iaq = Ia cos(δ + φ); Iad = Ia sin(δ + φ); and Ia = (Iaq2 + I2 ) (37)
ad
p
Iaa = Ia cos(φ); Iar = Ia sin(φ); and Ia = (Iaa 2 + I2 )
ar (38)
where σ = torque or power angle and φ = the p.f. angle of the load.
The phasor diagram Fig. 34 shows the two reactance voltage components Iaq ∗ xsq and
Iad ∗ xsd which are in quadrature with their respective components of the armature current.
The resistance drop Ia ∗ Ra is added in phase with Ia although we could take it as Iaq ∗ Ra
and Iad ∗ Ra separately, which is unnecessary as
Ia = Iad + jIaq
Actually it is not possible to straight-away draw this phasor diagram as the power angle σ is
unknown until the two reactance voltage components Iaq ∗ xsq and Iad ∗ xsd are known. How-
ever this difficulty can be easily overcome by following the simple geometrical construction
shown in Fig. 34(d), assuming that the values for terminal voltage V , the load power factor
(pf) angle φ and the two synchronous reactances xsd and xsq are known to us.
The resistance drop Ia ∗ Ra (length AB) is added to the tip of the voltage phasor
(OA) in phase with the current phasor (i.e. in a direction parallel to OQ ). Then we draw
50
line BC ( of length equal to Ia ∗ xsq ) and extend it to D such that BD will be (of length
equal to Ia ∗ xsd ) at the extremity B of Ia ∗ Ra and at right-angles to Ia . Draw OC and
extend it (to F ). From D draw the perpendicular DF on OC extended. Then OF represents
the induced voltage Et . The proof for this can be given as follows:. If DF is extended to G
such that this line is perpendicular to BG drawn parallel to OF, we have :
Neglecting the armature winding resistance, the power output of the generator is
given by:
P = V ∗ Ia ∗ cos φ (41)
It is clear from the above expression that the power is a little more than that for
a cylindrical rotor synchronous machine, as the first term alone represents the power for a
cylindrical rotor synchronous machine. A term in (sin 2σ) is added into the power - angle
characteristic of a non-salient pole synchronous machine. This also shows that it is possible
to generate an emf even if the excitation E0 is zero. However this magnitude is quite less
compared with that obtained with a finite E0 . Likewise we can show that the machine
develops a torque - called the reluctance torque - as this torque is developed due to the
variation of the reluctance in the magnetic circuit even if the excitation E0 is zero.
51
(a) Components of Ia
Eo
Iaq
σ Iaa
φ Iad V Eo
Iaqxsq
Ia Iar
(b) Phasor addition of
component drops
Iaq Iadxsd
σ Iara
φ
Iad V
Ia
Iaxsd Iaqxsq
(c) Phasor addition using Ia
Iaxsq
Iadxsd
σ Iara
φ D
V
Ia
C F
(d) Geometric construction E
of (c)
+ φ) G
(9 0-σ Η
σ A
φ V
o B
I
Q
52
the synchronous machine is driven by means of a prime mover (usually a DC motor in the
laboratory) at a speed close to the synchronous speed in the proper direction but not equal
to it. The armature is supplied with a low voltage 3-Phase balanced supply through a variac,
while the field circuit is kept open. The armature current varies between two limits since
it moves through, since the synchronously rotating armature mmf acts through the varying
magnetic reluctance paths as it goes from inter-polar axis to pole axis region. The values of
xsd and xsq are determined based on the applied voltage and the armature current values.
The ratio of applied voltage to the minimum value of the armature current gives the direct
axis synchronous reactance xsd which is usually the same as the synchronous reactance
xs that we usually determine from normal no-load and short-circuit tests as explained in
Sec. ?? The ratio of applied voltage to the maximum value of the armature current gives
the the quadrature-axis reactance xsq . For more accurate determination of these values the
oscillogram of the armature current and voltage can be recorded.
4. Load loss due to armature leakage flux causing eddy current and hysteresis losses in
the armature-surrounding iron.
1. The rotational losses, which include friction and windage losses, are constant, since the
speed of a synchronous generator is constant. It may be determined from a no-load
test.
2. The core loss includes eddy current and hysteresis losses as a result of normal flux
density changes. It can be determined by measuring the power input to an auxiliary
motor used to drive the generator at no load, with and without the field excited. The
difference in power measured constitutes this loss.
53
3. The armature and field copper losses are obtained as Ia2 Ra and Vf If Since per phase
quantities are dealt with, the armature copper loss for the generator must be multiplied
by the number of phases. The field winding loss is as a result of the excitation current
flowing through the resistance of the field winding.
4. Load loss or stray losses result from eddy currents in the armature conductors and
increased core losses due to distorted magnetic fields. Although it is possible to separate
this loss by tests, in calculating the efficiency, it may be accounted for by taking the
effective armature resistance rather than the dc resistance.
After all the foregoing losses have been determined, the efficiency η is calculated as,
kV A ∗ P F
η= ∗ 100% (44)
kV A ∗ P F + (total losses)
where η = efficiency,
The quantity (kVA*PF) is, of course, the real power delivered to the load (in kW) by
the synchronous generator. Thus, it could in general be stated as
Pout Pout
η= ∗ 100 = ∗ 100 (45)
Pin Pout + Plosses
The input power Pin = Pout + Plosses is the power required from the prime mover to
drive the loaded generator.
54
The development of synchronizing power depends on the fact that the armature
impedance is preponderating reactive. If it were not, the machines could not operate stably
in parallel: for the circulating current Is would be almost in phase- quandrature with the
generated e.m.f.’s, and would not contribute any power to slow the faster or speed up the
slower machine.
When both machines are equally loaded pn to an external circuit, the synchronizing
power is developed in the same way as on no load, the effect being to reduce the load of the
slower machine at the same time as that of the faster machine is increased. The conditions
are shown in Fig. 35, where I1 , I2 are the equal load currents of the two machines before the
′ ′
occurrence of phase displacement, and I1 , I2 are the currents as changed by the circulation
of the synchronizing current Is .
55
Load
V
v I vI
E1
2δ
Is
v
Zs1 Zs2 ls
v v
I1 θ
I2
Es
1 E1 E2 2
E2
Local circuit
(a) (b)
Es Es
I’1
E1
E2
E2 Is
2δ E1
δ δ I1.I2
Is
Is
Is
I’2
(c) (d)
56
104
100
98
0 25 50 75 100 125
Percent of full-load output power
general, the machines will have different synchronous impedance Zs1 , Zs2 ; different e.m.f.’s E1
and E2 and different speed regulations. The governors of prime movers are usually arranged
so that a reduction of the speed of the prime mover is necessary for the increase of the power
developed. Unless the governor speed/load characteristics are identical the machines can
never share the total load in accordance with their ratings. The governor characteristics
take the form shown in Fig. 36. If the two are not the same, the load will be shared in
accordance with the relative load values at the running speed, for synchronous machines
must necessarily run at identical speeds.
57
1. Since the demand of electricity varies during a day, also during the various seasons in a
year, a modern power station employs two or more units so that one or more alternators
can supply power efficiently according to the need. Installation of a single generator
of capacity equal to the installed capacity of a station will be uneconomic, as such a
generator will have to be run at a reduced load for certain periods of the day, and
also building of such a generator is difficult proposition. Further, routine maintenance
requires a unit to be shut down for a certain period of time and as such the capacity
requirement of the stand by unit in a power station with several alternators is less.
2. Connections of several stations by a grid is economic and advantageous. This reduces
the installed capacity of the stand by unit considerably, and enables economic distribu-
tions of load between several stations. Also, in a country like India, where considerable
amount of power is generated by harnessing waterpower,parallel operation of steam
and hydro-stations is essential to maintain continuity of supply throughout the year
and also to ensure the maximum utilization of water power resources.
For alternators in parallel, change in field excitation will mainly change the operating
power factor of the generator and has primarily no effect on the active power delivered by
the generators (change in power factor will change the total current of an alternator thereby
changing copper loss. The output active power will alter through a very small amount). The
control of active power shared between alternators is affected by changing the input power
to the prime mover. For example, in a thermal power station: having alternators driven
by steam turbines, an increase of throttle opening and thus allowing more steam into the
turbine will increase the power input; in a hydro station, the power input is controlled by
water inlet into the turbine. The prime-mover, speed-load characteristics thus determine the
load sharing between the alternators.
Consider for simplicity, a two machine case, consisting of two non-salient pole syn-
chronous machines (generators) 1 and 2 respectively coupled to prime-movers 1 and 2
Fig. 37 shows the speed-load characteristics of the prime- movers. Assume that initially the
two generators share equal active power and it is now required to transfer a certain amount
of power from unit 1 to unit 2, the total power remaining constant.
58
SPEED
c a b
2’
INPUT
INCREASED
d 1
BUS 2 P’L2 P’L1 INPUT
REDUCED
1’
ALTERNATOR PL2 PL1
~ ~ 0
PRIME MOVER
1 2
TOTAL ACTIVE POWER
LOAD ON LOAD ON
UNIT 1 UNIT 2 MACHINE 2 2PL MACHINE 1
(a) (b)
The initial operating points are indicated on the characteristic by points b and c, the
busbar speed (or frequency) being given by the point a. The load on each machine is PL .
the total load being 2PL . To reduce the load on unit 1, its input is decreased (by reducing
′
the throttle opening) so that the prime-mover characteristic is now given by 1 . The total
load being constant, the loads shared by the machines are
machine 1 → PL1 ,
machine 2 → PL2 ,
the total load being PL1 + PL2 = 2PL , and the bus frequency given by the point d is reduced.
To maintain the bus frequency constant at its original value (given by point a) the input to
′
unit 2 must be suitably increased so that its speed-load characteristic is given by 2 . The
final load sharing is thus given by
′ ′
machine 1→ PL1 , machine 2 → PL2
and
′ ′
PL1 + PL2 = 2PL (46)
59
For any load conditions represented e.g. by Fig. 38 the output per phase is P = V I cos φ.
The electrical power converted from mechanical power input is per phase
P1 = EI cos(φ + σ) (47)
Resolving E along I,
P1 = EI cos(φ + δ) = V I cos φ + Ir.I = V I cos φ + I 2 r (48)
The electrical input is thus the output plus the I 2 R loss, as might be expected. The prime
mover must naturally supply also the friction, windage and core losses, which do not appear
in the complexor diagram.
For a given load current I at external phase-angle φ to V , the magnitude and phase of E
are determined by Zs The impedance angle θ is arc tan(x3 /r), and using Fig. 38.
I = (E − V )∠zs = (E∠δ − V /0)/zs /∠θ (49)
= (E/zs )∠(δ − θ) − (V /zs )∠θ
when referred to the datum direction V = V ∠θ Converting to the rectangular form:
In large machines the resistance is small compared with the synchronous reactance
so that θ = arc tan(xs /r) ≃ 90◦ .Eqn. 50 and Eqn. 52 the simplify to P1 = P , where
P = P1 = E cos(θ + δ) ≃ (E/Xs )V sin δ (53)
60
v
Izs v
Ixs
δ V θ
v
v
Datum
φ Ir
I
Figure 38: Power conditions
Thus the power developed by a synchronous machine with given values of E, V and
Zs , is proportional to sin δ or, for small angles, to δ itself. The displacement angle δ rep-
resents the change in relative position between the rotor and resultant pole-axes and sis
proportional to the load power. The term load-, power- or torque-angle may be applied to
δ.
