AdvancedsGontrol Ce ,SAT ADVANCED CONTROL THEORY ‘ SECOND EDITION A.NAGOOR KANi ctin Tee Beart RBA PUBLICATIONS No.71 (Old No. 32), Seshachalam Street, Saidapet, Chennai - 600 015. Phone : (044) - 24333031, 24926472, Email : [email protected]. Website : www. rbapublications.comCONTENTS — a Paye No. Chapter - 1 Linear system design : 1.1, Introduction to dasign using componsaters «x. 1.2 Lag compensator 6 1.3 Load componsator M4 1.4 Lag-Lead componsator 63 1.5 PL, PD and PID contvoltars 83 1.6 Fandback compensation 109 1.7 Short quostions and answord sine a 427 18 197 Chapter - 2 Nonlinear systems 2.1 Introduction to nonlinear systoms 1 2.2 Describing functio 150 2.3 Describing function of doad-zona and saturation nontinoarity 151 2.4 Describing function of saturation nonlinoarlty 156 2.5 — Doscribing function of doad-zon0 nonlineadty 150 2.8 — Describing function of rolay with dond-vono and hystorosis 162 2.7 Describing function of backlash nonlinoarity 167 2.8 Describing function analysis of nonlinour systams...0u 172 2.9 Review of polar plot and Nichols plot. 175 2.10 Phase plano and phaso trajectories 193 2.11 Short questions and answi 25 212 Exorcises . 225 Chapter - 3 Sampled data control systems 3.1 Introduction .. 229 3.2 Sampling procos 231 3.3 Analysis of sampling procoss in froquo 292 3.4 Reconstruction of samplod signals using hold circuits 24 4.5 Discroto sequence (Discroto timo signal) 735 9.8 | 2-Transform .. “140 3.7 Linear discrote timo systems .. 268 4.8 Transfor function of LDS systom (Patso Tronstor Function} a 4.9 Analysis of samplor and zoro-ordor Old ss.ne a ui 3.10 Analysis of systoms with impulse samplingorton 3.11 Analysis of sampled data contral systems using z-trensfarm - 3.12 The x and s-domain relationship .. 3.13. Stability analysis of sampled data contral systems 3.14 Short questions and answars.. 3.15. Exercises... Chapter-4 State space analysis 4.1 Introduction 4.2 State space formulation . 4.3. State madel of linear syste 4.4 State diagram a 4.5 State space representation using physical variables 4.6 State space representation using phase variable: 4.7 State space representation using canonical variables... 4.8 Solution of state equations 4.9 State space representation of discrete time systems.. 4.10 Short questions and answers... 4.11 Exercises... Chapter-5 Analysis and design of control systems in state space 8.1 Definitions involving matrices 5.2. Eigenvalues. and eigenvectors. 5.3. Similarity transformation. 5.4 Cayley-Hamilton theorem 5.5. Transtormation of state model 5.6 Concepts of controllability and observabi 5.7 Gontrollable phase variable form of state model 5.8 Control system design via pole placement by state feedback 5.9 Observable phase variable form of state model 5.10 State observers... 5.11 Short questions and answers 5.12 Exercises Appendix | Bode plot. Appendix II Root locus .. Index * naa. 295 296 305 a8 319 320 321 323 325 MB 360 369 384 403 a7 421 422 424 426 436 449 460 470 £89 494 504 515 519 533 546Chapter 1 Linear Systam Design L Linkar Sysran DESIGN 1.1 _ INTRODUCTION TO DESIGN USING COMPENSATORS The control systems are designed to perform specific tasks. The requirements of control system are usually specified as performance specifications, The specifications are generally related to accuracy, reiative stability and spced of response. In time domain, the transient state performance specifications are given in terms of rise time, maximum overshoot, settling time and/or damping ratio. In frequency domain, the transient state performance specifications are given in terms of phase margin, gain margin, resonant peak and/or bandwidth. The steady state requirementare given in terms of error constants. When performance specifications are given for single input, single output linear time invariantsystems, then the system can be designed by using roat locus or frequency response plots. To be more precise, when time domain specifications are given, root locus technique is employed in designing the system. If frequency domain specifications are given, frequency response plots