Feeder Analysis & Modeling

1
 Part I
Feeder Modeling
INTRODUCTION

3
 Why Model Distribution Feeders?

 Planning for adequate system capacity

– New feeders and substations require significant
amounts of capital (millions to tens of millions)
– New substations are multi-year projects that require
well-executed project management
 Integration of distributed generation and bi-
directional power flow
 Integration of more complex control systems
– Individual homes responding to load curtailment signals
– Voltage and VAR control
– Substation and feeder device automation
– Advanced and adaptive protection elements

4
What Should a Model Be Able to Tell Us?

Whether:
System voltages, currents, and harmonics
are within established limits
Adequate system resources and
infrastructure are available to meet peak
demand and/or generation
The system is safe, secure, and reliable

5
Transmission Lines Vs. Distribution Lines

6
Transmission Lines
 Phase balance
– Lines has mutual influence over each others.
– Three phase lines will have mutual impedances between each phase.
– The mutual coupling between phases is different because the spacing
between phases are different
– For longer power lines without branches, wires are transposed
according to a transposing scheme as shown in the figure.
– Loads are balanced.
– Accordingly, Transmission studies can assume a balanced system
 R/X ratio
Transmission lines have small R/X ratio which leads to normally neglecting
the line resistance for power flow studies

7
 Distribution Feeders
 Phase balance
– Distribution lines are not transposed
– Distribution lines may be three-phase, two-phase
and single-phase
– Loading is not balanced
– Distribution system modelling must be done by
phase as distribution system are normally
unbalanced
 R/X ratios
Distribution feeders have a relatively higher R/X ratio
compared to transmission lines. Therefore, wire
resistance are not neglected for distribution feeder
modelling
8
 Why phase unbalance is studied?
Confirm that voltage unbalance is within
specifications.
Determine how to better balance the
loads, so no phase is heavily loaded
more than the others.
Reduce losses due to neutral currents.

9
 Overhead vs. Underground
Cost: Overhead’s number one Safety: Underground’s number
advantage. Significantly less cost, one advantage. Much less
especially initial cost. chance for public contact.

Longer life: 30 to 70 years vs. 20 to O&M: Notably lower

40 for new underground works. maintenance costs

Reliability: Shorter outage Reliability: Significantly fewer

durations because of faster fault short and long-duration
finding and faster repair. interruptions.

10
Underground Cables

11
Overhead vs. Underground

12
UNBALANCED DISTRIBUTION
FEEDER MODELLING
 Inductance of unequal phase spacing,
transposed lines
 The phase inductance of un-equal phase spacing and transposed lines is:
−7
Deq
L = 2 10 ln H/m
GMR
where : D eq = 3 D12 D23 D31

Deq
x = L = 2 (60)  2  10 ln−7
/m
GMR
Hence the phase
Deq
reactance will be: = 753.6  10 −7 ln /m
GMR
Deq Deq
x = 1609.3  753.6 10 −7 ln = 0.12134 ln /mile
GMR GMR

^ Deq
And the impedance: z = r + j 0.12134 ln /mile
GMR
 Modeling of Distribution Feeders
 The loading of a distribution feeder is inherently
unbalanced because of the large number of
unequal single-phase loads that must be served.
 An additional unbalance is introduced by the
unequal conductor spacing of the three-phase
overhead and underground line segments.
 If a distribution engineer is to be able to perform
accurate power-flow and short-circuit studies, it is
imperative that the distribution feeder be modeled
as accurately as possible.
 This means that three-phase models of the major
components must be utilized.
 Inductance of unequal phase spacing,
un-transposed lines
 The un-equal phase spacing and un-transposed lines will result in self and
mutual inductances between the line which can be written as
1
Ẑii = ri + j 0.12134 ln  / mile
GMR
1
Zˆ ij = j 0.12134 ln  / mile
Dij
 Zii is the self inductance of conductor i due to its current and Zij is the
mutual inductance of conductor i due to the current in conductor j.
 Inductance of unequal phase spacing,
un-transposed lines

 In a classic paper, John Carson developed a technique in

1926 whereby the self and mutual impedances for n
overhead conductors can be determined.
 Carson modified the previous equations with different
approximations till he reached the following equation:
Modified Carson’s equations:
^
 1 
zii = ri + 0.0953 + j 0.12134 ln + 7.93402  /mile
 GMR 
^  1 

zij = 0.0953 + j 0.12134 ln + 7.93402  /mile
 D 
 ij 
Phase Impedance Matrix for Overhead
Lines
The mutual and self impedance for a three phase feeder with
four wires are shown in the following figure:

Using KVL:     

Vag  V ' ag   z aa z ab z ac z an 
I a 
   '       
Vbg  V bg   zba zbb zbc zbn   I b 
V  = V '  +      . I 
 cg   cg   zca zcb zcc zcn   c 
Vng  V ' ng        I n 
 z na z nb z nc z nn 

primitive impedance matrix

 Phase Impedance Matrix for Overhead
Lines
In partitioned form:
     
 
Vabc   V ' abc 
   zij   zin   I abc 
V  =  ' +  .
 
 ng   V ng      z     n 
 I  
  nj   z nn  
  

Vabc  
=V '
abc   
+  zij .I abc  +  zin .I n 
   

   
0 = 0 +  znj .I abc  +  znn .I n 
   
For a grounded neutral, the voltages Vng and V’ng are equal to zero.
 Phase Impedance Matrix for Overhead
Lines
Solving for In:
−1