An obvious deduction from Eqn. 53 is that the greater the field excitation (corresponding
to E), the greater is the output per unit angle δ; that is, the more stable will be the operation.
61
we may presume that the generating plant maintains an invariable voltage and frequency
at all points.In other words a network has zero impedance and infinite rotational inertia. A
synchronous machine connected to such a network is said to be operating on infinite bus-bars.
Practically all synchronous motors and generators in normal industrial use on large
power supply systems can be considered as connected to infinite bus-bars, the former because
they are relatively small, the latter on account of the modern automatic voltage regulators
for keeping the voltage practical, constant at all loads. The behaviour of the synchronous
machine connected to infinite bus bars can be easily described from the electrical load dia-
gram of a synchronous generator.
Consider a synchronous machine connected to infinite bus bars (of constant-voltage, constant-
frequency) of phase voltage V, Fig. 39. Let the machine run on no load with mechanical and
core losses only supplied. If the e.m.f. E be adjusted to equality with V , no current will flow
into or out of the armature on account of the exact balance between the e.m.f. and the bus-
bar voltage. This will be the case when a synchronous generator is just parallel to infinite
bus bar. If the excitation,If is reduced, machine E will tend to be less than V , so that a
leading current Ir will flow which will add to the field ampere-turns due to direct magnetiz-
ing effect of armature reaction. Under the assumption of constant synchronous impedance,
this is taken into account by Ir Zs as the difference between E and V . The current Ir must
be completely reactive because the machine is on no-load and no electrical power is being
supplied to or by the machine, as it is on no-load. If now the excitation be increased, E will
tend to be greater than V . A current will therefore be circulated in the armature circuit,
this time a lagging current which will reduce the net excitation due to the demagnetizing
effect of armature reaction so that the machine will again generates a voltage equal to that
62
of the constant bus-bar voltage. The synchronous impedance drop Ir Zs is, as before, the
difference between E and V , and there should be only a zero-power-factor lagging current,
as the machine is running on no-load.
s of E
Irzs
L o cu
E
E V V V
Irzs
E
Ir Ir
Under Over
Normal excited
excited
Suppose the machine to be supplied with full-load mechanical power. Then as a gen-
erator it must produce the equivalent in electrical power: i.e. the output current must have
an active component Iaa corresponding to full-load electrical power. For an output at exactly
unity power factor, the excitation must be adjusted so that the voltage triangle E, V, Ia Zs ,
satisfies the conditions required, Fig. 40. If the excitation be reduced, a magnetizing reactive
component is supplied in addition, i.e. a leading current Iar , which assists the field winding
to produce the necessary flux. If the machine is over excited, a lagging reactive demagnetiz-
ing current component is supplied, in addition to the constant power component.
In Fig. 40 the IZs drop has been added in components corresponding to the current
components Iaa and Iar . For all three diagrams of Fig. 40, Iaa and Iaa Zs are constant, since
the electrical power supplied is constant. Only the component Iaa Zs (and therefore Iar )
varies with the excitation. Thus the excitation controls only the power factor of the current
63
t
s of E
Et
Locu
Irzs
V Iazs V I a zs V
Iazs
Et Irzs
Et Ir Ir
Ia Ia Ia
I I
64
supplied by the generator to the infinite bus-bars and not the active power.
From this diagram for different excitations, we can see that the extremities of the
phasor of E (indicated by dots) are seen to lie on the straight line shown dashed. Since
all three diagrams refer to full-load power, the dotted line becomes the locus of E and of
the excitation, to scale for constant power output. This is the basis of the electrical load
diagram, Fig. 42.
Ir Izs Izs
v
Et
v
Ixs Ixs
v
V1
v v
V V
v v -Et
v
V
Ixs
V
V
Ir
V Izs
Et vΙ
V Ι
σ φ
φ σ
σ
φ
(a) (b) (c)
vΙ
Generator
leading current v V1 Motor leading current
A generator working-on infinite bus-bars will become a motor if its excitation is main-
tained and the prime mover replaced by a mechanical load. The change in the phasor of Ia
is shown in the phasor diagrams Fig. 41(a and b). V is the output voltage of the machine,
furnished by the e.m.f generated. For the motor, the current is in phase-opposition to V ,
since it is forced into the machine against the output voltage. For convenience, the supply
voltage V1 (equal and opposite to V ) may be used when the motor is considered, and the
diagram then becomes that of Fig. 41(c). The retarded angle δ of E or −E is descriptive of
the fact that when the shaft of the machine is loaded, it falls slightly relative to the stator
rotating field in order to develop the torque, required by the load.
Thus, the power-angle δ, Fig. 41, plays an important role in the operation of a syn-
chronous machine. Changes in load or excitation change its magnitude. When a machine
alters from generator to motor action, δ reverses; and when δ is caused to increase exces-
65
The electrical load diagram is shown in Fig. 42. The phasor V represents the constant volt-
age of the infinite bus-bars. At the extremity of V is drawn an axis showing the direction
of the Ia Zs drops—i.e. the voltage drops for unity-power-factor output currents. This axis
must be drawn at the angle θ = arc tan(Xs /r) to V , to scale along the axis is a distance
corresponding to, say, full load at unity power factor. At this point a line is drawn at right
angles to the axis. It is the locus of the E values for constant power, or constant-electrical-
power line. Other parallel lines are drawn for other loads, one through the extremity of V
itself corresponding to zero power output, others on the right-hand side of V corresponding
to negative power output, i.e. input to the machine as a motor.
Electrical load
per unit
2.0 1.5 1.0 0.5 0 0.5 1.0 1.5 2.0
θ >
m p n Generator Motor >
VI c
V2/zs δ
t
E
V = Et / zs a ρ
u.
m
p.
unity p.f
>
0
>
> >
2.
1.5
δ V
1.0
900 - θ
0.5
xs/r
pemax v θ = arctan pemax
Limit of stability 0
The diagram solves Eqn. 52. Consider the full-load unity-power- factor case in
Fig. 40, and multiply each complex voltage by the constant (V /Zs ). This gives the inset
in Fig. 42, from which V I = P = mp = mn − np. Now mn = (EV /Zs ) sin(90◦ − θ + δ) and
np = (V 2 /Zs ) cos θ, so that P is given directly by Eqn. 52.
If the excitation be fixed, the extremity of the e.m.f. vector E, will have a circular
66
locus as indicated by the circular arcs struck with O as centre. Taking 1.0 per unit E as
that for which E = V on no load and no current, the per-unit excitation for any other
loading condition can be found from the diagram. Thus with 1.5 per unit excitation, the
machine will work on full-load power as a generator with a power factor of cos 8◦ lagging; on
half-fun-load power with a power factor of cos 42◦ lagging; and on zero power output with
a power-factor of zero lagging, as shown by the lines pa, pb and pc. The variation of the
power output (controlled by the input from the prime mover in the case of a generator and
by the load applied to the shaft for a motor) with constant excitation is thl1S accompanied
by changes in the load power factor.
If the generator be provided with greater mechanical power with say, 150 per cent (or
1.5 per unit) excitation, then the output power increases with reducing power factor from
lagging values until, with an output (for this case) of 1.2 per unit power (see Fig. 42), the
power factor becomes unity. Thereafter the power increases with a reducing power factor-
now leading. Finally the excitation will not include any more constant-power lines, for the
circle of its locus becomes tangential to these. If more power is supplied by the prime
mover, the generator will be forced to rise out of step, and synchronous running will be lost.
The maximum power that can be generated is indicated by intercepts on the limit of stabil-
ity. The typical point Pemax on the left of the load diagram is for an excitation of 1.5 per unit.
67
Izs V
-Et
V
V
Ixs
V
V1
Ir O-curve
V
I Ia
σ
φ
Figure 43: Synchronous motor with constant output and variable excitation -Leading current
The O-curves are circular arcs, because Eqn. 55 represents the equation to a circle.
Writing
68
Izs
V V
V1 Ixs V
-Et
O-curve
I
σ
Figure 44: Synchronous motor with constant output and variable excitation- Unity p.f
Ir V
Izs
V
Ixs
V1 -Et
V
Ia σ I
Figure 45: Synchronous motor with constant output and variable excitation-Lagging current
69
f.
u.p
=0
1
P
st
on
=c
1
Armature current I
c a
E.M.F Et
Figure 46: Synchronous motor with constant output and variable excitation-V-curves
70
The construction of the mechanical load diagram is given in Fig. 47. Let OM = V1 /2r
to scale: draw with M as centre a circle of radius OM. This circle, from Eqn. 55, corresponds
to P I = 0, a condition for M which the circle radius is V1 /2r. The circle thus represents
the current locus for zero mechanical power. Any smaller circle on centre M represents the
current locus for some constant , mechanical power output P1 .
Qm current locus
for
V12 P1 P1=const
/
4r2
- r
M
Qn
V1/2r
Mech.power P1
P1=0
Again there are in general two values O-CURVES of current for each power output
P1 , the smaller OQn in the working range, the greater OQm above the limit of stability.
If P1 /r = V12 /4r 2 , there is a single value of current I = V1 /2r corresponding to the max-
imum power P1m = V12 /4r. The power circle has shrunk to zero radius and becomes in
fact the point M. The efficiency is 50 per cent, the I 2 R loss being equal to the mechanical
output. Such a condition is well outside the normal working range, not only because of
heating but also because the stability is critical. The case corresponds to the requirement of
71
The completed mechanical load diagram is shown in Fig. 48, with the addition of
OR = V /Zs drawn at angle arc cos(r/Zs ) to OM. Circles drawn with R as centre represent
constant values of E1 /Zs , or E, or the field excitation.
200
150
per
150
cal power
cen
100
t of
100
nor
rmal mechani
φ=0ο
ma
50
Moto
50
l ex
ο
45
φ=
φ=
cita
0 45 R
ο
ti
θ
on
percent of no
50 0
Gen
50
φ=
ο
φ=0ο
45
45
100
φ=
ο
100
150 150
The current loci in Fig. 48 are continued below the base line for generator operation. The
horizontal lines of constant mechanical power are now constant input (from the prime mover)
and a departure from unity-power-factor working, giving increased currents, increases the
I 2 R loss and lowers the available electrical output. The whole system of lines depends, of
course, on constant bus-bar voltage. The circular current loci are called the O - curves for
72
Maximum power
er
w
po
50 00% 0%
1 15
ad
%
y
bilit
lo
of sta
it
o
Lim
N
Generator Armature current Motor
φ=0 ο
Motor
r Generator
lo
we
po
o
N
%
50
0%
10
Lim
0%
it o
fs
15
tab
φ=
ility
0
ο
73
constant mechanical power. Any point P on the diagram, fixed by the percentage excitation
and load, gives by the line OP the current to scale in magnitude and phase. Directly from
the O-curves, Fig. 48, the V -curves, relating armature current and excitation for various
constant mechanical loads can be derived. These are shown in Fig. 49.