like Bode plots are used in designing the system. The first step in design is the adjustment of gain to mect the desired specifications. In practical systems, adjustment of gain alone will not be sufficient to meet the given specifications. In many cases, increasing the gain may result in poor stability or instability In such cases, it is necessary to introduce additional devices or components in the system ter the behaviour and to meet the desired specifications. Such a redesi ofa suitable device is called compensation. A device inserted into the system for the purpose of satisfying the specifications is called compensator. The compensators basically introduce pole and/or zero in open loop transfer function to modify the performance of the system.Advanced Control Theory ification The design problem may be stated as Jolayes: When asetol specimens tare given jor aaystem . then o suifable compensator should be desianec's overall system will meet the given specification cs The compensation schemes used lor 16s) feedback contral system is either series nor parallel (feed back} tion. In series compensation, the compensator with transfer function. G(s)is placed in series with plant. In feedback compensation, the signal from some clement is feedback to the input and 2 compensator swith transfer function G¢spis placed in the resulting inner feedback path, “The series and parallel compensation schemes are shown in fig Lal and 1.2 respectively. The choice between series compensation and parallel compensation Fig. 1.2 : Feedbackiparal depestscn cower 1. Nature of signals in the system. 2. Power levels at various painls. 4. Dosigner’s experiance. 3. ‘Components available. 5, Economic considerations and 30.07. ‘The compensator may be electrical, mechanical, hydraulic, pneumatic oF other type of device armetwork, Usuially, an electric network electronic Ucvice serves as compensator in many control systems. ‘The different types of electrical or clectronic compensatars used are Lug compensator, Lead compensator ane Lay-teud compeusater, In control systems, compensation & quired in the following situations. ]. When the system és absolutely unstable, then compensetion is required 10 stobilize the system and also fo meet the desired performance, 2, When the system is stable, compensotion is provided fo obtain the desired performance. “The systems with type number 2nd above are usually absolutely unstable system: lence for systems with type number 2 and above, lead compensation is required, hecaus the lead compensator increases the margin of stability, l.Chapter 1 Linear System Desig pher Lard, slay. te pe of com pense ni pens petfarinanee He partieaar che the fallaysing sections Inv systens with ty ceases any a Ue th wl may be used to obtain the desired or is based on the factors thatare discussed in ROOT LOCUS APPROACH TO CONTROL SYSTEM DIES y: toot locus, the desired hehaviour is specified in terms of transient ications andl steady - state error requirement (the steady = state error is (1 in terms of error constants for stndard inputs, while the transient rement iy specified in terms.of peuk overshoot, settling time, rise time, uvsient respoulse specifications van be translated into desired N Ind Hesponse pipe Inorder to meet the desired specifications, the root loci are reshaped so that they ronigh the points where the dominant clased loop poles are located. The root loci haped hy intraducing a compensator. The compensator will add a pole and/or a the open loop tanster function af the syste The addition of a pole to the open-loop transfer function has the effect af pulling the root Locus to the right, whieh reduce the relative stability of the system and increase loop transfer function has the effect em is either unstable or stable but has undesirable transient response charaet lead compensatorean be employed to modify the root locus. The trarisfer mn of lead compensator will have a zero (compensating zeta) and a pole (compensating, pale). ‘The compensator zero ca ent reaponse specification which is eliminated by the compensati on real axis