MESLEKI

Yfl&IJNCI
Or, . Er,svn

/)I1..J:
'tf1L/N

{(L(atfYttl

#,1;.. ,/fol) ~

AU

-ro f\-r\ C
AA

cOi'J

l' i2.Q L-

(O+OMDt;k

Kvmro\)

!?:ilock

(Gn(l~)

i",p"~

.{.

~----

_-1- _

..-

Dvlp"!

C;'i$1

TX+j j

COl"'lpo.ro.kir
SVMl"'\.i~

PO,'At-

(1

cy.i \o-

Y\A

e \e WlCt"
N1Gt

Ko.[~\\u~~\(

eJ~wtoi1l)

D\stv,bCl,(\C~
Re9t-I(A~ cl

glYNcV

e+b)
OLrtrUfLI'"-

I"f'u.(-

((r:'(;"!~

~----"\
r~edbC\ck
C-

1.~

O\C,M~_.~ AeC\.$v(~f/\(j\'i-

( be. .. bej le :

---f'ced
bcic../c
1v>(! )

~~e.J~O('No('d

~eeJ bo i 'N~~ ~\V~

P OS i' tt' ve..

c.k.

....i

de.fi

bes

) i

cil'f1C-.

feelba.c.\L

Ger; 0\ 0-1'0
G-e.\\

bG5/010i
\c..

S/1JG./
Ll

0):;13

\0.1\\ L,

~eed bOc.\C.

be.:ikl'\ (:(\

s i'\jQ (

Ll~3\,) 0-1'\ i (', \

',/D\.s}u('bClfl~(!' lf\pub

Bazuc\!

8~r~~\e,

C (f'o (':

lfa tc,

Sensor)

De+ec.:tor:

D \J2:)or8"
( Gen~\ di'cr

j'T

('(J.j\S

\ duCer ' ,

D" "\,,"h.~r~eC; onu~

o \ Mc.
61'(

b i("' e ri e 'J;
bi'c..irril't1e

b,' C, n1/I1I'

ene(J;

dOi1u~ turu)

CljSit

olorol

tooIMlom!)

HOl1ipiJlafor,'
II'f'I'L~

Robot)

rohof
-ZJC\W'\Oii"\

kDI
.::.:.e>ic,b,

te..spOf'\se..:
r~$po\'1se
I,

~ f'cq. i.) efi C~

te s po~ (\ .s. te

; fr c.J:::o r1-5
Ce.v'D

,/~ree
-//~o('.(eJ

seibes+'
"2.0r\o(i

(''-jiM
\\4

eel/abi
c.e.v'o.br

b1

response;

('e~iM
L

seii'be:s+
2.0

rej /fI1
rC:j

/{c
Jc.irY1
!-

A ,::.'1'1 h1 \ ~

C evD

b 111 j (1 lap

VY1

l'eJi

NI

!~e.\Jo..b,

unit

st-cp
('Cltvlp

fv(\c..+-ri:H')',

ei(;M
B/I'I'('Yl
17;

bSo.iVlCk

~O(\k..S'~dc;i{1\J f'or'iic :;;:/J

O 1'1

U();t

"flvtlt.l/on:
IJ

re;~ M(:Jl1

/'r un;)

imf'

Ise

fUM.}io

B ir/dJ

ci "r be

!O/1!o,'d o/) LJ

Lop}Clce

IrClris]lorm:

toplQtez.

dBn;;~'"U mi;

S tL\. bl'li f:j : k%r/ik.

<;0.\

n ,

OSJt::C-il ve::.

oF

l'riJ"TOJ4..llt7/C.

CONTI?OL

.sY..srb~

Fe.dfol'I.'<JDtd
Fee.d b~ck

Col1frol

:ifetl'heS)ef'lle./i'

kCfl1rol

(Ol1fto(

! Geri

b~sktf)!.L,ko(1+ro\

k o i7 /Lo i

( {ci/ci wl
di\sei
kO"LJvV!

tC

"Z

,;:)'1

li i ci (i

.Jr:

d t.\
kui
JO(\\
IOri '

~G('

MCi\\<.i'ti~

\.ttde.

koli/rDJ

. .st's~jur1; oalenyiiiidci

Cot1trol : DC3ru

C\~()

c--.

..:io.jisa \ ~u 'i ~ Oeini ii,...

Kontrol U.jo.rlc.it"1C\\I

Kori+i'o(
KOIJt({)

L eo r 1lI'''!J
COt'T1fl

0,(1 fro

i }

63(~r1me.1,'
&1'0/.s0c:Jo.r),

u+erCori

i-roi;

kon tr~ {

~er<\o+e

COfrfrol:

()~aJcl-o.ll

!<urflO()d..q

.3

COf)t('o\

-Lt

~M

10,3

CO a ira /;

sil il!kj,. vef" /,'

AU/il

/co LLL ro'

Di'.sctek

Cctitfo/;,

veri'li'

kol1ffDI
,,,k
cr-

/.$ISO

(SI"3/e i"f'vf
C

o;;:i'!jle

cvfpv/)
)

!Jiri:;

i; -kk ,,;kis/;
cr

rti vii,' /flpvI

Ifwl/,' ou f~ c,rf

o/c ~/f/ ~ Ii'

Cri

L~ 1/

S\MO

Te.l:: 3\'rl'~
cr k g,'rf
-<

~ok q,ki~ Ii
li-

Miso
bisc.re+e.

{eK

cr

i ki

Jr

.f-ime.

.s~s~elYj:.

Ajnk

-zaMt.llill

sJ.c"-lle('

i ii
,

r i
i

CHAPTER

ONE
INTRODUCTION TO CONTROL SYSTEMS

1.1 INTRODUCTION
The automatic control of machines and processes is fundamental to the successful operation of modern industry. Modern manufacturing, processing and transportation systems are heavily dependent on automatic control systems. The benefits of automatic control include more consistent operation, greater safety for the process or machine and operating personnel and reduced operating costs due to improved utilization and reduction in manpower requirements. The need for automatic control continues to grow in terms of the range of applications and performance requirements. The development of high performance civil and military aircraft, missile technologyand space vehicles has placed great demands on the speed and accuracy of attitude control systems. In manufacturing, the developing use of robots and automated production has further increased theneed for reliable, high performance control systems. In the process industry, stricter requirements for produet quality, energy efficiency and pollution leve]s p]ace tighter limits on process control syste.ms. The reducing cost and increased performance of digitaJ coniputers has had a significant impact on the way control systems are designed and implemented. Powerfu] mini- and mieroeomputers with superb graphics capabilities are readily available for designing and simulating control systems. Computers are also widely used for implementing automatic control in an increasing variety of industrial and domestic applieations. Even masS-produced consumer products can indude powerfuI microprocessor systems to monitor and control the system. For example, the average music centre or video camera contains more computing power and automatic control than a typical engineering ]aboratory of the mid ]960s.

1.2 BASIC CONTROL SYSTEM HUMINOLOGV

A cQi1lrol sysiem consIsts of a comroller and apiani. We use the general term plant to deseribe the machine, veliicle or process which is being controlled. The controller can

0i

2 CONTROLYSTEM DESfGN S ANDSfMULA nON

fNTRoDucnoN TO CONTROL YSTEMS S 3

Seqiieiitial and Quantitative

Control elements elements Measurement Controller Plani

Control Systems

Figure LI

A general

control

system

be a person, in which case we have a manual control system. Alternatively, iR an automatic control system the controller is a device, electronic circuit, computer, or mechanical linkage, etc. Figure 1.1 shows the general arrangement. The interface between the plant and the controller requires, actuators (control elements) to provide the control action. In addition instrumentation, detectors and sensors (measurement elements) are neededto provide information about the plant status to the controller. The information passing between the controller and the plant is in the form of signals. These signals can be very diverse, for example electrical, pneumatic or mechanical, etc. The term 'transmitter' is commonly used to describe the measurement element in a process control system because the transmitter sends an electrical or pneumatic signal representing the measured value to the controller. Controllers are usual1y implemented electronically, either using analogue circuits or a digital computer (microprocessor). Pneumatic and hydraulic controllers are alsa to be found. Actuators are commonly pneumatic, electric or hydraulic depending on the application and power level required. The behaviour and performance of a control system depends on the interaction of all the elements. The individual components cannot generalJy be considered in isolation. The plant itself is probably the most important element in any control system; the best controlJer in the world cannot make an inadequate plant operate welL.