74
6 Synchronous motor
S N
N
T
N S
S Stationary
rotor poles
Figure 50: Force of attraction between stator poles and rotor poles - resulting in production
of torque in clockwise direction
try to maintain this alignment ( as per lenz’s law) and hence the rotor pole tries to follow
the stator pole as the conditions are suitable for the production of torque in the clockwise
direction. However the rotor cannot move instantaneously due to its mechanical inertia, and
so it needs sometime to move. In the mean time, the stator pole would quickly (a time
duration corresponding to half a cycle) change its polarity and becomes a South pole. So
the force of attraction will no longer be present and instead the like poles experience a force
75
of repulsion as shown in Fig. 51. In other words, the conditions are now suitable for the
S N
N
T
N S
S Stationary
rotor poles
Figure 51: Force of repulsion between stator poles and rotor poles - resulting in production
of torque in anticlockwise direction
production of torque in the anticlockwise direction. Even this condition will not last longer
as the stator pole would again change to North pole after a time of 2f1 . Thus the rotor will
experience an alternating force which tries to move it clockwise and anticlockwise at twice
the frequency of the supply, i.e. at intervals corresponding to 2f1 seconds. As this duration is
quite small compared to the mechanical time constant of the rotor, the rotor cannot respond
and move in any direction. The rotor continues to be stationary only.
Thus the synchronous rotor cannot start rotating on its own or usually we say
that the synchronous rotor has no starting torque. So, some special provision has to be
made either inside the machine or outside of the machine so that the rotor is brought to near
about its synchronous speed. At that time, if the armature is supplied with electrical power,
the rotor can pull into step and continue to operate at its synchronous speed. Some of the
commonly used methods for starting synchronous rotor are described in the following section.
76
Basically there are three methods that are used to start a synchronous motor:
• To reduce the speed of the rotating magnetic field of the stator to a low enough value
that the rotor can easily accelerate and lock in with it during one half-cycle of the
rotating magnetic field’s rotation. This is done by reducing the frequency of the applied
electric power. This method is usually followed in the case of inverter-fed synchronous
motor operating under variable speed drive applications.
• To use an external prime mover to accelerate the rotor of synchronous motor near to its
synchronous speed and then supply the rotor as well as stator. Ofcourse care should
be taken to ensure that the direction of rotation of the rotor as well as that of the
rotating magnetic field of the stator are the same. This method is usually followed in
the laboratory- the synchronous machine is started as a generator and is then connected
to the supply mains by following the synchronization or paralleling procedure. Then
the power supply to the prime mover is disconnected so that the synchronous machine
will continue to operate as a motor.
• To use damper windings or amortisseur windings if these are provided in the ma-
chine. The damper windings or amortisseur windings are provided in most of the
large synchronous motors in order to nullify the oscillations of the rotor whenever the
synchronous machine is subjected to a periodically varying load.
Each of these methods of starting a synchronous motor are described below in detail.
If the rotating magnetic field of the stator in a synchronous motor rotates at a low enough
speed, there will be no problem for the rotor to accelerate and to lock in with the stator’s
magnetic field. The speed of the stator magnetic field can then be increased to its rated op-
erating speed by gradually increasing the supply frequency f up to its normal 50- or 60-Hz
value.
This approach to starting of synchronous motors makes a lot of sense, but there is a big
problem: Where from can we get the variable frequency supply? The usual power supply
systems generally regulate the frequency to be 50 or 60 Hz as the case may be. However,
variable-frequency voltage source can be obtained from a dedicated generator only in the
77
olden days and such a situation was obviously impractical except for very unusual or special
drive applications.
But the present day solid state power converters offer an easy solution to this. We now
have the rectifier- inverter and cycloconverters, which can be used to convert a constant fre-
quency AC supply to a variable frequency AC supply. With the development of such modern
solid-state variable-frequency drive packages, it is thus possible to continuously control the
frequency of the supply connected to the synchronous motor all the way from a fraction
of a hertz up to and even above the normal rated frequency. If such a variable-frequency
drive unit is included in a motor-control circuit to achieve speed control, then starting the
synchronous motor is very easy-simply adjust the frequency to a very low value for starting,
and then raise it up to the desired operating frequency for normal running.
When a synchronous motor is operated at a speed lower than the rated speed, its internal
generated voltage (usually called the counter EMF) EA = Kφω will be smaller than normal.
As such the terminal voltage applied to the motor must be reduced proportionally with the
frequency in order to keep the stator current within the rated value. Generally, the voltage
in any variable-frequency power supply varies roughly linearly with the output frequency.
The second method of starting a synchronous motor is to attach an external starting motor
(pony motor) to it and bring the synchronous machine to near about its rated speed (but not
exactly equal to it, as the synchronization process may fail to indicate the point of closure of
the main switch connecting the synchronous machine to the supply system) with the pony
motor. Then the output of the synchronous machine can be synchronised or paralleled with
its power supply system as a generator, and the pony motor can be detached from the shaft
of the machine or the supply to the pony motor can be disconnected. Once the pony motor
is turned OFF, the shaft of the machine slows down, the speed of the rotor magnetic field
BR falls behind Bnet , momentarily and the synchronous machine continues to operate as a
motor. As soon as it begins to operates as a motor the synchronous motor can be loaded in
the usual manner just like any motor.
This whole procedure is not as cumbersome as it sounds, since many synchronous mo-
tors are parts of motor-generator sets, and the synchronous machine in the motor-generator
set may be started with the other machine serving as the starting motor. More over, the
starting motor is required to overcome only the mechanical inertia of the synchronous ma-
chine without any mechanical load ( load is attached only after the synchronous machine is
78
paralleled to the power supply system). Since only the motor’s inertia must be overcome,
the starting motor can have a much smaller rating than the synchronous motor it is going
to start. Generally most of the large synchronous motors have brushless excitation systems
mounted on their shafts. It is then possible to use these exciters as the starting motors. For
many medium-size to large synchronous motors, an external starting motor or starting by
using the exciter may be the only possible solution, because the power systems they are tied
to may not be able to handle the starting currents needed to use the damper (amortisseur)
winding approach described next.
As already mentioned earlier most of the large synchronous motors are provided with damper
windings, in order to nullify the oscillations of the rotor whenever the synchronous machine
is subjected to a periodically varying load. Damper windings are special bars laid into slots
cut in the pole face of a synchronous machine and then shorted out on each end by a large
shorting ring, similar to the squirrel cage rotor bars. A pole face with a set of damper wind-
ings is shown in Figure..
When the stator of such a synchronous machine is connected to the 3-Phase AC sup-
ply, the machine starts as a 3-Phase induction machine due to the presence of the damper
bars, just like a squirrel cage induction motor. Just as in the case of a 3-Phase squirrel cage
induction motor, the applied voltage must be suitably reduced so as to limit the starting cur-
rent to the safe rated value. Once the motor picks up to a speed near about its synchronous
speed, the DC supply to its field winding is connected and the synchronous motor pulls into
step i.e. it continues to operate as a Synchronous motor running at its synchronous speed.
The behavior of a synchronous motor can be predicted by considering its equivalent circuit
on similar lines to that of a synchronous generator as described below.
79
The equivalent-circuit model for one armature phase of a cylindrical rotor three phase syn-
chronous motor is shown in Fig. 52 exactly similar to that of a synchronous generator except
that the current flows in to the armature from the supply. All values are given per phase.
Applying Kirchhoff’s voltage law to Fig. 52,
Ia Ra jXl jXas
If
jXs R
DC
VT Ef source
Field winding
Figure 52: Equivalent-circuit model for one phase of a synchronous motor armature
Xs = Xl + Xas (59)
80
jIaxs
Iazs
) VT
-φ i
( 90
IaRa δ
IaZs
φi
jIaxs
Ef
Ia
IaRa
A phasor diagram shown in Fig. 53, illustrates the method of determining the counter
EMF which is obtained from the phasor equation;
Ef = VT − Ia Zs
The phase angle δ between the terminal voltage VT and the excitation voltage Ef in
Fig. 53 is usually termed the torque angle. The torque angle is also called the load angle or
power angle.
Except for very small machines, the armature resistance of a synchronous motor is relatively
insignificant compared to its synchronous reactance, so that Eqn. 61 to be approximated to
VT = Ef + jIa Xs (62)
The equivalent-circuit and phasor diagram corresponding to this relation are shown
in Fig. 54 and Fig. 55. These are normally used for analyzing the behavior of a synchronous
81
Since the left side of Eqn. 64 is an expression for active power input and as the winding
resistance is assumed to be negligible this power input will also represent the electromagnetic
power developed, per phase, by the synchronous motor.
Thus,
Eqn. 66, called the synchronous-machine power equation, expresses the electro mag-
netic power developed per phase by a cylindrical-rotor motor, in terms of its excitation volt-
age and power angle. Assuming a constant source voltage and constant supply frequency,
Eqn. 65 and Eqn. 66 may be expressed as proportionalities that are very useful for analyzing
the behavior of a synchronous-motor:
P ∝ Ia cos θ (69)
P ∝ Ef sin δ (70)
82
jXs
Ia
To AC
VT Ef
source
VT
δ
Efsinδ φi
φi jIaXs
Ia
IaXscosφi
Ef
Figure 55: Phasor diagram model for a synchronous-motor, assuming armature resistance is
negligible
83
6.3.3 Effect of changes in load on armature current, power angle, and power
factor of synchronous motor
The effects of changes in mechanical or shaft load on armature current, power angle, and
power factor can be seen from the phasor diagram shown in Fig. 56; As already stated, the
applied stator voltage, frequency, and field excitation are assumed, constant. The initial
load conditions, are represented by the thick lines. The effect of increasing the shaft load
to twice its initial value are represented by the light lines indicating the new steady state
conditions. These are drawn in accordance with Eqn. 69 and Eqn. 70, when the shaft load
is doubled both Ia cos φi and Ef sin δ are doubled. While redrawing the phasor diagrams
to show new steady-state conditions, the line of action of the new jIa Xs phasor must be
perpendicular to the new Ia phasor. Furthermore, as shown in Fig. 56, if the excitation is
not changed, increasing the shaft load causes the locus of the Ef phasor to follow a circular
arc, thereby increasing its phase angle with increasing shaft load. Note also that an increase
in shaft load is also accompanied by a decrease in φi ; resulting in an increase in power factor.
As additional load is placed on the machine, the rotor continues to increase its angle
2(Ia1cosφi1)
Ia1cosφi1
VT
δ2 δ1
φi1 Xs
jI a1 Ef1sinδ1
s
Ef1
2X
φi2
jIa
Figure 56: Phasor diagram showing effect of changes in shaft load on armature current,
power angle and power factor of a synchronous motor
of lag relative to the rotating magnetic field, thereby increasing both the angle of lag of
the counter EMF phasor and the magnitude of the stator current. It is interesting to note
that during all this load variation, however, except for the duration of transient conditions
84
whereby the rotor assumes a new position in relation to the rotating magnetic field, the
average speed of the machine does not change. As the load is being increased, a final point
is reached at which a further increase in δ fails to cause a corresponding increase in motor
torque, and the rotor pulls out of synchronism. In fact as stated earlier, the rotor poles at
this point, will fall behind the stator poles such that they now come under the influence of
like poles and the force of attraction no longer exists. Thus, the point of maximum torque
occurs at a power angle of approximately 90◦ for a cylindrical-rotor machine, as is indicated
by Eqn. 68. This maximum value of torque that causes a synchronous motor to pull out
of synchronism is called the pull-out torque. In actual practice, the motor will never be
operated at power angles close to 90◦ as armature current will be many times its rated value
at this load.