such that Iwo dominant closed loop pok Ifthe pole is locuted faraway from zero then it will not be effective in suppressing the noise. If the pale is tao close to zero then it will not allow the zero to do its job. In order to avoid this conflict, the pole is located at 3 to 10 times the value of zero location. ‘The lag compensator is employed when a stable system has satisfactory transient response characteristics but unsatisfactory stendy state characteristics, L.e., error requirement. The transfer function of lag compensator will have a zero (compensating zero) and a pole (compensating pole). In order to preserve the int response characteristics the compensating pole and vero should have negligible effect on shape of raot lacus. This is achieved by placing the compensating pole and zeta very close to each other, If the pale and zero are located close to the origin then the error constant will increase which will reduce the steady state error. hy adjusting the gain, In eel.Advanced Control Theory _ —_ both the transient and steady-state tion will improve the transient The lag-l characteristic ‘ad compensator is employed when © not satisf he lead compen: Tesponse and lag compensation will reduce the steady state error. nique is that the information about ined from the ____Theadvantage in design using root locus tech closed loop transient response and frequency response are dircetly obtai pole-zero configuration of the system in the s-plane. FREQUENCY RESPONSE APPROACH TO CONTROL SYSTEM DESIGN The objective of frequency domain design is to reshape the frequncy-response characteristics so that the desired specifications are satisfied. ‘The frequency domain design can be carried out using Nyquist plot, Bode plot or Nichols chart. But Bode plot are popularly used for design because they are easier to draw and modify. In design using bode plots the desired performance specifications are given in terms of frequency domain specifications and steady-state error requirement. The stability requirement is specified in terms of phase margin and resonant peak. The transient response requirements are specified in terms of ‘gain cross over frequency, bandwidth and resonant frequency. The error requirement is specified in terms of static error constants (K,. K, or K,). Note : In case the transient response specifications are given in time domain, we can translate them into frequency domain specifications using the following formulae. 26 [acrar of Phase margin, y= tan” = 1006 1 Gain cross over frequestey, ge =O n rf lac -2f ps faci frequency, ®, =n ii 20? 1 Bandwidth, op» =n [o-2 4 fac! -avaP ‘The low frequency region of bode plot provides information regarding the steady and high frequency region provides information regarding the transient- “The medium frequency (or mid-frequency) range provides information vevarding lative stability Therefore, the lov requeney region ofthe bode plot is reshaped by lag compensation to improve steady state performance. The high frequency region of the bode plot is reshaped by Icad compensation to improve transient-state performance. = Resonant peak, M, state performance state performance.es ____ Chapter 1 Linear System Design When the system requires improvement in both steady-state and transient state, sation can be employed to alter both the low and high frequency regions {The primary function of lend compensator is to reshape the frequency response Set ane Provide sufficient phase lead angle to offset the excessive phase lag associated with the components ofthe plant The primary function of lag compensator ito provide attenuation in the high frequency region to achieve sufficient phase margin, The advantage in frequency-domain design is that the effects of disturbances, Samer Oise and plant uncertainties are relatively easy to visualize and assess in frequency domain. Another advantage of using frequeney response is the ease with whiny experimental information can be used for design purposes. A disadvantage of frequency-response design is that it gives us information on closed-loop system's transient response indirectly, while the root locus design gives this information directly. 