A sequential control system involves logic control functions. The sensors monitoring the plant provide switehed outputs which produce only on/off signals. For example an automatic door may be litted with limit switches to detect the position of the door and an infrared detector with a switched output to sense an approaching person. The control funetion involves the use of logical rules so that the actuatars operate in the coneet sequence and at the conect time. Sequential control systems are comman in factory automation, automatic warehouses and the control of batch operations. The design of sequential control systems involves problems in logic and is not covered in this book. The objectives of a quantitative control system are dilferent. This type of control system is concemed with controlling the actual value of some plant quantity. Measurement elements provide quantitative information to the controiier rather than just on/olf signals. The division between sequential control and quantitative control can be vague and same systems can be considered in both ways. As an example, a modern automatic washing machine clearly involves sequential control to switch the various solenoid valves and pumps on and oIT in the required progression for the selected wash program. Quantitative control is used for the wash drum rotational speed. Here the actual drum speed is measured and controlled by altering the power delivered to the motoL A less clear example is the thennostatic control of wash temperature. The thermostat switches the heater element on and off depending on whether the wash temperature is too law or too high. The control signal is clearly on/olT but since it is the actual value of the temperature which is important the system can alsa be considered as quantitative. The behaviour of a quantitative control system depends fundamentaiiy on the rate and extent to which the plant responds to the control action. Such dynamic behaviour is dIfficult to predict and the design of quantitative control systems to achieve acceptable response is no trivial matter. This book is concemed with the behaviour and design of quantitatIve control systems.

1.3 OBJECTIVES OF AUTOMATIC CONTROL SYSTEMS

Regulatioii

Feedback In everyday life, feedback occurs when we are made aware of the consequences of our actions. Feedback is so natural that we take it for granted. Imagine trying to accomplish the simpJest of tasks without feedback, for example, trying to waJk without visual feedback. Feedback not only gives verilication of our actions: it allows us to cope with changing environment by adjusting our actions in the presence of unforeseen events -and changing conditIons, Feedback has similar advantages when applied to automatic control. Feedback occurs in automatic control systems when the control action depends upon the measured state of the madiine or process being controlled. Feedback gives an automatic control system the abIlity to deal wIth unexpected disturbances and changes in the plant behaviour.
:ci

A control system for maintaining the plant output constant at the desired value in the presence of external disturbances is called a regulator. Disturbances will cause the plant output to deviate and the regu]ator must apply control action or control effort to attempt to maintain the plant output at the reference value with the minimum of error. Feedback is fundamental to regulation becaus~ only feedback can provide information about the actuaJ plant output. A good regulator will minimize,the effects of disturba~s on the plant output. Tr:iijectory F ollowing Quite often a control system is required to make the plant output [o]]ow a certain profile or trajectory. A ser va system is a control system specificaliy designed 10 foiiow a changing

r.~
g:

;8

4 CONTROL SYSTEM DESIGN AND SIMULATION

INTRODUCTiON

TO CONTROL SYSTEMS 5

reference value. The servo problem, as it is called, is of major concem in transportation, defence and manufacturing systems. The servo must apply control effort to mak e the plant output follow the desired path with the minimum of erroL it is clear that the regulation and servo problem s are very similar and indeed many control systems give good regulation against disturbances and dose following of a changing reference,

J
-i

~
1.4 CONTROL STRATEGIES
In order to examine some different control strategies control problem. Figure 1.2 shows a tank holding liquid process being supplied requires a constant head feed and to keep the tank leve! constant at some reference level. inIet to alter or modulate the flow rate. let us consider a simple level for feeding some process. The so a control system is required A valve is located in the tank

Open-Ioop Control
The simplest strategy is to have a dial on the inlet valve, By experiment the valve can be moved to different positions and a note made of the dial position and the corresponding level in the vessel. The dial can be calibrated in 'metres'. Thus if it is decided to operate at a different !evel, the valve can be moved to the corresponding position on the diaL. This strategy is termed open-Ioop control. Open-loop control is simp!e and will work well provided there is no change in the flow of liquid from the vessel and all other parameters affecting the level in the vessel remainconstanL

be determined by examining the calibration curve fot the new Dowand then opening or clasing the inIet valve accordingly, This strategy is termed feedforward control. feedforward control requires a set of calibration curves or a model of the relatianship between the valve position, outDowand level. The relationship can be obtained experimentally by measuring the leve! for various outflow rates and inlet valve positions. Alternative!y the model can be formulated from a theoretical analysis of the tank, Another consequence of this strategy is that a measurement of the outlet Dow rate is required to calculate the necessary change in the position of the inlet valve. Feedforward control is an improvement over simple open-Ioop control. But it only caters for the one variable which is being monitored (in this ease the outflow) and relies on a good model of the planL if the model is inaccurate or thebehaviour of the systern varies with time then the feedforward strategy may not work too weii. Disturbances can originate from many causes which may not be included in the model, or are not monitored. For example, the supply pressure upstream of the inle! valve may change or the density of the liquid could alter. THese variations wiii cause the relationship between inlet valve position and tank inDow to change and so the tank level will be incorrect, The feedforward strategy wiii not correct for these factorso

Feedback Control
Rather than adding more feedforward measurements to compensate for these other factors, the obvious solution tomaintain the level in the vessel is to monitor the level itself and :idjust the inlet valve if the level deviates from .the desired value, Such a feedback strategy is error driven in that the control effort is a function of the difference between the required level and actual level. The relationship between the error and the Feedback control, unlike feedforward, can give control effort is called the controllaw. regulation against unmeasured or unmodelled disturbances, 011/0ff Feedback

Feedforward Control
The major cause of disturbances affecting the tank level is likely to be changes in the tank outDow rate. An increased outDow will cause the level to drop. A mare sophisticated strategy is to use a set of calibrations over a number of outflow rates. By monitoring the outDow rate when the plant is in operation, the correct position of the valve can

Control

---t--i
T
i

Inflow

","" '"""'i

rJ

J_

r_
-10c"".
Figuce 1.2 Example plani for levc1 control

The simplest method of monitoring the level is by means of a level switch (Doat switch). The level switch is mounted in the tank at the desired level. The switch produces a binary (on/off) signal that indicates whether the !evel is above or below the required value. The signal can be used to operate the inlet valve directly, When the level is above the reference the inlet valve is closed and when belo,w it is opened, The control law in on/off control therefore switches the control effort between extremes depending on the sign of the errOL On/off control certainly overcomes the criticisms of the open-Ioop and feedforward strategies, Whatever the cause of the change in level, if the deviation in the level is large enough to activate the switch then control action will be applied to correct the situation, The required level (refereneie) in this simple scheme is determined by the position of the level switch on the tank, On/off control requires only very simple equipment in the form of level switches and a simple solenoid type actuator to open or shut the valve, Tliere are several problems with on/of[ control. One problem concems the violent Ductuations in inlet Dow as the valve switches between fuJly open and fuJly shut. These now changes may appear as a significant disturbance to any process feeding the tank,

'f'i

~
'~

CONTROL SYSTEM DESIGN AND SIMULATION

INTRODUCTION

TO CONTROL SYSTEMS 7

Oisturbances

Another problem concerns plan ts which do not respond immediately to the control effort appIied, Such delays are comman in more complex plants and particularly in temperature control systems, Any delay in the plant response means that the process output will continue to rise even after the upper switching limit is reached, EventualIy the output will respond to the control effort and start to fall, but of course the process output will once again continue ta drop past the lower switching limiL Thus the precision of on/off control depends heavily on' any plant delay, Even with negligible delay, the tank level will be constantly fluctuating, if the switch is very sensitiye it only requires a smail change in level to change state, A sensitiye switch wiIl cause the inlet valve to switch between fully open and fully closed very frequently as the level cycIes around the required setting, Alsa anywaves or ripples on the Iiquid surface could cause a very sensitiye switch to be activated very rapidly, The frequency of switching can be traded off against a lass of precision in the level control by using a less sensitiye switch or even by using two switches, one at high level to dose the inlet valve and anather at low level to open il.

~
Th

Model and calculalion of feedforward conlribution Feedforward contribulion Feedback _

Measurement

:il

.~.