I n  = − znn  


. znj .I abc 
   
          −1    
Vabc  = V ' abc  +   zij  −  zin . znn  . znj  .I abc 
       
 

Vabc  = V abc 
'   
+   zabc  .I abc 
 
where

Zˆ 
abc
          −1    
=   zij  −  zin . znn  . znj  
       
 
 Phase Impedance Matrix for Overhead
Lines
 zaa zab zac 

Z abc  =  zba zbb zbc 
 zca zcb zcc 
 For a distribution line that is not transposed, both the diagonal terms
of above equation will not be equal and the off-diagonal elements
will not be equal too. However, the matrix will be symmetric.
 The phase impedance matrix for a open delta connections with
phase a and c only available will be:

 zaa 0 zac 

Z abc  = 0 0 0  
 zca 0 zcc 
 Phase Impedance Matrix for Overhead
Lines
 For The diagram of a three phase system without a neutral wire is as
shown in the following figure:

Vag  Vag   zaa zab zac   I a 

     z z z  I 
 bg 
V =  bg 
V +  ba bb bc   b 
V  V   zca zcb zcc   I c 
 n  m
cg cg

In condensed form this equation can be written as:

VLG abc n = VLG abc m + Z abc I abc 
 Phase Impedance Matrix for Overhead
Lines
Symmetrical Components:
It is the method by which an unbalance system of “n” related
phasors can be resolved into “n” systems of a balance phasors.

For the three-phase power system, symmetrical components convert

unbalanced voltages and currents to three sets of balanced sequence
components: positive sequence, negative sequence, and zero sequence.

Sequence Voltages
Unbalance Voltages Balance
Systems
zero sequence
Vag Zero
Sequence positive sequence
Vcg
Symmetrical Three
Positive
components Sequence balanced
Systems
Vbg
Negative
Sequence

negative sequence
 Phase Impedance Matrix for Overhead
Lines
Sequence Impedance:
Vag  1 1 1  VLG 0 
   2  
 bg 
V = 1 a a VLG1 
V  1 a a2  VLG 2 
 n 
cg 
In condensed form this equation can be written as:
VLG abc  = A. VLG 012 
The phase line currents are defined in the same manner:

I abc  = A. I 012 

The sequence line-to-ground voltages as a function of the phase line-to-
ground voltages :
VLG012  = A .VLGabc 
−1

1 1 1 
where: A−1 = 1 1 a a2 
3
1 a 2 a
 
 Phase Impedance Matrix for Overhead
Lines
Sequence Impedance:

Since: I abc  = A. I 012  and VLG abc n = VLG abc m + Z abc I abc 
VLG012  = A−1.VLGabc 
Then:
VLG012 n = A−1.VLG abc m + A−1.Z abc . A. I 012 

Z 012 
Where:
 Z 00 Z01 Z02  Z00 = zero sequence impedance
Z 012  = Z10 Z11 Z12  Z11 = positive sequence impedance
 Z 20 Z21 Z22 
Z22 = negative sequence impedance
 Example 1
An overhead three-phase distribution line is constructed as
shown in the following figure. Determine the phase impedance
matrix and the positive and zero sequence impedances of the
line. The phase conductors are 336,400 26/7 ACSR and the
neutral conductor is 4/0 6/1 ACSR.
 Example 1, solution

The distance between phases are as follows:

The self and mutual impedances for phase (a) are

 Example 1, solution
The primitive impedance matrix

The primitive impedance matrix in partitioned form is:

 Example 1, solution
The impedance matrix in reduced form

The sequence impedance matrix is:

Clearly the sequence networks are not decoupled.

Z00 Important point to be made here is that, because
Z11 sequence networks are not decoupled, they offer
no real advantage over the abc network in terms
of analysis.
 Medium transmission lines
(for balanced system)
Recall the ABCD model for balanced medium transmission line:

 V Y   YZ 
Vs = VR + Z  I R + R  = 1 + VR + ZI R
 2   2 

VRY VsY
Is = IR + + , subsitute the value of Vs
2 2

 YZ   YZ  YZ
I s = Y 1 + V +
 R 1 + I R A = D = 1+
 4   2  2
B=Z
 YZ 
C = Y 1 + 
 4 
Exact Feeder Model
 Exact Feeder Model (Cont.)

VLG ac n  a  b VLG abc m 

 I   =  
 abc n  c  d   I abc m 


where: a = d  = U  + 1  Z abc  Yabc 

2
b = Z abc 
c = Yabc  + 1  Yabc  Z abc  Yabc 
4

 Z aa Z ab Z ac  Yaa Yab Yac  1 0 0

Z abc  = Z ba Z bb Z bc  Yabc  = Yba Ybb Ybc  U  = 0 1 0
Yca Ycc  0 0 1
 Z ca Z cb Z cc  Ycb
 Exact Feeder Model (Cont.)
The voltage at node m can be found from the voltage at node n

VLG 
abc m  a  b VLG abc n 
−1

 I   =   
  
 abc m   c  d   abc n 
I

Since: a. d  − b. c = U 

VLG abc m   d  − b VLG abc n 

 I   =   
  
 abc m   − c  a    abc n 
I

VLG abc m = a. VLG abc n − b. I abc n

and,
I abc m = −c. VLG abc n + d . I abc n
 Exact Feeder Model (Cont.)
Sometimes it is necessary to compute the voltage at Node m as
a function of the voltages at Node n and the currents entering
Node m.
VLG abc n  a  b VLG abc m 
 I   =   
  