Intuitively we can expect that increasing the strength of the magnets will increase the mag-
netic attraction, and thereby cause the rotor magnets to have a closer alignment with the
corresponding opposite poles of the rotating magnetic poles of the stator. This will obvi-
ously result in a smaller power angle. This fact can also be seen in Eqn. 68. When the shaft
load is assumed to be constant, the steady-state value of Ef sin δ must also be constant. An
increase in Ef will cause a transient increase in Ef sin δ, and the rotor will accelerate. As
the rotor changes its angular position, δ decreases until Ef sin δ has the same steady-state
value as before, at which time the rotor is again operating at synchronous speed, as it should
run only at the synchronous speed. This change in angular position of the rotor magnets
relative to the poles of rotating magnetic field of the stator occurs in a fraction of a second.
The effect of changes in field excitation on armature current, power angle, and power
factor of a synchronous motor operating with a constant shaft load, from a constant voltage,
constant frequency supply, is illustrated in Fig. 57. From Eqn. 69, we have for a constant
shaft load,
Ia1 cos φi1 = Ia2 cos φi2 = Ia3 cos φi3 = Ia cos φi (72)
This is also shown in Fig. 57, where the locus of the tip of the Ia phasor is a line
85
Ia3
Locus of
Ia phasor
jIa2Xs
IaCosφi
Figure 57: Phasor diagram showing effect of changes in field excitation on armature current,
power angle and power factor of a synchronous motor
86
Note that increasing the excitation from Ef 1 to Ef 3 in Fig. 57 caused the phase angle
of the current phasor with respect to the terminal voltage VT (and hence the power factor)
to go from lagging to leading. The value of field excitation that results in unity power factor
is called normal excitation. Excitation greater than normal is called over excitation, and
excitation less than normal is called under excitation. Furthermore, as indicated in Fig. 57,
when operating in the overexcited mode, |Ef | > |VT |. In fact a synchronous motor operating
under over excitation condition is sometimes called a synchronous condenser.
6.3.5 V curves
Curves of armature current vs. field current (or excitation voltage to a different scale) are
called V curves, and are shown in Fig. 58 for typical values of synchronous motor loads. The
curves are related to the phasor diagram in Fig. 57, and illustrate the effect of the variation
of field excitation on armature current and power factor for typical shaft loads. It can be
easily noted from these curves that an increase in shaft loads require an increase in field
excitation in order to maintain the power factor at unity.
The locus of the left most point of the V curves in Fig. 58 represents the stability
limit (δ = −90◦ ). Any reduction in excitation below the stability limit for a particular load
will cause the rotor to pullout of synchronism.
87
Eqn. 74 is based on the phasor diagram and the assumption that Ra is negligible. It is to
be noted that instability will occur, if the developed torque is less than the shaft load plus
friction and windage losses, and the expression under the square root sign will be negative.
The family of V curves shown in Fig. 58 represent computer plots of Eqn. 74, by tak-
ing the data pertaining to a three-phase 10 hp synchronous motor i.e Vph = 230V and
Xs = 1.2Ω/phase.
160
oad
PF
dl
ty
140 e
rat
ni
U
0%
Armature current Ia A/phase
15 oad
120 t e dl
ra
Stability limit 0%
10
100
80
oad
e dl
rat
60 %
50
lagging Leading
40
power power factor
ad
factor
lo
20
o
N
0
0 50 100 150 200 250
Excitation voltage Ef V/phase
The flow of power through a synchronous motor, from stator to rotor and then to shaft
output, is shown in Fig. 59. As indicated in the power-flow diagram, the total power loss for
the motor is given by
88
Pfield
Pshaft
N
or
Pin
ot
R
S
Pgap Pstray
Pscl Pfcl Pf,w
Pcore
89
Induction Machines
1 Introduction
The induction machine was invented by NIKOLA TESLA in 1888. Right from its incep-
tion its ease of manufacture and its robustness have made it a very strong candidate for
electromechanical energy conversion. It is available from fractional horsepower ratings to
megawatt levels. It finds very wide usage in all various application areas. The induction
machine is an AC electromechanical energy conversion device. The machine interfaces with
the external world through two connections (ports) one mechanical and one electrical. The
mechanical port is in the form of a rotating shaft and the electrical port is in the form of
terminals where AC supply is connected. There are machines available to operate from three
phase or single phase electrical input. In this module we will be discussing the three phase
induction machine. Single phase machines are restricted to small power levels.
Click on the following steps in sequence to get a graphical picture. It is suggested that
the reader read the text before clicking the link.
• Consider a cosine wave from 0 to 360◦. This sine wave is plotted with unit amplitude.
• Now allow the amplitude of the sine wave to vary with respect to time in a simisoidal
fashion with a frequency of 50Hz.Let the maximum value of the amplitude is, say, 10
units. This waveform is a pulsating sine wave.
• Now consider a second sine wave, which is displaced by 120◦ from the first (lagging). . .
• and allow its amplitude to vary in a similar manner, but with a 120◦ time lag.
• and allow its amplitude to change as well with a 240◦ time lag. Now we have three
pulsating sine waves.
icpk = Im cos(2π.50.t − 240◦ ) (3)
Let us see what happens if we sum up the values of these three sine waves at every angle.
The result really speaks about Tesla’s genius. What we get is a constant amplitude travelling
sine wave!
In a three phase induction machine, there are three sets of windings — phase A winding,
phase B and phase C windings. These are excited by a balanced three-phase voltage supply.
This would result in a balanced three phase current. Equations 1 — 3 represent the currents
that flow in the three phase windings. Note that they have a 120◦ time lag between them.
Further, in an induction machine, the windings are not all located in the same place.
They are distributed in the machine 120◦ away from each other (more about this in the
section on alternators). The correct terminology would be to say that the windings have
their axes separated in space by 120◦ . This is the reason for using the phase A, B and C
since waves separated in space as well by 120◦ .
When currents flow through the coils, they generate mmfs. Since mmf is proportional to
current, these waveforms also represent the mmf generated by the coils and the total mmf.
Further, due to magnetic material in the machine (iron), these mmfs generate magnetic flux,
which is proportional to the mmf (we may assume that iron is infinitely permeable and
non-linear effects such as hysterisis are neglected). Thus the waveforms seen above would
also represent the flux generated within the machine. The net result as we have seen is
a travelling flux wave. The x-axis would represent the space angle in the machine as one
travels around the air gap. The first pulsating waveform seen earlier would then represent
the a-phase flux, the second represents the b-phase flux and the third represents the c-phase.
This may be better visualized in a polar plot. The angles of the polar plot represent the
space angle in the machine, i.e., angle as one travels around the stator bore of the machine.
Click on the links below to see the development on a polar axes.
• This plot shows the pulsating wave at the zero degree axes. The amplitude is maximum
at zero degree axes and is zero at 90◦ axis. Positive parts of the waveform are shown
in red while negative in blue. Note that the waveform is pulsating at the 0 − 180◦ axis
and red and blue alternate in any given side. This corresponds to the sinewave current
changing polarity. Note that the maximum amplitude of the sinewave is reached only
along the 0 − 180◦ axis. At all other angles, the amplitude does not reach a maximum
of this value. It however reaches a maximum value which is less than that of the peak
occuring at the 0 − 180◦ axis. More exactly, the maximum reached at any space angle
θ would be equal to cosθ times the peak at the 0 − 180◦ axis. Further, at any space
angle θ, the time variation is sinusoidal with the frequency and phase lag being that
of the excitation, and amplitude being that corresponding to the space angle.
• This plot shows the pulsating waveforms of all three cosines. Note that the first is
pulsating about the 0 − 180◦ axis, the second about the120◦ − 300◦ axis and the third
at 240◦ − 360◦axis.
• This plot shows the travelling wave in a circular trajectory. Note that while individual
pulsating waves have maximum amplitude of 10, the resultant has amplitude of 15.
If f1 is the amplitude of the flux waveform in each phase, the travelling wave can then
be represented as
2π 2π 4π 4π
f (t) = f1 cos ωt cos θ + f1 cos(ωt − ) cos(θ − ) + f1 cos(ωt − ) cos(θ − )
3 3 3 3
3
= f1 cos(ωt − θ) (4)
2
1. what is the interpretation of the pulsating plots of the animation? If one wants to
know the ‘a’ phase flux at a particular angle for all instants of time, how can it be
obtained?
2. What will this time variation look like? It is obviously periodic. What will be the
amplitude and frequency?
Since the flux pattern is varying sinusoidally in space, as the flux waveform rotates, the
flux linkage varies sinusoidally. The rate of variation of this flux linkage will then be equal
to the speed of rotation of the air gap flux produced. This sinusoidal variation of the flux
linkage produces a sinusoidal induced emf in the rotor coil. If the coil is short circuited, this
induced emf will cause a current flow in the coil as per Lenz’s law.
Now imagine a second coil on the rotor whose axis is 120◦ away from the first. This is
shown in fig. 2. The flux linkage in this coil will also vary sinusoidally with respect to time
and therefore cause an induced voltage varying sinusoidally with time. However the flux
linkages in these two coils will have a phase difference of 120◦ (the rotating flux wave will
have to travel 120◦ in order to cause a similar flux linkage variation as in the first coil), and
hence the time varying voltages induced in the coils will also have a 120◦ phase difference.
A third coil placed a further 120◦ away is shown in fig. 3. This will have a time varying
induced emf lagging 240◦ in time with respect to the first.
When these three coils are shorted upon themselves currents flow in them as per Lenz’s
law. The mechanism by which torque is produced may now be understood as follows. The
diagram in fig. 4 shows a view of the rotor seen from one end. Positive current is said to
′ ′
flow in these coils when current flows out of the page in a, b, c conductors and into a , b
′
and c respectively.
If we look at the voltage induced in these coils as phasors, the diagram looks as shown
in fig. 5. The main flux is taken as the reference phasor. Considering that the induced emf
is −dψ/dt where ψ is the flux linkage, the diagram is drawn as shown.
As usual, the horizontal component of these phasors gives the instantaneous values of
the induced emf in these coils.
Let these coils be purely resistive. Then these emf phasors also represent the currents
flowing in these coils. If we consider the instant t = 0, it can be seen that
These currents act to produce mmf and flux along the axes of the respective coils. Let
′
us consider the space around b and c coil sides. The situation is shown in fig. 6.
The resulting flux pattern causes a tendency to move in the anticlockwise direction. This
is easy to see through the so called whiplash rule. Alternatively, since the force on a current
c’ b’
b c
a’
carrying conductor is F = q(v X B), it can be seen that the torque produced tends to rotate
the rotor counter-clockwise. The magnitude of the torque would increase with the current
magnitude in the coils. This current is in turn dependent on the magnitude of the main field
flux and its speed of rotation. Therefore one may say that motion of the main field tends to
drag the rotor along with it.