~ t

Total control

~. ~
,~

~ ~

contributiOn\ Feedback controller

i
>

+ ( 0eiio;ower . ,i amplification
and actuator

Plant

Conlrolled variable

i
~
{
}

'"

Feedback loop

Measured value

Measurement

Modulating Feedback Control

Rather than switch the inlet valve open and shut, a more subtIe appro'ach is to inch the valve by an amount which depends on the difference between the actual Ievel and the desired level. This strategy can be termed moduZating feedback control. Modulating control implies a more elaborate level measurement and valve actuator. In the first place it requires a signal related to the actuaIlevel (Le, a level transmitter), Secondly the valve actuatar must be able to open and dose the inIet valve gradually (moduIate the valve opening). Furthermore the valve itself must have a smooth characteristic so that its resistance to Dow is infinitely variable. The control law can profoundly effect the way in which a feedback control system behaves. SimpIe on/off control can sametimes give acceptable performance; most domestic heating systems are controlled this way. The important characteristic of modulating control is that it is capable of providing a range of control effort and can produce smail, as well as large, conections. With a well designed control law, feedback control can provide go ad regulation and trajectory following. One of the primary functions of a control engineer is to design or select an appropriate control law for the plant which gives acceptable performance. Figure 1.3 shows a general black diagram of acontrol system with both feedforward and feedback control. The box labelled 'feedback controller' is where the feedback contribution to the control effort is produced. if there is a feedforward component, is added to the feedback contribution to make up the total control effart signal. it

Figure 1.3 General black diagram of a control system

dangerous levels. Safety systems are usually separate from the normal control system and can be treated independently in terms of strategy and implementation, Safety systems usually involve detection of unsafe or potentially unsafe conditions using on/off or switching sensors. Standby equij:iment may need to be started if there is a failure or the plant may need to be shut down safely. Safety systems involve the same problems of binary logic as sequentiaI control systems. ' Quite often control systems can be designed to fail safe. With the correct design, failures of instrumentation and control equipment can result in safe rather than unsafe situations. For example consider the level control problem examined above, It may be that the liquid held in the tank is corrosive or otherwise dangerous. lt would therefore be unsafe if the tank were to overflow due to a control system failure. The fail safe philosophy is to design the control system and associated instrumentation such that a failure in any one element causes the control errort to act in a safe way (i.e. the tank inlet valve should dos e )~ For a fail safe measurement, a failure in the measuring element should produce the same signal as exists in the dangerous condition, In the case of the level control system a level transmitter failure should look like a high level in the tank. This can be achieved by using a law signal to represent a high level (a reverse acting transmitter). Loss of the measurement signal would thus look ]ike a high level in the tank and so the controller would shut the valve preventing tank overflow. The fail safe philosophy can alsa be applied to the control system actumor. The argument here is simpler, the actuator should move to the safe pasition in the event of an actuatar power failure. For thetank inlet control valve this means that the vaIve should dose if the actuator signal is lost. This can be, accomplished by making the actuatar push against a spring to open the valve (i.e. a spring return actuator), The spring would then close the valve in the absence of actuator power,

--

1.5 SAFETY

--

One overriding consideratian in control system design and imp]ementatIon is safety. This may relate to the reliability and robustness of the control strategy, but more often deals with the monitoring of exceptiona! conditions and the subsequent alarm and safety systems associated with the detection of a malfunction or of a variable that has reached

\x)

8 CONTROL SYSTEM DESIGN

AND

SIMULATION Disturbances Temperature controller

INTRODUCTION Disturbances Fuel Ilow

TO CONTROL SYSTEMS

The above arguments are based on the premise that the tank should not be allawed to overllow. if the Iiquid held in the tank is essential to the downstream process (i.e. lubricating oil or coolant) then the tank should not be allawed to mn dry in the event of a faiIure. The situation is now reversed and the fail safe argument implies that a direct acting level transmitter should be used. Similarly the control valve must be designed to open, should the actuator power fai!. Safety considerations are very dependent on the process or system, and notwithstanding their importance, can only be touched upon ina general text of this nature. Experience and a thorough knowledge of the system operation are required before sensible decisions on safety aspects can be made. it must be emphasized that most control engineers spend far rnore time considering safety aspects than in designing the control system for normaloperation.

i---------~-------i

Process outlet temperature Process lurnace \

\ .i
!

i
output

Control

Relerence temperature

i+Di1 i' __ \_:~~~~ ] \ 3!im

it

Errar

Measured temperature '\ Controller Temperature transmitter

1.6 EXAMPLES OF CONTROL SYSTEMS

Furuace Control
Furnaces are use d in the process industry for heating feedstocks prior to further treatment. For example, in an oII refinery, cmd e oil is heated before it enters the cmde distillation column where it is split up into fractions which eventually produce marketabIe products such as aviatIon [uel or road bItumen. The initial heating is carefully controlled as the subsequent fractionation is quite dependent on the degree of vaporization of the crude oil [eed. Figure lA shows a simplified diagram of a process furnace. The temperature of the process fluid is measured at the outlet of the fumace by a temperature transmitter which sen ds a signal to the temperature controller. The controller has a dial so that the desired

Figure 1.5

Biock diagram

for temperature

control of a process

(urnace

Heated process-Iluid to treatment plant Temperature transmitter

i
Pump i.---Measured

value of the process outlet ternperature (the reference ternperature) can be set. The controller compares the measured temperature with the reference temperature and the fuel flow is increased or decreased accordingly. The exact relatianship between the error in temperature and the fuel valve mavement is determined by the control law. A black diagram of the control system is shown in Fig. 1.5. The main source of disturbances are changes in the process fluid flowrate. Changes in throughput are inevitable for nuqierous reasons. For example, if there is unsufficient feedstock or if there is a sudden increase in demand because of a weather change or even fluctuations in the stock market. Feedforward control would be relatively easy to implement by monitoring the flow of process fluid and adding in a component to the feedback control ef[ort. Turning to the safety of this [urnace control system, [ail safe precautions can be taken to ensure that a control equipment failure results in the burner shutting down. However there are many other safety considerations to be considered in fired equipment such as this. For example, during normaloperation disturbances in the process fluid could cause the furnace outlet temperature to rise. If the correcting control effort is excessive, the fuel valve may be shut of[ causing a bumer flarne failure. When the fuel flow is restored by the controller opening the valve, the potential for an explosion will exist because o[ the unignited fuel entering the not fumace. Solutions to this problem can Include placing limits on the [uel control valve travel, use of name detectors and automatic re-ignition of the extinguished flame.

Fueloil

ControJ,.. valve

__ _+-----

Temperature contrailer

V __

temperature

Control of a Robotic Arm

Robats are increasingly used for materials handling, automatic assembly and fabrication. Robotic arms normally have several joints or axes each fitted with an actuator enabljng the arm to move in a variety of ways to pasition and orientate the gripper. The actuators are controlled by a dedicated computer so that the conect sequence o[ motions is carried out. The computer normally does this by replaying a stored sequence of desired motions Sametimes more sophisticated systems use televisian cameras or toueh sensors to heIp decide on the required gripper motion,

Desired/reference temperature

To other consumers

Figure

1.4

Process

furnace

ls>

W -8

INTRODUCTION
LO

TO CONTROL

SYSTEMS

]]

CONTROL

SYSTEM

DESIGN

AND

SIMULATION

Reference pasitlon

Gripper
r r r
r

Digrtallyencoded

Servomotor and gearbox

i
:~
-j:!

i
r

arm position

Control
______________________

computer

i
1

j~.

Figure 1.7 Black diagrain of robotic arin control system (one axis only)

.~

i
i

Figure 1.6 Robotic ann

ii ]

.~

In practice however, the control with a range of loads.

law is usually chosen to give adequate

performance

..:::....

The control computer must control all the robotjoints simuItaneously. Sametimes each axis has its own separate microprocessor which receives commancls from the main control computer. However implemented, the scheme for the PQsitioning of each axis is similar, so the controlaf just one arm joint will be examined as shown in Fig. 1.6. In this example, the joint is driven by a de electric motor (servomotor) through a gear box. A de motor is a flexible actuatorwhich can be driven in either directian at various speeds by altering the directian and magnitude of the motorcurrent. The current is supplied to the motor by a power amplifier which in turn receives its input from the computer or microprocessor. The computer deals with digital data and so the interface to the power amplifier requires a digital to analogue converter. The angle of the arm joint is monitored by an encoder. An en cad er is a transducer which ppoduces a digital output representing the measiired angle which can be directly interfaced tb the computer. As the computer retrieves the sequence of desired robot motions, the required positions for the axis are passed to a part of the program which implements the control law. The control cakiilation involves subtracting the measured arm angle from the required arm angle to find the angular pasition error. The computer then uses this error to detennine the magnitude and directian of the control effort. This calculated control errort is converted to an analogue signal which Is applied to the inpcit of the power amplifier to diive the servomotor and hence rediice the error. Figure 1.7 shows the arrangement in black diagram form. The main requirement in this control system is that of fast and acciirate trajectory following. The controller can use the rate of change of measured pasition to calculate the yelocity of the joint and so reduce the motor current if the arm is maving tf1ti fast. The control law in a high-performance robot may be q uite compiex and will be tailored to the way in which the arm is expected to respond. A difficulty arises because the arm may respond differently when carrying a load than when empty-handed. The motor will require more current to accelerate and decelerate with the load than wlthout. The control computer may know whether the gripper is holding a load and in theory could then compensate by using more current.