Since:
 abc n   c  d   abc m 
I

Then VLG abc m = a−1VLG abc n − bI abc m 

VLG abc m = A. VLG abc n − [ B].I abc m

Where: A = [a]−1

[ B ] = [a ]−1.[b]
 Exact Feeder Model (Cont.)
♦ Since the mutual coupling between phases on the line
segments are not equal, there will be different values of
voltage drop on each of the three phases.
♦ As a result, the voltages on the distribution feeder become
unbalanced even when the loads are balanced.
♦ The degree of voltage unbalance can be defined according
to the following formula:

Maximum deviation from average

Vunbalance = .100%
Vaverage
 Example 2
A balanced three phase load of 6000 kVA, 12.47 kV, 0.9 lagging
power factor is being served at Node m of a 10,000 ft. Both the
phase impedance matrix and shunt admittance matrix for the
line are given below. Determine the generalized line constant
matrices, [a], [b], [c] and [d]. Using the generalized matrices
determine the line-to-ground voltages and line currents at the
source end of the line segment.
 Example 2, solution
The 10,000 ft line segment impedance and admittance matrices are as
follows:

The generalized constant matrices are as follows:

Although the diagonal elements

of matrix a is not exactly 1, it is
very close to one, for example:
 Example 2, solution
This example demonstrates that the phase admittance matrix can be
neglected.
The magnitude of the line-ground voltage at the load is

12470
VLG = = 7199.56
3

The magnitude of the load current is

6000
I m
= = 277.79
3.12.47

7538.71.57 
VLG abc n = [a].VLG abc m + [b].I abc m = 7451.25 − 118.3
7485121.93 
 Example 2, solution
Calculation of the voltage unbalance:

The voltage in per unit

The currents at node n will be:

 Short transmission lines
(for balanced system)
Is RLine XLine
IR

Vs VR
Z

Vs = VR + ZI R A = D =1
Is = IR B=Z
C =0
Modified Feeder Model
 Modified Feeder Model (Cont.)

VLG abc n  a  b VLG abc m 

 I   =    
 abc n  c  d   I abc m 
where:
a = d  = U  +  Z abc  Yabc 
1
2
a = d  = U 
b = Z abc  b = Z abc 
c = Yabc  + 1  Yabc  Z abc  Yabc 
4
c = 0
 Modified Feeder Model (Cont.)
If the line is a three-wire delta, the voltage drops down the line must be in
terms of the line-line voltages and line currents.
Vab  Vab  Vdrop _ a  Vdrop _ b 
V  = V  + V   
 bc   bc   drop _ b  − Vdrop _ c 
Vca  n Vca  m Vdrop _ c  Vdrop _ a 
   
Vdrop _ a   Z aa Z ab Z ac   I line _ a 
where      
Vdrop _ b  =  Z ba Z bb Z bc . I line _ b 
V  Z Z cc   I line _ c 
 drop _ c   ca
Z cb
 

Vab _ n = Vab _ m + Vdrop _ a − Vdrop _ b Vab _ n = Van _ n − Vbn _ n

but,
Van _ n − Vbn _ n = Van _ m − Vbn _ m + Vdrop _ a − Vdrop _ b Vab _ m = Van _ m − Vbn _ m

Van _ n = Van _ m + Vdrop _ a

Vbn _ n = Vbn _ m + Vdrop _ b
 Example 3
The line of the previous example will be used to supply an unbalanced load
at Node m. Assume the voltages at the source end are balanced three phase
at 12.47 kV line-line. The balanced line-ground voltages are:

7199.560 
VLG abc n = 7199.56 − 120
7199.56120 

The unbalanced currents measured at the source end are given by:
 I a  249.97 − 24.5 
 I  = 277.56 − 145.8
 b  
 I c  n 305.5495.2 

Determine the line-ground and line-line voltages at the load using the
modified line model. Also, find the complex power of the load.
 Example 3, solution
The line-ground voltages at the load end are:
VLG abc m = AVLG abc n − [ B]I abc m

Since the source currents are equal to the load currents (why):
 Example 3, solution

The line-line voltages at the load end can be computed by:

 V .I
*
 1735.4223.03
S a  
ag a
  
 S  = 1 V .I *  = 1920.2624.25  kVA
 b  1000  bg b   
 S c  Vcg .I c  2104.4722.11
*
Cable structure

Fig. 1: single phase cable

Cable History
Series Impedance of Underground
Lines
 Two types of underground cables exist:
– Concentric Neutral Cables

 Underground residential distribution normally

has concentric-neutral cables.
 Concentric-neutral cables are also used for
three-phase mainline applications and three-
phase power delivery to commercial and
industrial customers.
 Series Impedance of Underground
Lines
– Tape Shielded Cables

 It is used for high current, three-phase applications.

 Other applications include: Three-conductor power
cables, fire-resistant cables, extra flexible cables, and
submarine cables.
 Concentric Neutral Underground Cables
 The geometric mean radius of the concentric neutral underground
cable can be expressed by
Phase Conductor

GMRcn = k GMRS . k . R k −1 Insulation

Jacket
R
d od − d S
R= ft d od dc Concentric Neutral Strand
24
Insulation Screen
rs ds
rcn = /mile
k
R

where,
GMRS is the geometric mean radius of the solid concentric neutral strands,
k the number of the concentric neutral strands,
R the radius of the circle passing through the center of the concentric neutral Strands,
dod the outside diameter over the concentric neutral strands in inch,
dS the diameter of the concentric neutral strand in inch,
rcn equivalent resistance of the concentric neutral
rs resistance of single neutral strand
Concentric Neutral Underground Cables

 The spacing distances of the cables configuration can be expressed by

– Spacing between concentric neutral and its own phase conductor: Dij = R
– Spacing between concentric neutral to an adjacent concentric neutral: Dij = Dnm
– Spacing between concentric neutral to an adjacent phase conductor:

Dij = k Dnm
k
− Rk

Where: Dnm is the center to center distance between the phases conductors.