When the rotor is free to move and begins moving, the motion reduces the relative speed
between the main field and the rotor coils. Less emf would therefore be induced and the
torque would come down. Depending on the torque requirement for the load, the difference
in speed between the rotor and the main field settles down at some particular value.
From the foregoing, the following may be noted.
1. The torque produced depends on a non-zero relative speed between the field and the
rotor.
eb ec
V
1200
300
F
V
1200
900
V
ea
2. It is therefore not possible for the rotor to run continuously at the same speed of the
field. This is so because in such a condition, no emf would be induced in the rotor and
hence no rotor current, no torque.
3. The frequency of currents induced in the rotor coils and their magnitude depends on
this difference in speed.
These are important conclusions. The speed of the main field is known as the synchronous
speed, ns . If the actual speed of the rotor is nr then the ratio
ns − nr
s= (5)
ns
are designed to operate at about less than 4 percent slip at full load.
It is instructive to see the situation if the rotor resistance is neglected and is considered
to be purely inductive. The phasor diagram of voltages and the currents would then look as
shown in fig. 7.
V
V
300
V
V
300
At t = 0, one can see that current in a phase coil is at negative maximum, while b and
c phases have positive current of 0.5 units. Now if we consider the current flux profiles at
′
coil sides a, b , c, the picture that emerges is shown in fig. 8.
Since main flux at the a coil side is close to zero, there is very little torque produced
′
from there. There is a tendency to move due to the b and c coil sides, but they are in
opposite directions however. Hence there is no net torque on the rotor. This brings up
another important conclusion — the resistance of the rotor is an important part of torque
production in the induction machine. While a high resistance rotor is better suited for torque
production, it would also be lossy.
X a
b’
V
V
V
10
4 Construction
In actual practice, the three coils form three windings distributed over several slots. These
windings may be connected in star or delta and three terminations are brought out. These
are conventional three phase windings which are discussed in greater detail in the chapters
on alternators. Such windings are present n the stator as well as rotor. A photograph of
the stator of an induction machine is shown in fig. 9. A close up of the windings is shown in
fig. 10.the several turns that makeup a coil are seen in this picture. The three terminations
are connected to rings on which three brushes make a sliding contact. As the rotor rotates
the brushes slip over the rings and provide means of connecting stationary external circuit
elements to the rotating windings. A schematic of these arrangements is shown in fig. 13. A
photograph of a wound rotor for an induction machine is shown in fig. 11. Fig. 12 shows a
close up of the slip ring portion. Brushes are not shown in this picture.
Induction machines, which have these kinds of windings and terminals that are brought
out, are called slip ring machines. The reader may note that in order that torque is produced
current must flow in the rotor. To achieve that, the stationary brush terminals must either
be shorted, or connected to a circuit allowing current flow. Sometimes a star connected
resistor bank is connected so that the developed starting torque is higher. There are also
other forms of power electronic circuitry that may be connected to the rotor terminals to
achieve various functions.
The popularity of the induction machine however, stems from another variety of rotor
11
12
Rotor shaft
v
Winding Brushes(stationary)
on rotor
Sliding
Contact Stationary terminals
13
that is used. This rotor has slots into which copper or aluminium bars are inserted. These
bars are then shorted by rings that are brazed on to each of the rotor ends. Figure 14 shows
a simple schematic.
Such a rotor is called squirrel cage rotor. This rotor behaves like a short-circuited winding
and hence the machine is able to perform electromechanical energy conversion. This type
of rotor is easy to manufacture, has no sliding contacts and is very robust. It is this feature
that makes induction machine suitable for use even in hazardous environments and reliable
operation is achieved. The disadvantage of this type of rotor is that the motor behavior
cannot be altered by connecting anything to the rotor — there are no rotor terminals.
Fig. 15 shows a photograph of a squirrel cage rotor. The rotor also has a fan attached
to it. This is for cooling purposes. The bars ( white lines on the surface) are embedded in
the rotor iron which forms the magnetic circuit. The white lines correspond to the visible
portion of the rotor bar.
Sometimes two rotor bars are used per slot to achieve some degree of variability in the
starting and running performances. It is to make use of the fact that while high rotor
14
resistance is desirable from the point of view of starting torque, low rotor resistance is
desirable from efficiency considerations while the machine is running. Such rotors are called
double cage rotors or deep-bar rotors.
To summarize the salient features discussed so far,
1. The stator of the 3 - phase induction machine consists of normal distributed AC wind-
ings.
2. Balanced three phase voltages impressed on the stator, cause balanced three phase
currents to flow in the stator.
3. These stator currents cause a rotating flux pattern (the pattern is a flux distribution
which is sinusoidal with respect to the space angle) in the air gap.
4. The rotating flux pattern causes three phase induced e.m.f.s in rotor windings (again
normal ac windings). These windings, if shorted, carry three phase-balanced currents.
Torque is produced as a result of interaction of the currents and the air gap flux.
5. The rotor may also take the form of a squirrel cage arrangement, which behaves in a
manner similar to the short-circuited three phase windings.
15
5 Equivalent Circuit
It is often required to make quantitative predictions about the behavior of the induction
machine, under various operating conditions. For this purpose, it is convenient to represent
the machine as an equivalent circuit under sinusoidal steady state operating conditions. Since
the operation is balanced, a single-phase equivalent circuit is sufficient for most purposes.
In order to derive the equivalent circuit, let us consider a machine with an open circuited
rotor. Since no current can flow and as a consequence no torque can be produced, the
situation is like a transformer open-circuited on the secondary (rotor). The equivalent circuit
under this condition can be drawn as shown in fig. 16.
Rs Xls Rr Xlr
Rm Xm
This is just the normal transformer equivalent circuit (why? ). Measurements aregenerally
made on the stator side and the rotor, in most circumstances, is shorted (if required, through
some external circuitry). Since most of the electrical interaction is from the stator, it makes
sense to refer all parameters to the stator.
Let us consider the rotor to be shorted. Let the steady speed attained by the rotor be ωr
and the synchronous speed be ωs . The induced voltage on the rotor is now proportional to
the slip i.e., slip times the induced voltage under open circuit (why? ). Further, the voltage
induced and the current that flows in the rotor is at a frequency equal to slip times the
stator excitation frequency (why? ). The equivalent circuit can be made to represent this by
shorting the secondary side and is shown in fig. 17.
′ ′
Rr and Xlr refer to the rotor resistance and leakage resistance referred to the stator side
(using the square of the turns ratio, as is done in transformer). The secondary side loop is
excited by a voltage sE1 , which is also at a frequency sf1 . This is the reason why the rotor
16
′ ′
Rs Xls Rr sXlr
Rm Xm E1 sE1
′
leakage is sXlr now . The current amplitude in the rotor side would therefore be
′ sE1
Ir = p ′ (6)
Rr + (sXlr2 )
′
2
′ E1
Ir = q (7)
R′r2
+ (Xlr2 )
′
s2
Equation 7 tells us that the rotor current is the same as the current flowing in a circuit
′ ′
with a load impedance consisting of a resistance Rr /s and inductive reactance Xlr . This
current would also now be at the frequency of E1 (stator frequency). Note that the slip no
longer multiplies the leakage reactance. Further this current is now caused by a voltage of
E1 itself (no multiplying factor of s). Hence the transformer in fig. 17 can also be removed.
Since, with this, the conversion to slip frequency is no longer there, the equivalent circuit
can be represented as in fig. 18.
This is then the per-phase equivalent circuit of the induction machine, also called as exact
equivalent circuit. Note that the voltage coming across the magnetizing branch is the applied
stator voltage, reduced by the stator impedance drop. Generally the stator impedance drop
is only a small fraction of the applied voltage. This fact is taken to advantage and the
magnetizing branch is shifted to be directly across the input terminals and is shown in
fig. 19.
17
′
Rr ′
Rs Xls s
Xlr
Rm Xm
Rm Xm
This circuit, called the approximate equivalent circuit, is simple to use for quick calcula-
tions.
′
Rr
The resistance term s
could be split into two parts.
′ ′
Rr ′ R (1 − s)
= Rr + r (8)
s s
With this equation the equivalent circuit can be modified as shown in fig. 20.
Dividing the equation for the rotor current by s and merging the two sides of the trans-
former is not just a mathematical jugglery. The power dissipated in the rotor resistance (per
′ ′
phase) is obviously I22 Rr . From the equivalent circuit of fig. 20 one can see that the rotor
′
current (referred to stator of course) flows through a resistance Rr /s which has a component
′ ′
Rr (1 − s)/s in addition to Rr , which also dissipates power. What does this represent?
18
′ ′
Rs Xls R Xlr
′
Rm Xm Rr (1−s)
s
From the equivalent circuit, one can see that the dissipation in Rs represents the stator
loss, and dissipation in Rm represents the iron loss. Therefore, the power absorption indicated
by the rotor part of the circuit must represent all other means of power consumption -
the actual mechanical output, friction and windage loss components and the rotor copper
′
loss components. Since the dissipation in Rr is rotor copper loss, the power dissipation in
′
Rr (1 − s)/s is the sum total of the remaining. In standard terminology, dissipation in
′
• Rr /s is called the air gap power.
′
• Rr is the rotor copper loss.
′
• Rr (1 − s)/s is the mechanical output.
In an ideal case where there are no mechanical losses, the last term would represent the
actual output available at the shaft. Out of the power Pg Transferred at the air gap, a
fraction s is dissipated in the rotor and (1 − s) is delivered as output at the shaft. If there
are no mechanical losses like friction and windage, this represents the power available to the
load.
19
1. When the machine is run on no-load, there is very little torque developed by it. In an
ideal case where there is no mechanical losses, there is no mechanical power deveoped
at no-load. Recalling the explanations in the section on torque production, the flow
of current in the rotor is indicative of the torque that is produced. If no torque is
produced, one may conclude that no current would be flowing in the rotor either.
The rotor branch acts like an open circuit. This conclusion may also be reached by
reasoning that when there is no load, an ideal machine will run up to its synchronous
speed where the slip is zero resulting in an infinite impedance in the rotor branch.
2. When the machine is prevented from rotation, and supply is given, the slip remains at
unity. The elements representing the magnetizing branch Rm &Xm are high impedances
′ ′
much larger than Rr & Xlr in series. Thus, in the exact equivalent circuit of the
induction machine, the magnetizing branch may be neglected.
From these considerations, we may reduce the induction machine exact equivalent circuit
of fig.18 to those shown in fig. 21.
′ ′
Rs Xls Rs Xls R Xlr
′
Rm Xm Rr (1−s)
s
These two observations and the reduced equivalent circuits are used as the basis for the
two most commonly used tests to find out the equivalent circuit parameters — the blocked
rotor test and no load test. They are also referred to as the short circuit test and open
circuit test respectively in conceptual analogy to the transformer.
20
The behaviour of the machine may be judged from the equivalent circuit of fig. 21(a). The
current drawn by the machine causes a stator-impedance drop and the balance voltage is
applied across the magnetizing branch. However, since the magnetizing branch impedance is
large, the current drawn is small and hence the stator impedance drop is small compared to
the applied voltage (rated value). This drop and the power dissipated in the stator resistance
are therefore neglected and the total power drawn is assumed to be consumed entirely as
core loss. This can also be seen from the approximate equivalent circuit, the use of which is
justified by the foregoing arguments. This test therefore enables us to compute the resistance
and inductance of the magnetizing branch in the following manner.