i ,
j

~
1.7 CONTROL SYSTEM i;>YNAMICS, MODELLING, SIMULATION ANDDESIGN
The response of any natural system to stimuli is not instantaneous, it takes time for the system to respond. We are all aware that before we can make a cup of instant coffee, it tak es time for the kettle to boil after we have switched it on. When acar hits a bump in the road the suspensian bounces or oscillates for a while before it settles back to an equilibrium position. When we hit the accelerator pedal of our car, it takes time for it to accelerate to a new speed, evep. if we happen to own the latest Lamborghini! The way the response of a system evolves as function of time to a particular stimulus is termed the dynamic response. The dynamic behaviour ol systems can be modelled by differential equations. Models can be obtained by applying the basic equations of physics and mechanics to the system. This approach can beapplied to systems where the underlying principles are dear and where the system is sufficiently simple, or can be broken dow n into simple subsystems, to use a 'first principles' approach. Anather approach is to observe the behaviour of the system in its normal working enyironment or introduce test signals. A model can then be proposed from the observations of the input/oiitput behaviour. The solution of differentiaJ equations by hand methods is difficult and so a range of numerical methods have been developed which can be implemented on digita! computers. Many computer techniques and packages exist for solvlng differentia] equations and simulating the behaviour of dynamic systems. Control system design involves simulating the dynamic behaviour-of feedback systems, but it alsa involves other techniques which require the use of computers. Dynamic simulation a!one is not enough. Design method s for control systems have evolyed since the 1930s. At first methods were devloloped which were based on short-cut hand calculations. Analogue computers were used to do the final simulation and verify the design. This state of affairs did not change significantly unli! the early 1950s when the first digital computers came int o commercia] and scientific use. Computer programs

CJ

12

CONTROL

SYSTEM

DESIGN

AND

SIMULATION

INTRODUCTION

TO CONTROL

SYSTEMS

13

were wrjtten to solve sets of differential equations and results produced on listing paper. Graphic displays were yirtuaiiy nonexistent and the interaction between the designer and the computer was yery poor. Even with the development .of minicomputers in the mid 1960s the situation was not significantly better. Graphics display terminals first became widely available in theearly 1970s. Their use for control system design was generally restricted to larger universities and industria! companies. User interaction was still poor, and often the design suites were not well integrated. The 1980s saw the arrival of the 'personal computer' (PC) and the availability of graphics terminals of high quality. Many software houses adapted existingprograms to the PC, but really did not tak e advantage of its potential as a design tool. The ancestry of the programs was dearly vIsible. In addition to the improyements and cost reduction of the hardware, the second half of the 1980s saw an increasing awareness of well design ed (user friendly) software. Business sodftware with programs such as word processors, spreadsheets and databases was the first area in which this development occmred. Later, as graphics capabilities grew, computer-aided design paekages for draughting, three-dimensional modelling, printed eireuit design, ete., appeared. Computer-Aided Control System Design (CACSD) software was slow to respand to this trend sinee a fairly large investment was required with a more restricted market. Taday there are several good quality, low cost packages available.

//
Oil cooler HotoH

Cooling air

Cooled oH

~ i
Temperature controller Figure P1.3

V02\---------control system

TT ) Temperature transmitter

Oil cooler temperature

PROBLEMS
1.1 Classify presenI: (a) (b) (c) (d) the following into sequential or quanlitative control systems and identify whether feedback is

The thermosta!;c temperalure control in a central heating boiler. The programmer controlling a typical central heating system. A cruise control system, as Iltted to a modern car. The over-speed cutout Iltted to acar engine.

1.2 Examine lhe fail safe requirements for the inslrumentaiion and actuators in the furnace temperature control syslem shown in Fig. 1.4. You may assume that it is unsafe for the Process fluid to overheat. 1.3 Consider the oil cooling system in Fig. P1.3. The oil temperature is controlled by throttling the cooler bypass valve. Safety considerations indicate thai control equipment failure should not cause lass of the cooling function. Determine the required aclion for the temperature transducer and the control valve. What action will the canIraller take if the signal from the temperature transducer reduces in value?

'"'
""""

o/-kr
.s w;t1j/~

on
f)

,'/7terva i

o/- hrne

a pea dl)i..JM

orid
\J;

f~e

lffofr'o/1

1rci. 1-,'0 (), r~reCij:

//7

Vi 6('0,,+1'0(\:
Oset' Ila tr'o 1'\
I,

-ri'+re~ im
~. L ri\M

Ic.JuarlaMak

fa

va i : o. rci. li k.

{~ s,,"i"j"
pluek "

",Me
q..eb'p

b.",bXiorJs
b\('ub()().

","d

D( \U or d..,

.f1,."",

tJ

Iixl!.d

eiid

Des""

v~

Freedom
"cl M ber

(Ser'ties+\;l
o-f

Oerccesi)
eoo re! ,'o Q -I-es aif paris
oP r< 'tli
Ci

-i hec
a-C onj

m" o" ,"cWI

; n J <ep ",,,de n -i-

ired
-Ik

io dda",ine
sjS+eM

OlMpldecb
o -P- lime

~he posifl~f\S

oF

o~

sj,dvYl
of

I"S lar'/-

de

.,'/leS

-(he ci'tJ re e

fre do ro

e : lft1c/e
Xi ~',

ca.r+cs laf)

CDoudi/1dfc-s

Tk

word',,,a-le,,
Motio

ne ce.$S"C:j

+o

decri h~

-tk
Q

oP

s<l~+C"'; CO~$+; /-u i~

se.+

oP (jene ro 1,'2ec1

"DordiMle.!
~erle
Kt:>t)rt:.f

(Geuca/",e d eODrdiflofd,'

NH tr~il"'if)

,'1'9c1

fltw

Ol'S ere.. +C

o.nd

S~sk",-<
ect.\\ej -

"-la-h

C\' ~'tl'+c:
oC"

tl "",i'ef
pon.HY1!:.J.(:('

o~

dC3r.,-es

elf freedo'"
cuid. -\-'no.se

GfC:

diserek

(umped

.s.,:jstcrT!-S)

~~-\'"

"'-" in ~i,,"-k "u"" bel'

e~d <_~te"" 9~

4rcede""

o.re

e o.lkd

con ti()u(}u-s aC' d,',d-('; b0ie.d

sjskm-s.

~\QS+jC

\;()~

det i i:t.:Ji' o ('\ c u('v~

Raf:mOIli'c

Moh'00
rII"-I-,,ii,

( Hor,-no n;k
""'~

HCH'e:ke1)

OS<-;\\'~H::'i'
D~

repc:a1 O~

Ns.J!

res,,/or!:J,

cH

in -Ihe cMc:

II S'Mp\,:
\ <i

pel'\dulut'l'\

"'Oj c!isploj c,,"siderabl'"

MDt.D" <luri,
e1j'tJa/

Irre3"lad(jJ
ke -

) os
1.(2

-\:hoc c<>-<e IYIDllon

"f
Th,z

5rD""J

0/1 e arfh

IJ-ua

fi~

is repea/d

alkr

!/lfu-o/als / er I/me. . J of! per/[)clh: ... MD ,/0/)

if

i.s

wlk..}

peri'c:icL'c..

Mo/rOf).

w:
aca /",rci{/oti ls CfHfG
i't
Celf)

(uo,ooriioral

be

seel1

+ha f

ro

fhe

d/.s~}acemel)

such
J.

cl vi hl'a

kol)

/s

kl1Dl:t.JrJ

as .s,'fri e. i 13J pl

ho.fmOJ1,'C-

mo/ion.

The
1Y\o+to (\,

~\loWi'll~

c ~C \e.
(Cevn'~)

Li

The.
posi'ti'c>n

mol/eme.nt

o~ c\ v;brct~i\nj
dlre.cflon,
+hu\
1-0

bod.,j
-the

frorll

,'+is meaA +c>

i~ Me

M<"i'\) 4h<n
(" '"

ih

ex+. M~

f' o" ih' " f1 i" oth<of' d/ce di",,)

a cjcJe. p vi hra f,'ot)

"'.ol

b ct<ok

ea" b cA i1 A.

p~r\'d

e~

O.sC.il\Cltllot'\:

-,

fi~ .J.,',"~ -laken

.fh feci nJ

-lo c.oMp)~f~ "r

aiie
{-;Mi:'

(t)cI~
fXriDd

ci. ,,'oti 01'>

Clr'lcl..

ts k~o"'l'1 ,,~
bjZ.

o P o sc i'iLa 1-/0"

,'s

deno1ed

freqveflt}j
I'.