If the distance Dnm is much larger than R, then Dij can be assumed to be Dnm
Then we use the modified Carson’s equations:
^
 1 
zii = ri + 0.0953 + j 0.12134 ln + 7.93402  /mile
 GMR 
^  1 

zij = 0.0953 + j 0.12134 ln + 7.93402  /mile
 D 
 ij 
 Example 4
 Three concentric neutral cables are buried in a trench
with spacing as shown in the following figure. The
cables are 15 kV, 250,000 CM stranded AA with 13
strands of #14 annealed, copper wire neutral. The
outside diameter of the cable over the neutral strands
is 1.29 inches. Determine the phase impedance matrix
and the sequence impedance matrix.
 Example 4, Solution
4 5 6

R R R

1 2 3

From the cable appendix table:

250,000 AA phase conductor:
GMR = 0.0171 ft., Diameter = 0.567”, Resistance = 0.41 /mile
# 14 copper neutral strands:
GMR = 0.00208 ft., Resistance = 14.8722 /mile, and Diameter = 0.0641”

d od − d S 1.29 − 0.0641
R= = = 0.0511 ft.
24 24

GMRcn = k GMRS . k . R k −1 = 13 (0.00208)(

. 13)(
. 0.0511)
13 −1
= 0.0486 ft

rS 14.8722
rcn = = = 1.144 /mile.
k 13
 Example 4, Solution
 The phase conductors are numbered 1, 2 and 3. The concentric
neutrals are numbered 4, 5 and 6. The conductor-conductor and
concentric neutral-concentric neutral spacing are:
D12 = D21 = D45 = D54 = 0.5 ft D23 = D32 = D56 = D65 = 0.5 ft

D31 = D13 = D64 = D46 = 1.0 ft

4 5 6

R R R

1 2 3

The spacing between conductors and their concentric neutrals

are:
D14 = D25 = D36 = R = 0.0511 ft
 Example 4, Solution
 Since the radius R is much smaller than the spacing between
cables, the distances between concentric neutrals and adjacent
phase conductors are just the center-center distances between
conductors
D15 = D51 = 0.5 ft 4 5 6

R R R

D26 = D62 = 0.5 ft 1 2 3

D61 = D16 = 1.0 ft

The self impedance

^
 1 
zii = ri + 0.0953 + j 0.12134 ln + 7.93402  /mile
 GMR 
^
 1 
z11 = 0.41 + 0.0953 + j 0.12134 ln + 7.93402  = 0.505 + j1.456 /mile
 0.0171 
 Example 4, Solution
The self impedance of the concentric neutral
^
 1 
z 44 = 1.144 + 0.0953 + j 0.12134 ln + 7.93402  = 1.2393 + j1.3296 /mile
 0.0486 

The mutual impedance between cable # 1 and cable # 2 is:

^
 1 
z12 = 0.0953 + j 0.12134 ln + 7.93402  = 0.0953 + j1.0468 /mile
 0.5 
The mutual impedance between cable # 1 and its concentric neutral
^
 1 
z14 = 0.0953 + j 0.12134 ln + 7.93402  = 0.0953 + j1.3236 /mile
 0.0511 
The mutual impedance between concentric neutral of cable # 1 and
concentric neutral of cable # 2
^
 1 
z 45 = 0.0953 + j 0.12134 ln + 7.93402  = 0.0953 + j1.0468 /mile
 0.5 
Example 4, Solution
  Zˆ 
 Zˆ ij
Zˆ = 
  Zˆ 
in

 Zˆ nj nn

[Znj] = [Zin]T
Example 4, Solution
 Tape Shielded Underground Cables
 The geometric mean radius of the tape shielded underground
cable can be expressed by
T
ds −
GMRshield = 1000 ft.
24

rshield = 7.9385  108 /mile
d s .T

Where:
dc the outside diameter of phase conductor in inch,
dS the outside diameter of the tape shield in inch,
T Thickness of copper tape shield in mils
rsheild resistance of tape shielded
 Example 5
 A single phase circuit consists of a 1/0 AA, 220 mil
insulation tape shielded cable and a 1/0 Cu neutral
conductor as shown below. The single phase line is
connected to phase b. Determine the phase impedance
matrix. Assume  = 2.3715 X 10-8 -meter
 Example 5, solution
 Cable data: Outside of the tape shield = 0.88 inches,
R = 0.97 /mile, GMRp = 0.0111 ft., Tape shield
thickness = 5 mils.
 Neutral data: Resistance = 0.67 /mile and GMRn =
0.01113 ft.,

rshield = 7.9385  10 8
/mile
d s .T
−8
2.3715  10
rshield = 7.9385 108 = 4.2785 /mile
0.88.5
T 5
ds − 0.88 -
GMRshield = 1000 = 1000 = 0.0365 ft.
24 24
 Example 5, solution
 The conductors are numbered such that:
– #1 = 1/0 AA conductor
– # 2 = tape shield
– # 3 = 1/0 copper ground

D11 = GMRP = 0.0111

 1 
z11 = 0.0953 + 0.97 + j 0.12134 ln + 7.93402  = 1.065 + j1.509 / mile
 0.0111 
D22 = GMRShield = 0.0365

 1 
z 22 = 0.0953 + 4.2786 + j 0.12134 ln + 7.93402  = 4.374 + j1.365 / mile
 0.0365 
 Example 5, solution

D12 = GMRShield = 0.0365

 1 
z12 = 0.0953 + j 0.12134 ln + 7.93402  = 0.0953 + j1.365 / mile
 0.0365 
 Cable Ampacity
 The national electric code NEC has also defined
ampacity as “the current in amperes a conductor can
carry continuously under the conditions of use
without exceeding its temperature rating”.