Let applied voltage = Vs . Then current drawn is given by
Vs Vs
Is = + (9)
Rm jXm
Vs , Is and Ps are measured with appropriate meters. With Rm known from eqn. 10, Xm
can be found from eqn. 9. The current drawn is at low power factor and hence a suitable
wattmeter should be used.
In this test the rotor is prevented from rotation by mechanical means and hence the name.
Since there is no rotation, slip of operation is unity, s = 1. The equivalent circuit valid under
these conditions is shown in fig. 21(b). Since the current drawn is decided by the resistance
and leakage impedances alone, the magnitude can be very high when rated voltage is applied.
Therefore in this test, only small voltages are applied — just enough to cause rated current
to flow. While the current magnitude depends on the resistance and the reactance, the power
drawn depends on the resistances.
The parameters may then be determined as follows. The source current and power drawn
may be written as
Vs
Is = (11)
(Rs + Rr ) + j(Xs + Xr′ )
′
′
Ps = |Is |2 (Rs + Rr ) (12)
21
In the test Vs , Is and Ps are measured with appropriate meters. Equation 12 enables us
′ ′
to compute(Rs + Rr ). Once this is known, (Xs + Xr ) may be computed from the eqn. 11.
Note that this test only enables us to determine the series combination of the resistance
and the reactance only and not the individual values. Generally, the individual values are
′ ′
assumed to be equal; the assumption Rs = Rr , andXs = Xr suffices for most purposes. In
practice, there are differences. If more accurate estimates are required IEEE guidelines may
be followed which depend on the size of the machine.
Note that these two tests determine the equivalent circuit parameters in a ‘Stator-referred’
sense, i.e., the rotor resistance and leakage inductance are not the actual values but what
they ’appear to be’ when looked at from the stator. This is sufficient for most purposes as
interconnections to the external world are generally done at the stator terminals.
22
In order to estimate the speed torque characteristic let us suppose that a sinusoidal voltage
is impressed on the machine. Recalling that the equivalent circuit is the per-phase represen-
tation of the machine, the current drawn by the circuit is given by
Vs
Is = (14)
R′r ′
(Rs + s
) + j(Xls + Xlr )
where Vs is the phase voltage phasor and Is is the current phasor. The magnetizing
′
current is neglected. Since this current is flowing through Rsr , the air-gap power is given by
′
2 Rr
Pg = |Is |
s
′
Vs Rr
= ′ (15)
(Rs + Rsr )2 + (Xls + Xlr )2 s
′
′
The mechanical power output was shown to be (1 − s)Pg (power dissipated in Rr /s). The
torque is obtained by dividing this by the shaft speed ωm .Thus we have,
′
Pg (1 − s) Pg (1 − s) R
= = |Is |2 r (16)
ωm ωs (1 − s) sωs
where ωs is the synchronous speed in radians per second and s is the slip. Further, this
is the torque produced per phase. Hence the overall torque is given by
23
′
3 Vs2 R
Te = . ′ . r (17)
ωs (Rs + Rr 2
) + (Xls + Xlr ) s
′
s
The torque may be plotted as a function of ‘s’ and is called the torque-slip (or torque-
speed, since slip indicates speed) characteristic — a very important characteristic of the
induction machine. Eqn. 17 is valid for a two-pole (one pole pair) machine. In general,
this expression should be multiplied by p, the number of pole-pairs. A typical torque-speed
characteristic is shown in fig. 22. This plot corresponds to a 3 kW, 4 pole, 60 Hz machine.
The rated operating speed is 1780 rpm.
80
70
60
50
Torque, Nm
40
30
20
10
0
0 200 400 600 800 1000 1200 1400 1600 1800
speed, rpm
We must note that the approximate equivalent circuit was used in deriving this relation.
Readers with access to MATLAB or suitable equivalents (octave, scilab available free under
GNU at the time of this writing) may find out the difference caused by using the ‘exact’
equivalent circuit by using the script found here. A comparison between the two is found
in the plot of fig. 23. The plots correspond to a 3 kW, 4 pole, 50 Hz machine, with a
rated speed of 1440 rpm. It can be seen that the approximate equivalent circuit is a good
approximation in the operating speed range of the machine. Comparing fig. 22 with fig. 23,
we can see that the slope and shape of the characteristics are dependent intimately on the
machine parameters.
Further, this curve is obtained by varying slip with the applied voltage being held con-
stant. Coupled with the fact that this is an equivalent circuit valid under steady state, it
implies that if this characteristic is to be measured experimentally, we need to look at the
torque for a given speed after all transients have died down. One cannot, for example, try
to obtain this curve by directly starting the motor with full voltage applied to the terminals
24
70
60 approximate
exact
50
Torque, Nm
40
30
20
10
0
0 200 400 600 800 1000 1200 1400 1600
Speed,rpm
and measuring the torque and speed dynamically as it runs up to steady speed.
Another point to note is that the equivalent circuit and the values of torque predicted
is valid when the applied voltage waveform is sinusoidal. With non-sinusoidal voltage wave-
forms, the procedure is not as straightforward.
With respect to the direction of rotation of the air-gap flux, the rotor maybe driven
to higher speeds by a prime mover or may also be rotated in the reverse direction. The
torque-speed relation for the machine under the entire speed range is called the complete
speed-torque characteristic. A typical curve is shown in fig. 7.1 for a four-pole machine,
the synchronous speed being 1500 rpm. Note that negative speeds correspond to slip values
greater than 1, and speeds greater than 1500 rpm correspond to negative slip. The plot also
shows the operating modes of the induction machine in various regions. The slip axis is also
shown for convenience.
Restricting ourselves to positive values of slip, we see that the curve has a peak point.
This is the maximum torque that the machine can produce, and is called as stalling torque.
If the load torque is more than this value, the machine stops rotating or stalls. It occurs at
a slip ŝ, which for the machine of fig. 7.1 is 0.38. At values of slip lower than ŝ, the curve
falls steeply down to zero at s = 0. The torque at synchronous speed is therefore zero. At
values of slip higher than s = ŝ, the curve falls slowly to a minimum value at s = 1. The
torque at s = 1 (speed = 0) is called the starting torque.
The value of the stalling torque may be obtained by differentiating the expression for
torque with respect to zero and setting it to zero to find the value of ŝ. Using this method,
25
80
60
40
20
Torque, Nm
0
−20
−60
−80
−100
−120
−1500 −1000 −500 0 500 1000 1500 2000 2500 3000
speed, rpm
′
±Rr
ŝ = p ′
(18)
Rr′ 2 + (Xls + Xlr )2
Substituting ŝ into the expression for torque gives us the value of the stalling torque Tˆe .
3V 2 1
Tˆe = s . p (19)
2ωs Rs ± Rs2 + (Xls + Xlr′ )2
the negative sign being valid for negative slip.
The expression shows that Tˆe is the independent of Rr , while ŝ is directly proportional to
′
′ ′
Rr . This fact can be made use of conveniently to alter ŝ. If it is possible to change Rr , then
we can get a whole series of torque-speed characteristics, the maximum torque remaining
constant all the while. But this is a subject to be discussed later.
′
p ′
We may note that if Rr is chosen equal to Rs2 + (Xls + Xlr )2 , ŝ, becomes unity, which
′
means that the maximum torque occurs at starting. Thus changing of Rr , wherever possible
can serve as a means to control the starting torque.
While considering the negative slip range, (generator mode) we note that the maximum
torque is higher than in the positive slip region (motoring mode).
26
Consider a speed torque characteristic shown in fig. 25 for an induction machine, having
the load characteristic also superimposed on it. The load is a constant torque load i.e.,the
torque required for operation is fixed irrespective of speed.
120
motor
100
80
torque, Nm
load
1
60
2
40
20
0
0 500 1000 1500 2000 2500 3000
speed, rpm
The system consisting of the motor and load will operate at a point where the two
characteristics meet. From the above plot, we note that there are two such points. We
therefore need to find out which of these is the actual operating point.
To answer this we must note that, in practice, the characteristics are never fixed; they
change slightly with time. It would be appropriate to consider a small band around the
curve drawn where the actual points of the characteristic will lie. This being the case let
us considers that the system is operating at point 1, and the load torque demand increases
slightly. This is shown in fig. 26, where the change is exaggerated for clarity. This would
shift the point of operation to a point 1′ at which the slip would be less and the developed
torque higher.
27
120
100
1’
2’
80
Te
torque, Nm
60
1 2
40
20
0
0 500 1000 1500 2000 2500 3000
speed, rpm
The difference in torque developed △Te , being positive will accelerate the machine. Any
overshoot in speed as it approaches the point 1′ will cause it to further accelerate since
the developed torque is increasing. Similar arguments may be used to show that if for
some reason the developed torque becomes smaller the speed would drop and the effect is
cumulative. Therefore we may conclude that 1 is not a stable operating point.
Let us consider the point 2. If this point shifts to 2′ , the slip is now higher (speed is
lower) and the positive difference in torque will accelerate the machine. This behavior will
tend to bring the operating point towards 2 once again. In other words, disturbances at
point 2 will not cause a runaway effect. Similar arguments may be given for the case where
the load characteristic shifts down. Therefore we conclude that point 2 is a stable operating
point.
From the foregoing discussions, we can say that the entire region of the speed-torque
characteristic from s = 0 to s = ŝ is an unstable region, while the region from s = ŝ to s = 0
is a stable region. Therefore the machine will always operate between s = 0 and s = ŝ.
28
The reader is referred to fig. 7.1 which shows the complete speed-torque characteristic of the
induction machine along with the various regions of operation.
Let us consider a situation where the machine has just been excited with three phase
supply and the rotor has not yet started moving. A little reflection on the definition of the
slip indicates that we are at the point s = 1. When the rotating magnetic field is set up due
to stator currents, it is the induced emf that causes current in the rotor, and the interaction
between the two causes torque. It has already been pointed out that it is the presence of the
non-zero slip that causes a torque to be developed. Thus the region of the curve between
s = 0 and s = 1 is the region where the machine produces torque to rotate a passive load
and hence is called the motoring region. Note further that the direction of rotation of the
rotor is the same as that of the air gap flux.
Suppose when the rotor is rotating, we change the phase sequence of excitation to the
machine. This would cause the rotating stator field to reverse its direction — the rotating
stator mmf and the rotor are now moving in opposite directions. If we adopt the convention
that positive direction is the direction of the air gap flux, the rotor speed would then be a
negative quantity. The slip would be a number greater than unity. Further, the rotor as
we know should be ”dragged along” by the stator field. Since the rotor is rotating in the
opposite direction to that of the field, it would now tend to slow down, and reach zero speed.
Therefore this region (s > 1) is called the braking region. (What would happen if the supply
is not cut-off when the speed reaches zero?)