0-1'

Ou\ /Lo O i'> "lAe t.

MM

ba

of
or

ej da
SIMf'~

p<r U/l!;'
f~

I/ed

.fh"

jre ~U,."c.3 oj DJ""'( <i hon

b,:j
-

ifrque""2:l

k,,~

0.(\(:1.

ls de no teJ
T -= ()

f
::.. ...--LU

iz
4

2.1f

i.
W

ci c on s J"tI t,

C~

1\ clli o
-i<>

b e . J Ne
cl ,'s lin 1\ v

/0

d "'0

le

fre q u "/Ic::J
/hr.. ;,~

is

C ct lIecl

-fhe

dN u10(' ~~

if iJ<nc:~

i's h

i~p" M

al'

/re

qUf\

C~

f = w I~Ti .
ls
meaSVee,}

w de/l,Aes
'"

the Cl"!ylar se

V'lzJoc,'/:j

of the

<:::JcI,<

(t1of/on

OliJ

rcidi().~S per

C1:JII , d

(ik
Pha5<:

<:\<;'$1)

GJ CMs,Je('
+""0
V;

Anf)lc.:

bro-\-orj

",ot,'O(lS .lccMt"'!

6~

)(2::

A2 .sin (wf

+~)
~

ci..,i) Ph~S~ A"jle.

A/"t1p/;fude"

.p

h 5e di' ~ ger(.()ce C h:ii. forki

<).

Na4v""\

P""q\le'\C,j' cm;b

i~ o
O"-l")

s::iste'"

is

drd"Cb"d
w;-\k

0.01 vJhich

"llo",ed
;~ oscill"les

-M ";b<~~e
O

"I,,, {reqve"c.3 or d.,,,,P,-" ~

s~st"""

""i'rh v1- ede r",.I

.(lre 'l')eI\C ~.
i)().;
LI

~o rt:eS

Is""

DWn

as ;ls ila tvr"

{lreedo n'1

A ,,;b<oIM.j

h"'''''n5 n d<,5rees () f
"a tu tQ \

h",.e. i ilI 5eM ro' \ ~ J ,'J ~i ne.. t

~c<q.le

ri

et' "oS

"f

vi bra f,'ot

15

CLASSIFICA TION OF VIBRA TiON

Vibration can he classified classifications are as follows.

in several

ways.

Some

of the

imfortant

Free and Forced Vibration rS:t'bt.~ JFree VIbration: if a system, after an initial disturbanee, is left to vibrate on its ,,-tI~~:Sii"\ own, the ensuing vibration is known as free vibration. No repeated extemal force acts on the system. The oscillation of a simple pendulum is an example of free vibration. (ensue: happen later or as a result; follow)

Forced VIliration, if a system is subjected to an extemal force (often, a

(c.of'lofll"l!\'.s'

i\+rt'~im

repeating type of force), the resulting vibration is known as forced vibration. The oscillation that arises in machines such as diesel engines is an example of forced vibration. if the frequency of the extemal force coincides with one of the natural

frequencies of the system, a condition known as resonance occurs, and the system undergoes dangerously largeoscillations. bui1dings, bridges, and airplanewingshave of resonance. Failures of sueh struetures as ~

been associated withthe occurrence

Undamped and Damped VIbratIon if no energy is lost or dissipated in friction or other resistance during oscillation, the vibration is known as undamped vibration. If any energy is lost in this way, on the other hand, it is called damped vibration. In many physical systems, the amount of damping is so small that it can he disregarded for most engineering purposes. However, consideration of damping becomes extremely important in analyzing vibratory systems near resonance.

tb
1
",!,',. :;':~

''''',',

.....

,
"'_'C

{;;;i;~'::: Y~-

'.

~':~I::;,.;:~~f':r:~~~~:~~j~ii~~~i~;'i:~~~~4~*ifJM; :;;,1i;~<~~(X~ti;X;,f;)(i);:,:;,;:;:/~}~;,;::<"
,.@:;f,~i~iKiJi:iti~;',:;

Force

Fiirc~

il

(~lA ifi:termini~tii:(ptri()uic)

exciiaiion

(h) ,\ ramlom i:~CifJtitJn

Deterministic and Randm VibratIn

(To(iirf/Ic

Ile

i<asigele T;'tresif'l1)

Hthe value of the excitation (force or motion) acting on a vibratory system is known at any given time, the excitation is ealled deterministic. The resulting vibration is known as deterministic vibration.

In some cases, the exeitation is, nondeterministie or random; the value of the exeitation at a given time cannot be predieted. In these cases, a large coqection of records of the excitation may exhibit some statistical regularity. it is possible to estimate averages such as the mean and mean square values oLthe excitation.

Examples of ranclam excitations are wind velocity, road roughness, and ground motion during earthquakes. If the excitation is random, the resulting vibration is called random vibration. In the case of raridom vibration, the vibratory response of the system is also random; it can be described only in terms of statistical quantities Figure 1.10 shows examples of deterministic and randam exc-itations.
1.6 Vibr.ilon Analy.l. ~OOJu",

13

LO)

,.Fb,iidi}
ol ... ecl, h lhmflin~(I1
Suo.j'l('n\i(lfI

Mau _nd

m03~~ mOnlenf ni h<wfi

of ~ M" ... r:J n'('r:; nr

rLu. e-;ih. ~
I:b\Jic-ir.\
;iond

incrri:i

d~nirinr

tir "C'lU\

'-------\ \
------_/

EJa!\fidlynf "'"hn'/ .

CcJ

SPRING ELEMENTS

(Yq~

EL

iz:

iMii \ ..ri )

A linear spring is a type of mechanical link which is generally assumed to have negligible mass and damping. A force is developed in the spring whenever there is relative motion between the two ends of the spring. The spring force is proportional to the amount of deformation and is given by
F=kx (1. 1)

where F is the spring force. x is the deformation (displacement of one end with respect to the other), and k is the coefticient ofproportionality, stiffness or spring constant. If we pIot a graph between F and x. the result is a stniight line according to Eq. called the spring

(i. i). The work done in aeforming a spring isstored as strain or potential energy
in the spring. Actual springs are i:ionlinear and fgllow Eq. (Ll) only within a certain range of deformation.

MASS OR INERTIA ELEMENTS The mass or inertia element is assumed to be a rigid body; it can gain or los e kinetic energy whenever the velocity of the body changes. From Newton's second law ofmotion, the product of the mass and its acceleration is eqmti to the force applied to the mass. Work is equal to the force multiplied by the displacement in the directian of the force and the work done on a mass is stored inthe form ofkinetic energy of the mass.

In most cases, we mu st use a mathematical model to represent the actual vibrating system, and there are often several possible models. As stated earlier, the purpose of the analysis determines which mathematical model is appropriate. Once the model is chosen, the mass or inertia elements of the system can be easily identified. For example, consider the cantilever beam with a tip mass shown in Fig. 1.14(a). For a quick and reasonably accurate analysis, the mass and damping of the beam can be disregarded; the system can be modeled as a spring-mass system, as shown in Fig. 1. 4(b). The tip mass m represents the mass element, and the elasticity of the beam denotes the stiffness of the spring. Next, consider a multistory building subjected to an earthquake. Assuming that the mass of the frame is negligible compared to the masses of the flooirs, the building can be modeled as a multidegree of freedam system, as shown in Fig. i. 18. The masses at the various floor levels represent the mass elements, and the elasticities of the vertical members denote the spring elements.
4

~"i

?/ "
ni, mi m~
/~

_i,
_i:

_i,
(1))

F~u,. 1.18 Idealilalion ot a mullislOlr niun1aegoool ti~ii syslem.

bui/eliiig as ~

DAMPING ELEMENTS A damper is assumed to have neither mass nor elasticity, and damping force exist only if there is relative velocity between the two ends of the damper. The energy or work iI1,putto a damper is eonverted into heat or sound: hence the damping element is noneonservative. The d~ping following types may be one or more of the

Viscous Damping. Viseousdamping is the most eommonly used damping (\Ji'-z,kO"2- S~f\tlM-) . meehanism in vibration analysis. This type of damping is present whenever a

viscous fluid flows through asIot, around a piston in a eylinder, or around the journal in a bearing. In viseous damping, the damping force is proportional to the veloeity of the vibrating body.