 From the previous definitions it is clear that

ampacity and temperature are related to each other.

 In order for us to understand ampacity we must

understand the thermal behavior of a cable carrying
current and its surrounding environment
 Cable Ampacity
 The temperature rise in a cable can be calculated
using the following formula:

 = W  T
where:  is the temperature difference across the material (like potential),
W is the flow of heat (like current) and T is the thermal resistance (like
electrical resistance)

 Sources of heat in cables:

1. Conductor Losses
2. Dielectric Losses
3. Sheath Losses
 Thermal Resistance
 A heat flow starts from a heat source through a thermal
resistance to the surrounding atmosphere of a lower
temperature.
 Thermal resistances are either internal or external.
 Internal thermal resistance are considered to be all those
non-metallic layers of cable such as insulation material,
conductor shield, insulation shield, etc.
 External thermal resistance is usually made up from the
surrounding environment of the cable such as soil thermal
resistance when a cable is buried underground, and the
border between the cable surface and air when the cable is
laid in free air.
Effect of Soil Thermal Resistivity and
Ground Temperature
Effect of Number of Circuits
 Overhead vs. Underground
Cost: Overhead’s number one Aesthetics: Underground’s
advantage. Significantly less cost, number one advantage. Much
especially initial cost. less visual clutter.
Longer life: 30 to 50 years vs. 20 to Safety: Less chance for public
40 for new underground works. contact.

Reliability: Shorter outage Reliability: Significantly fewer

durations because of faster fault short and long-duration
finding and faster repair. interruptions.
Loading: Overhead circuits can O&M: Notably lower
more readily withstand overloads. maintenance costs
Overhead vs. Underground
 Part II
Feeder Analysis and
Examples

20-Mar-21 73
 Introduction
 By analyzing a three phase feeder supplying a load,
we will be able to compute the voltages at the
receiving end (the customer) given the voltage at the
sending end (the substation) and given the loads.
 The sending end voltage at the substation is normally
controllable (and therefore known) through the tap
setting of the transformer or through the voltage
regulator.
 The receiving end voltage at the customer is normally
not controllable, although it is important since the
customer appliances require that the voltage
magnitude lie within a certain range in order to
properly operate.
 Voltage Drop Calculation for Primary
Circuit for Radial System

20-Mar-21 75
 Example 6

 Find the total voltage drop for the following single

phase feeder which has a total length of 500 m and
total Z = 0.02+j0.04  and the voltage at point A
is 250V

76
 Example 6, solution
Current in section AD is the vector sum of the three
load currents:
I AD = 50 + 100(0.8 − j 0.6) + 50(0.6 − j 0.8) = 160 − j100 A

The impedance of section AD is:

200
Z AD =  (0.02 + j0.4) = 0.008 + j0.016 
500
The voltage drop of section AD is:
VAD = (160 − j100)  (0.008 + j0.016) = 2.88 + j1.76 V

Current in section DC is :
I DC = 160 − j100 − 50 = 110 − j100 A

77
 Example 6, solution
The impedance of section DC is:
100
Z DC =  (0.02 + j0.4) = 0.004 + j0.008 
500

The voltage drop of section DC is:

VDC = (110 − j100)  (0.004 + j0.008) = 1.24 + j0.48 V

Current in section CB is :
I CB = 50(0.6 − j 0.8) = 30 − j 40 A

The impedance of section CB is:

200
Z CB =  (0.02 + j0.4) = 0.008 + j0.016 
500
78
 Example 6, solution

The voltage drop of section CB is:

VCB = (30 − j 40)  (0.008 + j0.016) = 0.88 + j0.16 V

The total voltage drop is:

Vdrop = 5 + j2.4 = 5.5525.6 V

79
 Voltage Drop Calculation for a Ring
System

20-Mar-21 80
 Example 7

 A single phase ring distributer ABC is fed at A. The loads

at B and C are 20 A at 0.8 p.f. lagging and 15 A at 0.6 p.f.
lagging respectively. The total impedance of each line as
shown below, find the current at each segment.

A
1+j1
1+j3

B C
1+j2
20 A, 0.8 15 A, 0.6
p.f. p.f.