1.2 LAG COMPENSATOR A compensator having the characteristics of a lag network is called a lag sonbensator Ifa sinusoidal signal is applied to a lag network, then in steady state the output will have a phase lag with respect to input. Lag compensation results ina large improvement in steady state performance but results in slower response due to reduced bandwidth. The attenuation due to the lag compensator will shift the gain crossover frequency to a lower frequency point where the phase margin is acceptable. Thus, the lag compensator will reduce the bandwidth of the system and will result in slower transient response. Lag compensator is essentially low pass filter and so high frequency noise signals are attenuated, Ifthe pole introduced by the compensator isnot cancelled by a zero in the system, then lag compensator increases the order of the system by one. S-PLANE REPRESENTATION OF LAG COMPENSATOR SR ETRESENTATION OF LAG COMPENSATOR The lag compensator has a pole at s=—1/BT anda zero at s = —1/T. The pole-zero plot of lag compensator is shown in fig 1.3. Here, B>1, so the zerd is located to the left of the pole on the negative real axis, The general form of lag compensator transfer functivn is given by equ (1.1). Fig e # Pole-zero plot of lag compensator . sz sth Transfer function of lag compensator, G()= T woe LT) S+P. sx ele: Br where, T > Oand B>1Advaneed Contre Theory a ee of ‘The zero of lag compensator, 7 =1/'T 1.2) The pole of lag compensator, py = ‘i = 3) (1.4) From equ(1.2) we get, T= — % From equ(1.3) we get, B= Pe REALISATION OF LAG COMPENSATOR USING ELECTRICAL NE’ ‘WORK The lag compensator can be realised by the R-C network shown in fig 1.4, Let, E(s)= Input voltage and E,(s) = Output voltage In the network shown in fig 14, the input voltage is applied to the series combination of R,, R, and C. The output voltage is obtained across series combination of R, and C. Fig 14: By voltage division rule, Electrical lag compensator . (Ry + Ke) (SCR, +) /sC E)= ES) Riek, +o BUS) Fer + Ry) Wisc =E(6) (sCR, +1) [scR, +R) +1] ransfer function of the electrical network is the ratio of output voltage to The input voltage. ‘Transfer function EA) _ CR3(8+ Yen,) of electrical network} E\(s) C(Ry FRIYS+ Kernen, ] 1.6) wT) - simaclSystem Design On comparing equations (1.6) and (1.7) we get, Ey) 1 (sep) Ey) Grgh) + Shere, T= RC and B= (Ry ARR 1.8) —__— — | | The transfer function of RC network as given by equ( 1.8) is similar to the general form with an attenuation of (1/B) (since B > 1, (1/B) < 1). If the attenuation is not required then an amplifier with gain B can be connected in cascade with RC network to nullify the attenuation, : FREQUENCY RESPONSE OF LAG COMPENSATOR Consider the general form of lag compensator, . + }) _(sT+I/T (i+sT) SOD GATED IBT CespT) ell) The sinusoidal transfer function of lag compensator is obtained by letting s = jo in equ (1.9). Bi jor) (1+ jap) When =0,G.00)=B al (1.11) Geljo) = 1.10) From equ(1.11) we can say that the lag compensator provides a de gain of B (here i> 1). Ifthe de gain of the compensator is not desirable then it can be eliminated by a suitable attenuation, Let us assume that the gain B is eliminated by a suitable attenuation network, Now, G(jc) is given by ‘The sinusoidal transfer function shown in equ( 12) has two comer frequencies and they are denoted as @,, and 0. Here, @) =1/BT and We =1/T Since, BT>T, 0.1 > 1 An 20 log P= 20108 ae arntAdv ‘ed Controf Theory —_——— 5 shown in fig 1.5. The h Odbupto ©,;,, Then it tis a straight line witha constant gain nitude plot of lag compensator i magnitude plot of Bode plot of GG) is a straight line through The approximate tv has a slope of -20 db of 20 log (1B). Leto= 2 dec upto oy and ater i ZG je) ant e@T = tan o BT o> 06 30 As oad 30 Approximate nitude plot $20 tog 1/0” IGGo 9 : indb 20 log (1B) -db 0°}. 