Coulomb or Dry Friction Damping. Here the damping force is eonstant in {J< vrv ,J cl r- Hui~e: S'04 M'V } magnitude but opposite in direction to that of the motion of the vibrating body. it is eaused by kinetic frietion between two sliding dry surfaees. Material or Solid or Hysteretic
le, .s6110MW)

Damping.

When materials are deformed,

U.{a{-;tefYIg..

energy is absorbed and dissipated by the materia1. This effect is due to frietion between the internal planes. whieh slip or slide as the deformatians take place.
5

20

1.0 FUNDAMENTALS OF VIBRATION 1.1 What is Vibration? Mechanlcal vlbratlon is a form of motlon oscillation. It occurs In all forms of machlnery and equlpment. it Is what you feel when you put your hand on the hood of acar, the engine of whlch Is runnlng, or on the base of an electrlc motor when the motor is runnlng. Perhaps the simplest illustratlon of a mechanlcal vibration is a vertical sprlng wlth welght, as shown in FIgure 1. In this posltion, the defiectlon of the spring from its free state Is just sufficlent to counterbalance the welght W. This defiectlon ls called the static deflection of the sprlng. The pasition in which the spring is at rest is #1. The sp ring Is then slowly extended to posltion #2, and released. The subsequent motlon of the welght as a function of time, when there is negllgible reslstance to the motlon, is wavy and repetltlve as shown in the graph. it exhlblts many of the basic characteristlcs of mechanlcal vlbratlons. The maximum dlsplacement from the rest or me an posltlon Is calied the AMPLITUDE of the vlbratlon. The vibratory motion repeats itself at regular Intervals (Al, A2, A3). The Interval of time wlthin whlch the motion sequence repeats Itself is called a CYCLE or PERIOD. The number of cycles executed in a unit time (for example, during one second or during one mlnute), is known as the the- spring-weight system is FREQUENCY.In a high-speed osclllatlon the frequency is high and conversely. When, as In Figure not drlven by an outside source,'the vlbration Is a FREEVIBRATION and the frequency Is called the NATURAL FREQUENCYof the system. In general, vlbratory motion mav or mav not be repetltive and its shape as a funetlon of time mav be simple or complex. Typical vlbratlons, whlch are repetltive and continuous, are those of the base or houslng of an electric motor, household fans, vacuum deaners and sewing rrrachlnes, for example. Vlbrations of short duration and variable intensity are frequently Initiated by a sudden impact or shock load; for example, rocket equlpment upon takeoff, equipment subjeet to Impaet and drop tests, a package falling from a height, or a Iading in a freight car. In many machines, the vibration Is not part of its regular or Intended operation and funetion, but rather it cannot be avoided. The task of vlbratlon isolation is to controi this unwanted vibration so that Its adverse effects are kept withln acceptable limlts.

i,

:Fi~iite1
T195

1.2 What Causes Vibration? It is an UNBALANCED fORCE, or system of The basic cause Is aiready evldent In the slmple mass-spring system of flgure forces (In Figure lt Is the spring force aeting on the weight) aeting on or through an ELASTIC OR RESILlENT MATERIAL (In Figure 1, this is the spring). The unbalanced force mav be due to mass unbalance, such as In an eccentrlcaily mounted rotor, or lt mav be due to the variable inertla forces In machinery, which does not move unlformly, e.g. crank-and-conneetlng-rod motion, I.inkages, cam-follower systems. In the latter, the speeds and dlreetlons of motian of machine parts are contlnuously changing, e.g. the needie motion in a household sewing machine, bucket motions In earth-movlng machlnery, ete. Force unbalance can arlse also from electrle, hydraulic and acoustlc sources, e.g. transformer hum, water hammer, a loudspeaker, ete.

i.

1.3 Adverse Effects of UncontrolledVibrations The objectlonable results of machlne vlbrations, If left uncontrolied, can be several: High stresses and foree levels mav be set up as a result of vibratlons and in extreme cases mav lead to part fallure. Such fallure can be sudden or gradualras~infatlgue.More frequently,therels Increased w.ear of parts anduns~tisfaetory equlpment performance. This requlres increased malntenance and mayaiso Involve downtlme of equipmenL For example, In a machine tool with excessive vibrations,_parts mav be inaccurately machined and subsequently rejected. In other casesi an inadequately cushioned machlne mav walk away on Its foundation. And finaliy, noise may become excesslve, Independent of stress levels, consumer produet acceptance maybe jeopardized, and w.orklng conditions mav become unacceptable, Usually, the objectionable
results are a combination of these circumstances.

1.4 Principles of Vibration and Shock Isolation In dlscussing vlbratlon Isolation, it is useful to identlfy the three basic elements of all vlbratlng systems: the equipment (component, machine, motor, instrument or part); the vlbratlon mount or Isolator (resillent member); and the base (fioor, base plate, concrete foundation, ete.); the vibratlon mount Is a resillent member (rubber pa d, sprlng or the like), whlch is interposed between the equlpment and the base. It is usualiy qulte smaiL. lf the equipment Is the souree of the vlbratlon, the purpose of the vlbratlon mount is to reduee the force transmitted from equlpment to base. The directlon of force transmission is from equipment to base. This is probably the most common case. lf the base is the souree of the vibration, the purpose of the vibratlon mount Is to reduce the vlbratlng motlon transmitted from the base to the equipmenL The direction of motion transmission Is from base to equipmenL This case arises, for Instance, In proteeting delicate measuring Instruments from vlbrating fioors, ete. In either case, the prlnclpie of the cushioning aetlon of the vibration Isolator Is the same. The Isolator is a reslllent member. It aets both as a time delay and a source of temporary energy storage, whlch evens out the force or motlon dlsturbance on one slde of the vlbratlon mount and transmits or meters out a lesser, controlled dlsturbance, at the other end of the mount. A good vibration mount, thereforei slows equipment response to a force- or motian disturbance. In engineering termsi the charaeterlstic of a good vibratlon mount is that the natural frequency of the equlpment wlth the mount Is substantially lower than the frequency of the vibratlon source (forcing frequency). The design of a suitable vlbratlon mount insures that this Is the case.
Conversely a poorly designed mount, having an undesirable frequency characteristic, can be WOrse than no mount at ali.

In addition to its funetion as a time delay and source of temporary energy storage, vibration mounts can also function as energy dlssipators or absorbers. This effeet Is usualiy produced by the damping charaeteristics of materials, viscous fiulds, slidlng frietion, and dashpots, T196

21

a7f!iough in general these mav or mav not be part of the isolatar. The damping, or energy-dissipating effect of an isolatar mav be
negligible or substantlal depending on the applkation. The main purpese of isolatar dampIng \s to re.duce or to attenuate tne. vlbratlons as rapicny as posslble. Damping \s partic"ularly important at certaln speeds, which cause a candition known as RESONANCE. Th'is occurs when the natural frequency of the equipment with isolatar (oincides with the frequency of the source of the vibration. For examp!e, jf an electric motor runs at 3600 RPM, then an isoJator-equipment natural frequency of 3600 cycles per minute corresponds to a conditlan of resanance, if a machine operates near resonancei or has to pass through aresonant speed in order to attain operat1ng speed, damping is important in preventing the buildup of vibration to an unsatisfactory level. In summary, then, a good vibration mount functions as_a time delay, temporary energy absorber and possibly to same extent as an energy dlsslpator, or damper. The engineering design of a vibration mount consists in identifying the characteristics of the souree of the vibration, the mechan"ical characteristics of the equipment and the determtnation of the rnount characteristics, in order to achieve a speclfied degree of vibration reduction.

1.5 Principles of Nois_e Reduction

A good vlbration mount can be effective in reducJng noise as we1Jas in reducing the transmission of forces and motions.

(a) What is Noise?