20-Mar-21 81
 Example 7, solution

 The same procedure followed in example 4 will be

followed here except the total voltage drop around the ring
will equal to zero.
 After solving the equations, the current that is flowing in
section AB will be 18.6-j13.53 A

20-Mar-21 82
Voltage Drop Calculation for Primary
Circuit Using K-factor
Example 8
 A primary feeder system is composed of a primary feeder and two
laterals as shown below. Each lateral is connected to a uniformly
distributed load with total kVA of 518 kVA. The maximum allowed
voltage drop is 4%. The voltage level of the main and laterals are
2400/4160 V. The distance between phases is 37 inches. Use ampacity
sized conductors but in no case smaller than AWG # 6 for reasons of
mechanical strength. Determine:
 The %VD at the end of laterals and main
 If 4% max VD criterion is exceeded, find a reasonable combination of
larger conductors for both the main and the laterals that will meet the
VD criterion
 Example 8, solution
The current in the lateral is: I 518
lateral = = 72 A
3  4.16
From table A1, AWG # 6 copper conductor with 130 A
ampacity is selected for the lateral.
1036
I main = = 144 A
The current in the main is: 3  4.16

From table A1, AWG # 4 copper conductor with 180 A

ampacity is selected for the main.

The K factor of both conductors can be calculated using

the following figure:

20-Mar-21 84
 Example 8, solution

From the figure, K(lateral) = 0.015 and K(main) = 0.01

1
%VDlateral =  K  kVA  L =
2
1 5760
 0.015  518  = 4.24%
2 5280

%VDmain = K  kVA  L =
3300
0.01 1036  = 6.48%
5280

85
 Example 8, solution

So the total voltage drop will be:

%VD = %VDlateral + %VDmain = 10.72%

And hence the voltage drop is more than the required value.

So there is a need to increase the cable size to reduce the voltage

drop and then re-calculate the voltage drop.

For this particular example, 4/0 and AWG#1 copper conductors

with ampacities of 480 A and 270 A for the main and lateral
respectively will satisfy the design needs.
 Secondary Distribution System
Example
 For the residential secondary distribution system shown in the Figure
Assume the following:

– The distribution transformer capacity is 75 kVA,

– All secondaries and services are single-phase, three wire, normally 120 /
240 V,
– All secondary lines are of # 2/0 AWG Al/Cu XLPE cable,
– Service drops are of # 1/0 AWG Al/Cu XLPE cable,
– All service drops are 100 ft long,
– All secondary lines are 200 ft long,
– Average lagging load power factor of 0.9.

 Determine:
– The total load on the transformer in kVA and in pu.
– The total steady-state voltage drop in pu at the most remote and severe
customer’s meter for the given annual maximum system load.

20-Mar-21 87
 Example (Cont’d)

20-Mar-21 88
 Table A: Single-phase 7200-120/240 V distribution
transformer data at 65C
Rated Core Loss Copper Loss R X Excitation
kVA kW kW pu pu Current
A

15 0.083 0.194 0.013 0.0094 0.014

25 0.115 0.309 0.0123 0.0138 0.015

37.5 0.17 0.4 0.0107 0.0126 0.014

50 0.178 0.537 0.0107 0.0139 0.014

75 0.28 0.755 0.0101 0.0143 0.014

100 0.335 0.975 0.0098 0.0145 0.014

20-Mar-21 89
 Table B: Twin concentric Al/Cu XLPE 600 V cable
data
~
K*
R(/1000 ft) per conductor X(/1000 Per unit voltage
Direct-burial drop per 104 A.ft
ft)
Size Ampacity
Phase Neutral per phase
A 90 % 50 %
conductor Conductor conductor
PF PF
90C 80C
2 AWG 0.334 0.561 0.0299 180 0.02613 0.01608
1 AWG 0.265 0.419 0.305 205 0.02098 0.01324
1/0 AWG 0.21 0.337 0.0297 230 0.01683 0.01089
2/0 AWG 0.167 0.259 0.029 265 0.0136 0.00905
3/0 AWG 0.132 0.211 0.028 300 0.01092 0.00752
4/0 AWG 0.105 0.168 0.0275 340 0.00888 0.00636
250 kcmil 0.089 0.133 0.028 370 0.00769 0.00573
350 kcmil 0.063 0.085 0.027 445 0.00571 0.00458
500 kcmil 0.044 0.066 0.026 540 0.00424 0.00371

20-Mar-21 90
 Solution
 Assume perfectly balanced loading in all three-wires single-phase circuits,
and nominal operating voltage of 240 V when computing currents from kVA
loads.

 The total load of the transformer is

ST = (3+2+8+6) + (5+6+7+4) + (6+7+8+10) = 72 kVA.
ST, pu = ST / Sbase = 72 / 75 = 0.96 pu.

 To calculate the total steady-state voltage drop in pu at the most remote and
service customer’s meter. We have to calculate the per unit voltage drops in
the transformer, the service line, and that of the most remote and severe
customer. This is carried out as follows:
For the transformer:
VDTr = ZI = (0.0101+j0.0143)*0.96 angle(-25.8)
= 0.0168 angle(28.9o)

Where R = 0.0101 pu and X = 0.0143 pu are found from Table A.

20-Mar-21 91
 Solution (Cont’d)
– For the service lines:
ISL = SSL, max / Voperating = (6+7+8+10)kVA / 240 V =31 / 240 = 129.17 A.

K= 0.0136 from Table B.

VDSL =K(ISL x l) x 10-4 = 0.0136 (129.17 x 200) x 10-4 = 0.03513 pu.

– For the service drops:

ISD = SSD, max / Voperating = 10 kVA / 240 V =41.67 A.

K = 0.01683 from Table B.

VDSD = K (ISD x l) x 10-4 = 0.01683 (41.67 x 100) x 10-4 = 0.007 pu.

– Total voltage drop:

VDT = VDTr + VDSL +VDSD =0.0168+ 0.03513 + 0.007 = 0.0589 pu.