26,60) nase]. /'. Phase-plot + -+ L 1 log © wT oT Fig 1.5 : Bode plot of lag @,) compensator ‘As @ is varied from 0 to ©, the phase angle decreases from 0 to a negative maximum value of 6, at © = @,, then increases from this maximum value to 0. The phase plot of lag compensator is shown in fig 1.5. It can be shown that the frequency at which maximum phase lag occurs is the geometric mean of the two corner frequencies. tii prt Typ From the bode plot of lag compensator , we observe that lag compensator has a de gain of unity while it offers a high fequeney gain of (1/8) {In decibels it is 20 log (1 By], Itmeans thatthe high frequency neise is attenuated in passing through the network id so. the signal to noise ratio is improved. A typical choice of B= 10. Frequency of maximum phase lag, © =J@q OaChapter 1 Linear Syste Design 4 The frequency «,, can be determined by differentiating 6 with respect to w and equating dé/de to zero as shown below. From equ(1.12) we get, Phase of G.(0), §= 2G,(jo)= tanto - tan“ oBT sn 114) On differentiating the equation (1.14) we get, a 1 ‘ een ge (1.13) do 14(@T) — 1+(opT)? 8 4 : 1 + A (ano) = When 0 = @,, dé/do = 0 [noe a Fl Hence, replace « by «,, in equ(1.15) and equate to zero. wo + —— = 1+ (mT) 1+(@,BTY TT _ opr 1+(@mT)? —-1+(@,8T)? On crossmultiplication we get, 1+(@phT? =f[0+ (0, pT=0 (BT) ~BlqT)? =B-1 Ho .T)B-)=G-0) of = sogwte TS TB Frequency corresponding to maximum phase lag, mq =1/TYB san (1.16) The maximum phase angle @, can be calculated from the knowledge of f§ and viceversa, The relations between 4” and B are derived below. From equ(|.14) we get, Go) = § = tan! oP ~ tan! @BT On taking tan on cither side we get, Note ; tan (A-B) tan A ~tan 1+ tan A tanl tan §=tan [tan @T — tan @BT] tan(tan!@T)—tan(tan wT) _ er opT = —tan(tan™ @T) = tan(tan” BT) _ _ . __oT(I-B) H+tan(tan oT). tan(tan" BT) T+02 Tp 14+ B(oT)?_ es Advanced Control Theory — =e D ote tans =p aT AL@= Om. 4 bo - 1 From equ (1.16) we get, ©m ==7y 8 TW 1.17) Maximum lag angle, m = ww Z| To find the value of B from 4, From the equ (1.17). it is evident that (1-B)and 2B are the two sides of right Je triangle as shown in fig 1.6. angled triangle. Hence construct aright ang With reference to figure 1.6, sind, =(1-B)/(1+B) sind, (1+B)=(1-B) c Ac=VBC? + AB? sing, *Psindg =1-B n/ |. “oper sind, +B=1—siny A _[nFaee B (indy +1) =1-siNb C7 aa -tesinés (1.18) Fle = [asp =0+8) Tiny J : RE FOR THE DESIGN OF LAG COMPENSATOR USING BODE PLOT PROCEDU: The following steps may be followed to design a lag compensator using bode plot and to be connected in series with transfer function of uncompensated system, G(s): 1: Choose the value of K in uncompensated system to meet the steady state error Step- requirement. Step-2 : Sketch the bode plot of uncompensated system. /R ix procedure 10 sketch bode plot). ystem. [Refer Appendix-I for the Determine the phase margin of the uncompensated system from the bode plot. Step-3: a If the phase margin does not satisfy the requirement then lag compensation is required.Step-4 2 Choose aw Chaptor 1 Linear Systom Dosiga Suitable value for the phase margin of the compensated system, 1 Let. yy = Desired phase margin as given in 5 and y, = ecifications, Phase margin of compensated system. Now, y, = y+ € Where © frequency. Additional phase lag to compensate for shift in gain crossover Choose an initial value of ¢ = 5°, Step-5 : Determine the new gain crossover frequency. Corresponding to a phase margin of y,on the bode plot of uncompensated system. Let, 6... = Phase of G(joo) at new gain erossover frequency, «,., Now, 7, = 180° + §,., (Ot) d= yy 180" The new gain crossover frequency, «,,, is given by the frequeney at which the phase of (jo) is 6,..