Sound is a vlbratlon of air. The air in this case is an elastic member. The vlbrations of the air have a frequencyand an intensity (loudness). The frequency can be expressed in eyeles per se~ond or eycfes per mlnute. The audible frequencles range from about 100 cycfes/see. to about 18,000 cyclesisec., although sens'itive human ears mav have a somewhat larger range. Intenslty or loudness, is measured in decibels. Asound intensity of 15 decibels would usually be regarded as quie~, whlle a decibel level of 60 and up iS usual1y regarded as loud and objectionable. Noise mav be regarded as objectionable sound. More speelfically, the decibells essentlally a comparison of the pressure of the sound to that of a standard or reference sound (.0002 micrabars, usuaiiy). In arder to arrive at a reasonable seale of values, the logarithm of thIS ratlo to the base 10 is used

and multipled by twenty. Typical valoes of levels of sound intenslty and noise intensity are shown in the following Tabies la and 1b:
TABLE la VALUES OF SOUND AND NOISE INTENSITY

Various industria! operations and related no ise levels recorded distances of from one to three feet from machine. * *

at

TABLE lb VALUES OF SOUNO AND NOISE INTENSITY

**From: Acoustical Enclosures Muffle Plant Nolse" by S.Wasserman and A.Oppernheim, P!ant Engineering, January

1965.

Mark's Mechanlcal Engineers' HandBoek, Sixth Editfon, Hill Book Co. ine., New York, 1958, Sedian 12,p.153 N to Specify Audible Naise" by E.A. Harris and W.E.

Machine Design Nov. 9, 1961, p.168. T197

(b) What Causes Ndise? A comman cause is the impad or vibration of a solid material, which sets alr In motion; for example, a hammer striking a nail, or a vibrating equipment pane!. In machlnery, in particular, there are many commoniy found sources of no ise. These are usually associated with the operating frequency of machine motions, e.g. the RPM of an eiectric motor or of gears, the rate of tooth engagement in gear teeth, the frequencies associated with reciprocatlng machinery, etc. lt is possible alsa that vibrations mav be generated in one part of the equipment, but mav set up noise and vlbratlon In anather part of the equipment, such as doors, panels, chassis, fiexible lines, printed-circuit boards ete. (c) The Adverse Effects of UncontrolJed Noise There are severaL. First,nolse may be an indication of faulty equipmentoperation, e.g. cracked parts, faulty bearings, excessiye rotor unbalance, Improper lubrication, loose parts, etc. It is possible also, how~ever, for a machine to fundian satisfadorily mechanlcally, but to be rejected by the Customer, If it is too nolsy, e.g. nOisy household appllances, alr condltloners, etc. Second, human efficiency and fatigue mav be adversely affected, for example in production Iines in a nolsy fadory, or in the office. Short of expenSive--ancJ"diffidJit iiiv~estigatibhs, agooirVibration mountcan often be-an effective way to reducenoise !evels to within acceptable limits. (d) How Can Noise be Reduced? There are many ways. One of the most practical and effectlve mav be the use of vibration mounts. As a general rule, a well design ed vibration Isolator will also help reduce ndise. In the case of panel fiutter, for example, a well designed vibratlon mount could reduce or eliminate the noise. This can be achieved by elimlnating the fiutter of the panel Itself, or by preventing its transmission to ground, or by a combination of the two. The range of audible frequencies is so high that the natural frequencies of a vibration mount can usually be designed to be well below the nolse-produeing frequency. In order to reduce nolse, try to Identify the source of the nolse, e.g. transformer hum, panel Rutter, gear tooth engagement, rotor unbalance, etc. Next identify the noise frequency. A vibration mount designed in accordance with the guidelines for vibration and shock control can then ad as a barrier either in not canduding the sound, or in attenuating the vlbration, whicli is the source of the noise. 2.0 BASIC DEFINITIONS AND CONCEPTS IN VIBRATION ANALYSIS 2.1 Kinematlc Characteristics COORDINATE - A quantity, such as a length or an angle, which helps define the position of a moving part. In Figure 1, x is a coordinate, which defines the position of the weight, W. DlSPLACEMENT - A change in pasition. lt is a vedor measured relative to a specified position, or frame of reference. The change in x (Figure 1) measured upward, say, from the bottom pasition, iS a dlsplacement. A displacement can be positive or negative, translational or rotatlona!. For example, an upward dlsplacement mav be positive; and a downward displacement negatiye. Similarly, a clockwlse rotation mav be positive and a counterclockwise rotatian negative. Units: inches, feet, or, in the case of rotations: degrees, radians, etc. VELOCm - The rate of change of dlsplacement. Units: in/sec, M.P.H., etc. Velocity has a diredion. lt is a vedor. lts magnitude is the speed. Angular velocity mlght be measured In radians/sec or deg/sec, elockwise or counterclockwise. ACCELERA110N - The rate of change of velocity. Units: in/see' etc. It is a veetor and has magnitude and direct"ion, e.g. 5 in/see' North. Angular acceleratlon mlght be measured in radians/see' or deg/sec', clockwlse or eounterclockwlse. T198

2'2-

VIBRATORY MOTION - An oScillating motior,; for example, that of the weight W, in figure 1. SIMPLE HARMONIC MOTION - A form of vibratory motlan. The motian as a fundian of the time iS of the form X = a sin wt, where a, w are constants. The maximum dlsplacement, a, from the mean pasition (X = O) is the amplitude; the frequency (rate at which motlan repeats Itself) Is w/2it eycles/sec, where has the dlmensions of reciprocal time, e.g. reclprocal seconds. The motlan is alsa cal/ed harmonic or sinusoidal motion. AMPLITUDE - flgure 2 shows a vibrating motian, which repeats itseif every T seconds. The maximum values of the displacement, x, from the reference pasition (x = O) are called amplitudes. These are (al, a2 ... ). The largest of these is ealled the pea k amplitude ..

FREQUENCY- Rate at whlch motian repeats itself per unit time. if the motian repeats itself every Tseconds, the frequency is l/T eycles per second. PERIOD, CYCLEThe intervai of time within which the motian repeats itself. in figure 2, this is Tsecongs. The term cycle tends to refer alsa to the sequence of events wlthln one period. STEADY-STATE MOTION - A periodic matlan of a mechanical system, e.g. a eontinuously vibrating penduium of constant amplitude. TRANSIENT MOTION - Amotion whieh changes with time in a non-periodic manner; orten the motian declines (attenuates) to a negligible value after a finite period oftime (e.g. impaet effects whieh decay with time, ete.). PERIODIC AND NON-PERlOoiC MOTIONS - Amotion, whieh repeats Itself is periodie; amotion, whleh does not repeat itself, is non-periodie. HARMONICS - Any motian can be considered as made up of a series of simple harmonic motions of different frequencies and amplitudes. The lowest-frequency component Is usually called the fundamental frequency; higher frequency components are ealled harmonics or super-harmonlcs. Thelr frequeneies are exact multlples of the fundamental frequency. Sametimes, components of the frequencies of which are a fraetion of driving frequencyare signlficant (e, g. the "half-frequency" whirl of rotating sharts, ete). Such components are calied subharmonics. PULSE - Usually a displacement-time or force-time funetion deseribing an input into a dynamical system. PULSE SHAPE - The shape of the tlme-dlsplacement or force-displaeement curve of a pulse. Typicaliy, this might be a square wave, a rectangular pulsei or a half sine-wave pulse. In general, however, the shape can be an arbitrary functian of the time.

Tl99

SHOCK MOTION - A motlon in which there iS a sharp, nearly sudden change in veioclty, e.g. a hammer blow on anaii, a package failing to ground from a height, ete. its mathematical ideallzatlon Is that of amotion in which the veioeity changes suddenly. The mathematical idealizatian of a sudden velocity change often represents a dose approxirnation to the real dynamic behavior of the system. 2.2 Rigid-Body Charaeteristics MASS - Weight in Ibs. divided by,the gravitational constant (g = 32.2 ft/see', or 386 in/see2). CENTER OF GRAVin - Point of support at which a body would be in balance. MOMENT OF INERTIA - The moment of inertla of a rigid body about a given axis in the body Is the sum of the produet of the mass of each volume eiement and the square of its distance from the axls. Units are in-ib-sec2, for example. Moments of inertia of the standardshapes are ,blated in h-andb06ks (see par.4.2, Mechanical System Characterlstics). If insteadof the mass of the element, the volume ls used, the result is also called a moment of inertia. Depending on the applieation, mass-, vOlume-, or area moments of Inertia can be used. PRODUCTUFINERTIA ' The 'produet ofinertia ofa rigid body abouttwo interseding, perpendicular axes in the body ls the sum of the produet of the mass (volumes,'areas) of a constltuent element and the produet of the distances of the element from the two perpendlcular axes. Units same as moment of inertia. Tabulations are available in handbooks. PRINCIPAL AXES OF INERTIA - At any po int of a rigid bOdy, mutually perpendicular (orthogonal) axes can be chosen so that the produets of inertla about these axes vanish. The orthogonal set ofaxes are called principal axes of inertia. These can often be identified by the symmetry of the bOdy, beeause the principal axes coincide with the axes of symmetry. (An axis of symmetry is a line in the body, such that the body can be rotated a fraetion of a turn about the line without changing its outline in space). 2.3 Spring and Compliance Characteristics TENSION - When a body Is stretehed from its free configuration, its partieles are said to be in tension (e.g. a stretched bar), The tensile foree per unit area is called the tensile stres s (Units: Ibs/in2), COMPRESSION - When a body is compressed from its free connguration, its partieies are said to be in compressian (e,g, a eolumn in axial loading). The compressive force per unit area is called the eompressive stress (Units: Ibs/in2). SHEAR - When a body is subjeet to equai and opposite forees, whieh are not collinear, the forees tend to shear the body in two. The body is said to be in a state of shear, e.g. a rubber pad under forees parallel to its upper and lower faees. The shear foree per unitarea is called the shear stress (Units: Ibs/in'). Most bodies in a state of stress are in tension, compresslan and shear
simultaneously, e,g. a beam in bendingo