20-Mar-21 92
 Solution (Cont’d)

It can be noticed from the previous

example is that the voltage was assumed
to be constant along the feeder. i.e.
240V.
This is not correct as there will be
voltage drop in the line.
However, the voltage drop will be very
small and exact calculation will be as in
the coming slide:

20-Mar-21 93
 Solution (Cont’d)
V = 1 – 0.0168 = 0.9832
– For the service lines:
ISL = SSL, max / Voperating = (6+7+8+10)kVA / (0.9832*240) V = 131.3 A.

K = 0.0136 from Table B.

VDSL = (ISL x l) x 10-4 = 0.0136 (131.3 x 200) x 10-4 = 0.03566 pu.

– For the service drops:

V = 0.9853 – .03566 = 0.9496

ISD = SSD, max / Voperating = 10 kVA / (.9496*240) V =43.88 A.

K = 0.01683 from Table B.

VDSD = (ISD x l) x 10-4 = 0.01683 (43.88 x 100) x 10-4 = 0.0074 pu.

– Total voltage drop:
VDT = VDTr + VDSL +VDSD =0.0168 + 0.03566 + 0.0074 = 0.0599 pu.
20-Mar-21 94
 Load Flow Analysis
A single phase example
 A single-phase lateral shown below has a line impedance
of 0.3+j0.6/mile. The loads are, S2=1500kW+j750kVar
and S3=900kW+j500kVar 1 2 3
3000’ 4000’

 The source voltage at node 1 is V1=7200V. Compute the

node voltages at buses 2 and 3 to within a tolerance of
0.001pu.
 Why is this problem not simply an exercise in single phase
circuit analysis?
 A single phase example

 Answer: Because to get V2 and V3, we need

currents I12 and I23. To get the currents, we need to
know the voltages V2 and V3!
 Therefore, we need an iterative technique to solve
for both V2 and V3.
Definitions:
– Vk= voltage at node k
– Ikj = current from node k to node j
– Ik =load current at node k.
 A single phase example-Solution
Technique
 The iterative approach can be applied as follows:
1. Assume node 3 voltage.
2. Forward sweep: Using the assumed or last
computed node 3 voltage, compute the node 1
voltage and node 1-2 current. If computed node 1
voltage is within tolerance of specified node 1
voltage, quit. Otherwise, go to step 2.
3. Backward sweep: Using the specified node 1
voltage & the computed node 1-2 current,
compute the voltage at the loads.
4. Go to 2.
 A single phase example-Solution
 To apply the algorithm, first we must obtain the
impedances from the per-mile impedances.

3000
Z12 = (0.3 + j 0.6) = 0.1705 + j 0.3409
5280
4000
Z 23 = (0.3 + j 0.6) = 0.2273 + j 0.4545
5280
 A single phase example-Solution
 Step 1:
 Assume node 3 voltage is V3=7200/0◦ .V
 Step 2: (Forward sweep)
 From node 3 voltage & node 3 power, we get node 3 load
current from S=VI*.
* *
 S3   (900 + j500) 10 
3
I 3 =   =   = 143 − 29.0 A
 V3   7200 
 The node 2-3 current is equal to the node 3 current.

I 23 = I 3 = 143 − 29.0 A
 A single phase example-Solution
 The node 2 voltage is:
V2 = V3 + I 23 Z 23
= 72000 + 143 − 29.0(0.2273 + j 0.4545)
= 7260.10.32
 Node 2 load current is
* *
 S2   (1500 + j 750)  103 
I 2 =   =   = 231.0 − 26.3
 V2   7260.10.32 
 Node 1-2 current is
I12 = I 2 + I 23
= 231.0 − 26.3 + 143.0 − 29.0
= 373.9 − 27.3
 A single phase example-Solution
 The node 1 voltage is:
V1 = V2 + I12 Z12
= 7260.10.32 + 373.9 − 27.3(0.1705 + j 0.3409)
= 7376.20.97

 Since it is specified in the problem that the source voltage is

7200V, we have an error of
Error=|7200-7376.2|=176.2.

 In per-unit (assuming a 7200 voltage base), we get

Error=176.2/7200=0.00245 pu.
So it is not within tolerance and we continue to step 3.
 A single phase example-Solution
 Step 3: (Backward sweep)
 We assume the voltage magnitude at node 1 is as specified,
7200 V. We will use the node 1 voltage angle computed in
the forward sweep. Therefore, V1=7200/0.97.
 The node 2 voltage is computed using the I12 current
computed in the forward sweep.
V2 = V1 − I12 Z12
= 72000.97 − 373.9 − 27.3(0.1705 + j 0.3409)
= 7084.00.306
 A single phase example-Solution
 The node 2 load current is determined next:
* *
 S   (1500 + j 750) 10 3

I 2 =  2  =   = 236.73 − 26.89
 V2   7084.00.306 
 The node 2-3 current is computed next.
I 23 = I12 − I 2
= 373.9 − 27.3 − 236.73 − 26.89
= 137.19 − 28.01
 The node 3 voltage is then:
V3 = V2 − I 23 Z 23
= 7084.00.306 − 137.19 − 28.01(0.2273 + j 0.4545)
= 7027.10
 A single phase example-Solution
Algorithm
 Step 4: Go to Step 2. Repeat forward sweep.