There is yet another situation. Consider a situation where the induction machine is
operating from mains and is driving an active load (a load capable of producing rotation by
itself). A typical example is that of a windmill, where the fan like blades of the wind mill
are connected to the shaft of the induction machine. Rotation of the blades may be caused
by the motoring action of the machine, or by wind blowing. Further suppose that both
acting independently cause rotation in the same direction. Now when both grid and wind
act, a strong wind may cause the rotor to rotate faster than the mmf produced by the stator
excitation. A little reflection shows that slip is then negative. Further, the wind is rotating
the rotor to a speed higher than what the electrical supply alone would cause. In order to
do this it has to contend with an opposing torque generated by the machine preventing the
speed build up. The torque generated is therefore negative. It is this action of the wind
against the torque of the machine that enables wind-energy generation. The region of slip
s > 1 is the generating mode of operation. Indeed this is at present the most commonly used
approach in wind-energy generation. It may be noted from the torque expression of eqn. 17
that torque is negative for negative values of slip.
29
From the torque equation of the induction machine given in eqn.17, we can see that the
torque depends on the square of the applied voltage. The variation of speed torque curves
with respect to the applied voltage is shown in fig. 27. These curves show that the slip at
maximum torque ŝ remains same, while the value of stall torque comes down with decrease
in applied voltage. The speed range for stable operation remains the same.
Further, we also note that the starting torque is also lower at lower voltages. Thus, even
if a given voltage level is sufficient for achieving the running torque, the machine may not
start. This method of trying to control the speed is best suited for loads that require very
little starting torque, but their torque requirement may increase with speed.
Figure 27 also shows a load torque characteristic — one that is typical of a fan type of
load. In a fan (blower) type of load,the variation of torque with speed is such that T ∝ ω 2 .
Here one can see that it may be possible to run the motor to lower speeds within the range
ns to (1 − ŝ)ns . Further, since the load torque at zero speed is zero, the machine can start
even at reduced voltages. This will not be possible with constant torque type of loads.
One may note that if the applied voltage is reduced, the voltage across the magnetising
branch also comes down. This in turn means that the magnetizing current and hence flux
level are reduced. Reduction in the flux level in the machine impairs torque production
30
60
V1
50
40
torque, Nm
V2
30
20 V3 o1
o2
10 o3
load
V1 > V2 > V3
0
0 500 1000 1500
speed, rpm
(recall explantions on torque production), which is primarily the explanation for fig. 27. If,
however, the machine is running under lightly loaded conditions, then operating under rated
flux levels is not required. Under such conditions, reduction in magnetizing current improves
the power factor of operation. Some amount of energy saving may also be achieved.
Voltage control may be achieved by adding series resistors (a lossy, inefficient proposition),
or a series inductor / autotransformer (a bulky solution) or a more modern solution using
semiconductor devices. A typical solid state circuit used for this purpose is the AC voltage
controller or AC chopper. Another use of voltage control is in the so-called ‘soft-start’ of the
machine. This is discussed in the section on starting methods.
31
The reader may recall from eqn.17 the expression for the torque of the induction machine.
Clearly, it is dependent on the rotor resistance. Further, eqn.19 shows that the maximum
value is independent of the rotor resistance. The slip at maximum torque eqn.18 is dependent
on the rotor resistance. Therefore, we may expect that if the rotor resistance is changed, the
maximum torque point shifts to higher slip values, while retaining a constant torque. Figure
28 shows a family of torque-speed characteristic obtained by changing the rotor resistance.
60
50
40 r2 r1
torque, Nm
r3
30
20
o1
o2
10 o3
r3 > r2 > r1
0
0 500 1000 1500
speed, rpm
Note that while the maximum torque and synchronous speed remain constant, the slip
at which maximum torque occurs increases with increase in rotor resistance, and so does the
starting torque. whether the load is of constant torque type or fan-type, it is evident that
the speed control range is more with this method. Further, rotor resistance control could
also be used as a means of generating high starting torque.
For all its advantages, the scheme has two serious drawbacks. Firstly, in order to vary
32
the rotor resistance, it is necessary to connect external variable resistors (winding resistance
itself cannot be changed). This, therefore necessitates a slip-ring machine, since only in
that case rotor terminals are available outside. For cage rotor machines, there are no rotor
terminals. Secondly, the method is not very efficient since the additional resistance and
operation at high slips entails dissipation.
The resistors connected to the slip-ring brushes should have good power dissipation ca-
pability. Water based rheostats may be used for this. A ‘solid-state’ alternative to a rheostat
is a chopper controlled resistance where the duty ratio control of of the chopper presents a
variable resistance load to the rotor of the induction machine.
The power drawn from the rotor terminals could be spent more usefully. Apart from using
the heat generated in meaning full ways, the slip ring output could be connected to another
induction machine. The stator of the second machine would carry slip frequency currents of
the first machine which would generate some useful mechanical power. A still better option
would be to mechanically couple the shafts of the two machines together. This sort of a
connection is called cascade connection and it gives some measure of speed control as shown
below.
Let the frequency of supply given to the first machine be f1 , its number poles be p1 , and
its slip of operation be s1 . Let f2 , p2 and s2 be the corresponding quantities for the second
machine. The frequency of currents flowing in the rotor of the first machine and hence in the
stator of the second machine is s1 f1 . Therefore f2 = s1 f1 . Since the machines are coupled
at the shaft, the speed of the rotor is common for both. Hence, if n is the speed of the rotor
in radians,
f1 s1 f1
n = (1 − s1 ) = ± (1 − s2 ). (20)
p1 p2
Note that while giving the rotor output of the first machine to the stator of the second,
the resultant stator mmf of the second machine may set up an air-gap flux which rotates in
the same direction as that of the rotor, or opposes it. this results in values for speed as
f1 f1
n= or n= (s2 negligible) (21)
p1 + p2 p1 − p2
The latter expression is for the case where the second machine is connected in opposite
phase sequence to the first. The cascade connected system can therefore run at two possible
33
′ ′
Rs Xls Rr sXlr
+
Rm Xm E1 sE1 Er
speeds.
Speed control through rotor terminals can be considered in a much more general way.
Consider the induction machine equivalent circuit of fig. 29, where the rotor circuit has been
terminated with a voltage source Er .
If the rotor terminals are shorted, it behaves like a normal induction machine. This is
equivalent to saying that across the rotor terminals a voltage source of zero magnitude is
connected. Different situations could then be considered if this voltage source Er had a
non-zero magnitude. Let the power consumed by that source be Pr . Then considering the
rotor side circuit power dissipation per phase
′ ′ ′
sE1 I2 cos φ2 = I2 R2 + Pr . (22)
Clearly now, the value of s can be changed by the value of Pr . For Pr = 0, the machine
is like a normal machine with a short circuited rotor. As Pr becomes positive, for all other
circuit conditions remaining constant, s increases or in the other words, speed reduces. As
Pr becomes negative,the right hand side of the equation and hence the slip decreases. The
physical interpretation is that we now have an active source connected on the rotor side
′
which is able to supply part of the rotor copper losses. When Pr = −I22 R2 the entire copper
loss is supplied by the external source. The RHS and hence the slip is zero. This corresponds
to operation at synchronous speed. In general the circuitry connected to the rotor may not
be a simple resistor or a machine but a power electronic circuit which can process this power
requirement. This circuit may drive a machine or recover power back to the mains. Such
circuits are called static kramer drives.
34
Sometimes induction machines have a special stator winding capable of being externally
connected to form two different number of pole numbers. Since the synchronous speed of the
induction machine is given by ns = fs /p (in rev./s) where p is the number of pole pairs, this
would correspond to changing the synchronous speed. With the slip now corresponding to
the new synchronous speed, the operating speed is changed. This method of speed control
is a stepped variation and generally restricted to two steps.
If the changes in stator winding connections are made so that the air gap flux remains
constant, then at any winding connection, the same maximum torque is achievable. Such
winding arrangements are therefore referred to as constant-torque connections. If however
such connection changes result in air gap flux changes that are inversely proportional to the
synchronous speeds, then such connections are called constant-horsepower type.
The following figure serves to illustrate the basic principle. Consider a magnetic pole
structure consisting of four pole faces A, B, C, D as shown in fig. 30.
A1
A A2
D B
C1
C
C2
Coils are wound on A & C in the directions shown. The two coils on A & C may be
connected in series in two different ways — A2 may be connected to C1 or C2. A1 with the
other terminal at C then form the terminals of the overall combination. Thus two connections
result as shown in fig. 31 (a) & (b).
Now, for a given direction of current flow at terminal A1, say into terminal A1, the flux
directions within the poles are shown in the figures. In case (a), the flux lines are out of the
pole A (seen from the rotor) for and into pole C, thus establishing a two-pole structure. In
case (b) however, the flux lines are out of the poles in A & C. The flux lines will be then have
to complete the circuit by flowing into the pole structures on the sides. If, when seen from
the rotor, the pole emanating flux lines is considered as north pole and the pole into which
35
i i
T1 T1
A A
D B D B
T2
C T2 C
(a) (b)
A1
i A2
A
T1
T2
C1
C
C2
(c)
they enter is termed as south, then the pole configurations produced by these connections is
a two-pole arrangement in fig. 31(a) and a four-pole arrangement in fig. 31(b).
Thus by changing the terminal connections we get either a two pole air-gap field or a four-
pole field. In an induction machine this would correspond to a synchronous speed reduction
in half from case (a) to case (b). Further note that irrespective of the connection, the applied
voltage is balanced by the series addition of induced emfs in two coils. Therefore the air-gap
flux in both cases is the same. Cases (a) and (b) therefore form a pair of constant torque
connections.
Consider, on the other hand a connection as shown in the fig. 31(c). The terminals T1
and T2 are where the input excitation is given. Note that current direction in the coils now
resembles that of case (b), and hence this would result in a four-pole structure. However,
in fig. 31(c), there is only one coil induced emf to balance the applied voltage. Therefore
flux in case (c) would therefore be halved compared to that of case (b) (or case (a), for that
36
matter). Cases (a) and (c) therefore form a pair of constant horse-power connections.
It is important to note that in generating a different pole numbers, the current through
one coil (out of two, coil C in this case) is reversed. In the case of a three phase machine,
the following example serves to explain this. Let the machine have coils connected as shown
[C1 − C6 ] as shown in fig. 32.
T1
C1 C2
Ta Tb
C3 C5
T1
T2
C4
C6
Tc
The current directions shown in C1 & C2 correspond to the case where T1 , T2 , T3 are
supplied with three phase excitation and Ta , Tb & Tc are shorted to each other (STAR
point). The applied voltage must be balanced by induced emf in one coil only (C1 & C2 are
parallel). If however the excitation is given to Ta , Tb & Tc with T1 , T2 , T3 open, then current
through one of the coils (C1 & C2 ) would reverse. Thus the effective number of poles would
increase, thereby bringing down the speed. The other coils also face similar conditions.
37
The expression for the synchronous speed indicates that by changing the stator frequency also
it can be changed. This can be achieved by using power electronic circuits called inverters
which convert dc to ac of desired frequency. Depending on the type of control scheme of the
inverter, the ac generated may be variable-frequency-fixed-amplitude or variable-frequency-
variable-amplitude type. Power electronic control achieves smooth variation of voltage and
frequency of the ac output. This when fed to the machine is capable of running at a controlled
speed. However, consider the equation for the induced emf in the induction machine.