SPRING CONSTANT - When a helical spring is stretched or compressed (amount x, say), the applied force, F, is proportional to the dispiacement, x (Hooke's law). The constant of proportlonality (k) is called the spring constant or gradient: F = kx. More generally, k is the ratio of a force inerement to the corresponding displacement inerement of the spring. In the usual helica! spring this ratio is a eonstant (independent of displaeement) within the elastic limit of the spring materia!. However, k may be variable, e.g. in a nubber pad, since the rubber is nearly incompressible. Units: Ibs/in. if the spring defieet in torsion, the units of k mav be in-Ib/radian or in-Ib/degree, ete. FORCE-DEfLECTION CHARACTERISTIC Of A SPRING - This refers to the shape of the force-defiection curve. The most familiar case is a straight line through the origin of coordinates (eonstant k). it is possibie, however, for the spring "constant" to vary. if it
increases

T200

23

with displacement (e.g. as in a rubber pad), the spring is called a h~rd spring.lfit decreases with displacement (e.g. as in a Belleville spring), the sprlng Is called a soft sprlng. ENERGYSTORAGE - This Is the area und er the force-defieetlon curve of the spring. It represents the strain energy stored iri the spring (Unlts: in-Ibs, or ft-Ibs, etc.). PRElOAD - A spring used in equipment may or may not have a rest (on the shelf) posltion in which it has its natural, free, or unstretched length. If its rest-position length is not its free length, the spring is in tension or compression. The amount of this tension or compression is called the preload. When measured in force units, it is a preload force; when measured in defiection from free posltion, it is a preload defiection. ELASTlC MODULUS (E) AND SHEAR MODULUS (G) - These are materlai properties, whlch charaeterize material compliance in tension or in compressian (E) and in shear (G). Theyare defined as the ratio of stress to strain, where strain refers to the change in length (or deformation) per unit length. E involves tensile or compressiye stress and G involves shear stress. Units: Ibs/in'. In many practical applications, E and G can be regarded as constants, withln alimit of material stress known as the proportion al llmit. Metals loaded below the proportion al limit are exampies. Rubber and plasties, however, usually have no well defined proportionai Iimit. 2.4 Damping, Friction and Energy-Dissipation Characteristics STATIC FRICTlON, SLIDING FRICTlDN, COULOMB DAMPING - These are all terms used for the frictional resistance eneountered when one body slides relative to another, e.g. a weight dragged on the ground. The frietional force is approximately proportion al to the contaet force between the two bodies and opposed to the directian of relative motion. The constant of proportionality, m, is known as the coeffieient of frietion. It a 10 Ib weight is dragged along a horlzontal fioor with a coefficient of friction, ~ 0.2, the frictional resistance is 0.2 x 10 ~ 2 Ibs. Sometlmes a distlnetion is made between the value of the coefficient of friction when motian is just impending (starting frietion) and the value during motion (kinetic friction). -The coefficient of friction in the latter case is generally somewhat lower. Table 2 shows typical values of the coefficient of friction for various materials and operating
conditions.

VISCOUS DAMPING - if a body moves relative to a second bOdy, viscous damping refers to a resisting force, which is proportional to the relative velocity between the two bodies and opposes the directlan of relative veioclty between them. The constant of proportionality is known as the coeffieient of viscous damping, c. Units: ibs per unit velocity, Le. ibs/(in/sec). Such damping is encountered, for exampie, in hydraulic dashpots and devlces, whlch meter a liquid through an orifice. The more viscous the nuid, the greater the damping. if c ~ 0.5 Ibs/(in/sec) and the body moves through a viseous fiuid at 10 in/see, the viscous damping force Is 0.5 x 10 ~ 5 Ibs. Typical example: hydraullc door c1osers. CRITICAL DAMPING - Vaiue of damping constant just sufficientiy high in a mass-sprlng-damplng system so as to prevent
vibration.

DAMPING RATlO - The ratio, of the damping constant to the critica i damping constant for that system. 2.5 Vibration Characteristics of Mechanical Systems MATHEMATlCAL MODEL - An idealized representation of the real mechanical system, slmplined so that it can be analyzed. The representation often consists of rigid masses and dashpots. Hopefully, the representation is sufficiently realistic so that the results of the analysis correspond reasonably c10sely to the behavior of the physical system from which it was derived. T201

LUMPED-AND DISTRIBUTED-PARAMETER SYSTEMS - In a lumped-paranieter system, the mass-, elastic- and damping properties are separated or lumped into distinet components each of which has only mass or oniy elasticity or only damping, but not more than one of these per component. In a distributed-parameter system, a component may have combined mass and elasticity and damping, distrlbuted continuously through the component. The latter model s ten d to be more realistic, but more diffieult to analyze. DEGREE OF FREEDOM- This Is the number of independent quantities (dimensions), which must be known in order to be able to draw the mechanlcal system in any one position, the fixed dimenslons of the system being known. The slmple mass-spring system of Figure 1 has one degree of freedom; a mechanical differential, for example, has two degrees of freedam; a rigid body moving freely in space has 6 degrees of freedom. FORCE AND MOTlON EXCITATION - If a force is applied to a dynamical system, it usually is a source of vibration (e.g. centrifugal force due toai1 unba"Wite-rbtbr).Thevibrations-are then said to 'be due to force exeitation. If, on the other hand, the foundation (or other part) of a machine is subject to a forced motion (vibration dr shock), the resulting machine vibration is said to be due to motian excitation, e.g.anearthquake actuating a seismograph. FREEVIBRATlDN--If theweight in-Figure 1 is'movedou\' of-i\'s equilibrium position, and released, the system wiil vibrate without
the actian of any external forees. Such an asciiiation is called a free vibration.

FORCED VIBRATION - If an external force is applied to the weight in Figure l, whleh causes it to vibrate (e.g. a force varying harmonically with time, say), the resulting motion of the spring-mass system is ealied a foreed vibration. if the base. which supports the spring, undergoes a forced motion, which in turn causes the weight to vibrate, the vibration is also forced. RANDOM VlBRATION - Equipment may be caused to vibrate by applled forces or motlons, the frequency (or frequency components) of which vary in arandam manner with time (e.g. wind gusts on a missiie). The resulting vlbration is called random. NATURAL FREQUENCY- When mechanical equipment vibrates freely, the resulting number of oscillations per unit time is called the frequency (cycles/see). According to whether the system is free without damping, or free with damping, the frequency is cailed the free-undamped natural frequency or the free-damped natural frequency. The natural frequency is a function of the mass distribution and compliance of the system. For a simple mass-spring system (Figure 1), which represents a reasonable approximation to many reai mechanlcal systems, the natural frequency is equal t0.a.-",,\i''-'~. ' radians per second,

IN

",il,,!

where k is the spring constant, Ibs/in; W is the weight, ibs; 9 is the gravitational constant, 386 In/sec'; and Xst is the statlc defiection of the sprlng, In. Thus, fiexlble systems tend to have low natural frequencies and rlgid systems tend to have high naturai frequencies. At the same time, the natural frequency can be changed by altering the compliance and mass distribution of the system. The simple expressions for natural frequency just giyen, yield the natural-frequency curve of the basic vibration chart given in Par. 4; Case A. In the ehart theyare plotted on a logarithmic scale and the frequency ls given in cycles per minute, rather than in radians per second. A system may have more than one natural frequency, in which case the lowest of these is often the most significant. In general, the number of natural frequeneies is equal to the degree of freedom of the system. FOReING FREQUENCY- The number of oscillations per un it time of an external force or displacernent, applied to a vibratlng system. T202