Key to understanding the forward and backward

sweep method is that
♦ The forward sweep is simply obtaining estimates
of currents in each of the segments. In the forward
sweep, voltage accuracy is not important.
♦ The backward sweep is concerned with accuracy
of resulting voltages.
 A single phase example-Solution
 Comments:
 We could have made backward sweep easier by using all
forward sweep currents. If we were to have done this, the
resulting node 3 voltage would have been 7026.0. If there is
more than one lateral emanating from a node, this is what
you must do because then you will have 2 unknowns in your
KCL equation at that node (one unknown for each lateral)
instead of only 1 as we had above.
 Although use of forward sweep currents in the backward
sweep calculation is an approximation, it is OK since the
tolerance test is on the result of the forward sweep where
exact currents are computed and used.
 Application to three phase
 Some preliminary comments
 Consider making the following assumptions in
regards to three-phase feeder analysis.
– The load is balanced.
– The phases are transposed (each phase occupies the
same physical position on the structure for 1/3 of the
length of the line).
 Application to three phase
 Under the above two assumptions, the self and
mutual inductances of the three conductors can be
represented as three equal self inductances for each
conductor. Because
– the three self inductances are equal and
– there is zero mutual coupling between phases,
 We can analyze such a situation using a per-phase
equivalent circuit
 In this case, the analysis for 3-phase circuits
proceeds as in the single-phase case.
 Application to three phase
 Unfortunately, these assumptions do not hold for
distribution feeders. Distribution circuits consist of single-
phase, two-phase, lines serving unbalanced loads.

 Therefore, for three-phase distribution circuits, it is the

case that the loading is always unbalanced, and
sometimes, the unbalance is fairly large.

 In addition, distribution circuits are usually not

transposed. As a result, the self inductances will not be
equal. In addition, we will have to account for mutual
inductance between conductors (including the neutral).

 Therefore proper mathematical model for the line is

required.
 Application to three phase-Illustration
 For now, in order to illustrate, we omit the calculations
involving the lines.
 The complex powers (per-phase) of a Y-connected load
are:
 2236.126.6 
S abc = 2506.028.6 kVA
 2101.425.3

 Assume that the nominal line-to-line voltage of the feeder,

12.47 kV, is the line-to-line voltage seen at the load.
Compute the currents in all three phases. Also, compute
the neutral current.
|VLN|=12,470/√3=7200
 Application to three phase-Illustration
 Therefore, assuming balanced voltages, we have:
 72000 
VLN ,abc = 7200 − 120
 7200120 
 The currents can be obtained using I=(S/VLN)*, i.e.,
 2236.110 3 26.6 * 
  
 72000  
 *  310.57 − 26.6 
2506.0 10 28.6   
3
I abc =   = 348.06 − 148.6
 7200 − 120    
 *  291.8794.7 
 2101.4 10 25.3  
3

    Note that the phase

  7200 120   currents are NOT equal!
 Application to three phase-Illustration
 What does this mean with respect to the neutral
current?
Ia + Ib + Ic =
310.57 − 26.6 + 348.06 − 148.6 + 291.8794.7
= 52.4 − 145.73

 So under unbalanced conditions, the neutral

conductor carries current, and this current is non-
negligible.
 Approach for feeders with laterals
 Consider the following distribution feeder that serves loads
using several different laterals.
1 2 3

 In this case, node 1 represents the substation. Node 3 is the

far-end of the feeder, and branch 2-4 represents a lateral
serving load at its far-end node 4.
 Node 2 is referred to as a junction node.
 Approach for feeders with laterals

 The forward and backward sweep method can be

used to determine voltages at nodes 3 and 4 given
specification of node 1 voltage together with the
complex values of the loads at nodes 3 and 4.
 Forward and Backward Sweep Method

 So the method is the same as previously given with the

following additional rules:
1. Begin the algorithm by assuming nominal voltages at all far-
end load nodes.
2. Compute all downstream currents (those currents flowing
from a junction node towards the load) before moving
upstream from the junction node.
3. Assign to the junction node a voltage computed based on the
forward sweep done on the last downstream branch
emanating from that junction node. (Note here that this is
arbitrary – we could take the voltage computed based on the
forward sweep from any of the downstream branches
emanating from the junction node).
4. Use forward sweep currents in all backward sweep
calculations.
 Forward and Backward Sweep Method
 Sequence of calculations:
1. Assume nominal voltages at nodes 3 & 4.
2. Compute the load current I23 into node 3. Forward
sweep to junction node 2.
3. Compute the load current I24 into node 4. Forward
sweep to junction node 2. Assign the resulting voltage
to node 2.
4. Compute load current I2 at node 2.
5. Compute I12=I23+I24+I2.
6. Forward sweep to node 1.
7. Using given voltage compute all downstream currents
(those currents flowing from a junction node towards
the load) before moving upstream from the junction
node.
 More Complicated Configuration
1

8 4 10 11

12
9 7 5

13

Feeder with multiple laterals

20-Mar-21 116
 References

[1] R. E. Brown, "Electric Power Distribution Reliability," Marcel Dekker, New

York, 2002.

[2] Beauty, H. Wayne, Electric power distribution systems: a non-technical guide,

Tulsa, Okla.: PennWell, 1998.

[3] T. Gőnen, Electric Power Distribution System Engineering, McGraw-Hill, New

York, 1986.

[4] Westinghouse Electric Corporation: Electric Utility Engineering Reference

Book- Distribution Systems, vol. 3, East Pittsburgh, Pa., 1965.

[5] C. Morrison, "A linear approach to the problem of planning new feeder points
into a distribution system," AIEE Trans., pt. III (PAS), Dec. 1963, pp. 819-832.