V = 4.44Nφm f (23)
where N is the number of the turns per phase, φm is the peak flux in the air gap and f is
the frequency. Note that in order to reduce the speed, frequency has to be reduced. If the
frequency is reduced while the voltage is kept constant, thereby requiring the amplitude of
induced emf to remain the same, flux has to increase. This is not advisable since the machine
likely to enter deep saturation. If this is to be avoided, then flux level must be maintained
constant which implies that voltage must be reduced along with frequency. The ratio is held
constant in order to maintain the flux level for maximum torque capability.
Actually, it is the voltage across the magnetizing branch of the exact equivalent circuit
that must be maintained constant, for it is that which determines the induced emf. Under
conditions where the stator voltage drop is negligible compared the applied voltage, eqn. 23
is valid.
In this mode of operation, the voltage across the magnetizing inductance in the ’exact’
equivalent circuit reduces in amplitude with reduction in frequency and so does the inductive
reactance. This implies that the current through the inductance and the flux in the machine
remains constant. The speed torque characteristics at any frequency may be estimated as
before. There is one curve for every excitation frequency considered corresponding to every
value of synchronous speed. The curves are shown below. It may be seen that the maximum
torque remains constant.
This may be seen mathematically as follows. If E is the voltage across the magnetizing
branch and f is the frequency of excitation, then E = kf , where k is the constant of
proportionality. If ω = 2πf , the developed torque is given by
′
k2 f 2 Rr
TE/f = 2 (24)
R′r ′ sω
s
+ (ωLlr )2
38
E/f constant
90
80
70
60 60 Hz
54 Hz
Torque, Nm
50
48 Hz
40
30 Hz
30
15 Hz
20
10
0
0 200 400 600 800 1000 1200 1400 1600 1800
speed, rpm
If this equation is differentiated with repsect to s and equated to zero to find the slip at
′ ′
maximum torque ŝ, we get ŝ = ±Rr /(ωLlr ). The maximum torque is obtained by substitut-
ing this value into eqn. 24.
k2
T̂E/f = ′ (25)
8π 2 Llr
Equation 25 shows that this maximum value is indepedent of the frequency. Further ŝω
is independent of frequency. This means that the maximum torque always occurs at a speed
lower than synchronous speed by a fixed difference, independent of frequency. The overall
effect is an apparent shift of the torque-speed characteristic as shown in fig. 33.
Though this is the aim, E is an internal voltage which is not accessible. It is only the
terminal voltage V which we have access to and can control. For a fixed V , E changes with
operating slip (rotor branch impedance changes) and further due to the stator impedance
drop. Thus if we approximate E/f as V /f , the resulting torque-speed characteristic shown
in fig. 34 is far from desirable.
At low frequencies and hence low voltages the curves show a considerable reduction in
peak torque. At low frequencies ( and hence at low voltages) the drop across the stator
impedance prevents sufficient voltage availability. Therefore, in order to maintain sufficient
torque at low frequencies, a voltage more than proportional needs to be given at low speeds.
39
v/f constant
80
70
60
50
Torque, Nm
48 Hz 54 Hz
40
30
20 30 Hz
15 Hz
60 Hz
10
0
0 200 400 600 800 1000 1200 1400 1600 1800
speed, rpm
Another component of compensation that needs to be given is due to operating slip. With
these two components, therefore, the ratio of applied voltage to frequency is not a constant
but is a curve such as that shown in fig. 35
With this kind of control, it is possible to get a good starting torque and steady state
performance. However, under dynamic conditions, this control is insufficient. Advanced
control techniques such as field- oriented control (vector control) or direct torque control
(DTC) are necessary.
40
0.9
0.8
0.7
voltage boost
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
fraction of rated speed
41
The first non-ideality is the presence of harmonics in the input supply given to the three
phase machine. The source may contain 3rd , 5th , 7th . . . harmonics. Note that due to the
symmetry of the waveform (f (t) = −f (t + T /2), where T is the period of the supply sine
waveform, even ordered harmonics cannot exist. Let the R phase supply voltage be given by
the expression
vR = V1m sin(ω1 t + φ1 ) + V3m sin(3ω1 t + φ3 )
+V5m sin(5ω1 t + φ5 ) + V7m sin(7ω1 t + φ7 ) + − − − (25)
Being a balanced three phase supply, we know that the waveforms of vY and vB are 120◦
and 240◦ shifted from vR respectively. It is further well known that if a waveform is shifted
by φ degrees, its harmonics are shifted by nφ degrees, where n is the order of the harmonic.
Thus the expressions for vY and vB would be
2π 2π
vY = V1m sin(ω1 t + φ1 − ) + V3m sin(3ω1 t + φ3 − 3. )
3 3
2π 2π
+V5m sin(5ω1 t + φ5 − 5. ) + V7m sin(7ω1t + φ7 − 7. ) + − − − (26)
3 3
4π 4π
vB = V1m sin(ω1 t + φ1 − ) + V3m sin(3ω1 t + φ3 − 3. )
3 3
4π 4π
V5m sin(5ω1 t + φ5 − 5. ) + V7m sin(7ω1 t + φ7 − 7. ) + − − − (27)
3 3
If we consider the third harmonic components of the three phase waveforms, and if vx3 (t)
is the third harmonic of phase x, we can see that
42
Therefore, all the three third harmonics are in phase. In a STAR connected system with
isolated neutral, these voltages cannot cause any current flow since all three terminals are
equal in potential. If the neutral point is connected to some point, then then current can flow
through the neutral connection. Such a connection is however rare in induction machines.
The machine is therefore an open circuit to third harmonics. In fact, one can see that any
harmonic whose order is a multiple of three, i.e., the triplen harmonics, as they are called,
will face an identical situation. Since the machine is an open circuit to triplen harmonics in
the excitation voltage, these do not have effect on the machine.
Let us now consider the fifth harmonic. From the equations above, one can see that
From eqns. 29 we see that the fifth harmonic of the excitation forms a negative sequence
system — B phase lags R by 120◦ and Y phase lags R by 120◦.
The MMF caused by a negative sequence excitation causes backward revolving flux pat-
tern (compared to the direction of the fundamental). The torque which it generates will act
as an opposing torque to that generated by the fundamental.
Looking at the seventh harmonic, we can see that
43
From eqns. 30, it is evident that the seventh harmonic components of the excitation form
a positive sequence system. The torque produced by these currents will therefore be additive
with respect to the fundamental component’s torque.
The actual effect of these harmonics on the induction machine would depend on the reac-
tance of the machine since at high frequencies, it is the reactance component that dominates
the inductance. Excitation voltage waveforms with considerable harmonic content may re-
sult when induction machines are controlled through inverters. Apart from the effects on
torque, these harmonics cause considerable heating in the machine and are hence a cause for
concern. These harmonics are called time harmonics since they are generated by a source
that varies non-sinusoidally in time.
Apart from this, there is another kind of harmonic generated in machines called space har-
monic. To understand that this behaviour, it is necessary to consider MMF/flux production
in the machine. It may not be out of place to recall once again that in all our foregoing
analysis we have assumed that both air-gap mmf and flux are sinusoidally distributed in
space.
Let us consider a single full-pitched coil excited by a sinusoidal voltage. The current
flowing through it is sinusoidal and hence the time variation of the mmf produced by it is
sinusoidal. But if we travel around the span of the coil, the MMF variation that we would
encounter is square. It is the amplitude of this square wave that varies sinusoidally in time.
The behaviour is depicted diagrammatically in fig. 36.
Let one more coil be connected in series to this, which is spatially displaced by the slot
angle. This is shown in fig. 37. The same current passes through both and hence the mmf
pattern generated by both would vary in tandem. However, they will be displaced by a
slot angle as far as spatial distribution is concerned. The resulting situation is also shown
in fig. 37 at a particular time instant. It can be seen that the resultant mmf waveform a
non-sinusoidal function of the space angle θ. The harmonics are functions of the space angle.
These are called space harmonics. Let us consider the effects of these.
Let the net flux waveform as a function of angle at an instant of time when unit current
flows in the coils be described by f (θ). Clearly f (θ) is a periodic function of θ with a period
equal to 2π. Therefore one may express this as a fourier series. If fA (θ) is the distribution
function for phase A,
44
1.5
0.5
current, A
MMF
0
90o 180o 270o 360o θ
−0.5
−1
−1.5
0 0.0025 0.005 0.01 0.0125 0.015 0.0175 0.02
time, s
(a) (b)
θ=90 ο
ο
θ=0
(c)
The distribution functions for phases B & C will be displaced from that of A phase by
120◦ and 240◦ respectively and hence are given by
2π 4π
fB (θ) = A1 sin(θ + φ1 − ) + A3 sin(3θ + φ3 ) + A5 sin(5θ + φ5 − ) + − − − (32)
3 3
4π 2π
fC (θ) = A1 sin(θ + φ1 − ) + A3 sin(3θ + φ3 ) + A5 sin(5θ + φ5 − ) + − − − (33)
3 3
Note that we have written these at a given instant of time when unit current flows. We
know that this current variation is sinusoidal in time. Considering the fifth harmonic, let
the resultant fifth harmonic variation is given by,
45
Consider the behaviour of this function. At t=0, the function a value of − 32 A5m at θ
=0. Now let ωt = π3 . At this instant, we find that the function reaches a value − 32 A5m at
π
θ = − 3∗5 . In other words the function f5 (t, θ) has shifted by an angle which is a fifth of the
value of ωt, in the negative direction. The fifth harmonic therefore rotates opposite to the
direction of the fundamental at a speed which is one-fifth of the fundamental.
Similarly, if we consider the seventh harmonic, it can be shown that the resultant distri-
bution is
3
f7 (t, θ) = − A7m cos(ωt − 7θ) (36)
2
By similar arguments as above we conclude that the seventh space harmonic rotates in the
same direction as that of the fundamental at one seventh the speed. In general , we may have
harmonics of the order h = 6n±1, where n is an integer greater than or equal to 1. Harmonics
orders generated by the addition operation move in the same direction as the fundamental
and those generated by the subtraction operation move in the opposite direction. The speed
of rotation is ω1 /h, where ω1 , is the synchronous speed of the fundamental.
46
The space harmonics, it may be noticed are a result of non-sinusoidal distribution of the
coils in the machine and slotting. These have their effects on the speed torque current of
the machine. An example speed-torque characteristic of an induction machine is compared
with its ideal characteristic in fig. 38. The effect of 5th , 7th , 11th and 13th harmonics have
been considered. It can be seen that these harmonics result in kinks in the speed-torque
characteristic near starting region.
80
without harmonic
60 11th with harmonic
40 5th
20
Torque, Nm
−20 7 th
13th
−40
−60
−80
−100
−120
−2000 −1000 0 1000 2000 3000 4000
speed, rpm
To understand the effect of these kinks, consider fig. 39, which shows the same speed
torque characteristic in the motoring region. A load characteristic is also shown, which
intersects the machine characteristic at various points. Note that point 1 is stable and hence
the machine would have a tendency to operate there, though the intended operating point
might be point 5. This tendency is referred to as crawling. A momentary reduction in load
torque conditions might accelerate the machine to print 2, which is unstable. The operating
point would then settle down at 3. The intended operating point may be reached if favourable
torque variations are there.
47
80
line 1
70
60
50 5
Torque, Nm
40
30
20
3
1 2 4
10
−10
0 200 400 600 800 1000 1200 1400 1600 1800
speed, rpm ‘
48