- Step-6 : Determine the parameter, B of the compensator. The value of § is given by the magnitude of GG) at new gain crossover frequency, 9... Find the db gain (A,,,) at new gain crossover frequency, 0, en A gal 20 LA, =20log8 (or) SEY = tog Bh «B= 10% Now, A,., = 20log B (or) o-=log BB Step-7 : Determine the transfer function of lag compensator. Place the zero of the compensator arbitrarily at 1/10" of the new gain crossover frequency, ©, the | r 1 Pee . ensator, Ze = = OE ©. Zero of the lag compens: eo pee 10 Now, T= gen Pole of the lag compensator, p, = 1/BT Transfer function G.6)= st -s( l+sT of lag compensator!Advanced Control Theory Step-8 Determine the open loop transfer function ofcompenseted system. The Up lag compensator is connected inseries with the plant as shown in fig 1.7. Te of lag. compensated syste Block diagram When the lag compensator is inserted in series with the plant, the open loop . i ae a crate is aenpBifed by the factor (‘= B> 1) Ie gnin produce not required then attenuator with gain 1/B can be introduced in series with the lag compensator to nullify the gain produced by lag compensator The open loop transfer function of the compensated system, 1 1, (+sT) (its) G,(3) = 5).G(s) = =. Gs 08) porose Bree yo) GespT) (8) nsated system. Calcu the compensated transf the actual function Stop-9 + Determine the actual phase margin of compen phase angle of the compensated system using at new gain crossover frequency, O- Let, $5 = Phase of G,G0) a O= Oy « margin of the compensated system, 7, = 180° + 6. he given specification then the design is dure from step 4-to 10 by taking © as 5° Actual phas If the actual phase margin satisfies 1 accepted. Otherwise repeat the proce more than previous design. PROCEDURE FOR DESIGN OF LAG COMPENSATOR USING ROOT LOCUS “The following steps may be followed to design a lag compensator using root focus and to be connected in series with the transfer function of uncompensated system. Step-1: Draw the root locus of uncompensated system. [ Refer Appendix-I1 for the procedure to construct root locus]. termine the dominant pole, s,. Draw a straight line through the origin with an 1 with respect to negative real axis. The intersection point of the inant pole, s,. ‘Step-2: De angle cos” straight line with root locus gives the dor Let this 3 : Determine the open loop gain of the uncompensated system at s = Step: gain be K. The open loop gain K at s= s,on root locus is given by jg Ritidutiot vector lengths from sq to-epen loop poles Product of vector lengths from sy to open loop zeros (vector length should be measured to scale) . aaChapter 1 Linear System Design ‘Step-4 : Calculate the parameter, B of the compensator. lis Let K,, = Velocity error constant of uncompensated system. and K,,= Desired velocity error constant. K, wo = Lt sG(s) Let A be the factor by which the velocity error constant of the system has to be increased where, A= K JK, t 7 Choose B such that it is 10 to 20% greaterthan A. “B= (1.1 to 1.2) A. Determine the transfer function of lag compensator. The zero of the lag compensator (1/T) is chosen to be 10% of the second pole of uncompensated system. Zero of the compensator, z, = 0.1 x (second pole of G(s)) Now, T= I/[-0.1 x ( second pole of G(s))] Pole of the lag compensator, p, =-1/BT. "Transfer function of lag compensator ‘Step-6 : Determine the open loop transfer function of the compensated system. The lag compensator is connected in series with the plant as shown in fig 1.8 Fig 1.8 : Block diagram of lag compensated system Open loop transfer fuction of compensated system | G(8)=G,(s).G(s) = TT ee) BT Step-7 : Check whether the compensated system satisfies the steady state error requirement. If it is satisfied, then the design is accepted otherwise repeat the design by modifying the locations of poles and zeros of the compensator,Advanced Control theory for function, G(s) = K/s(1+2s). A unity feedback system has an open loop trans! 5 40° and the steady Design a suitable lag compensator so that phase margin ! state error for ramp input is less than or equal to 0.2. SOLUTION Step- Step-2: + Calculation of gain, K. Given that, e,, ¢ 0.2 for ramp input ; Let e,,= 0.2 We know that, e, = I/K, for ramp input *. Velocity error constant, Ky By definition of velocity error constant, Ky = Lt) s G(S)H(s). Since the system is unity feedback system, H(s) ae ‘5S 2 Ky= Tt 560)" Les a9 Bode plat of uncompensated system Given that, G(s) = 5/s(1#2s) Lets=jo, -.G(Qjw) = S/ja(1+j20), Magnitude plot ‘The corner frequency is, , = 1/2 = 0.5 rad/sec ‘The various terms of G(jo) are listed in table-1, Also the table shows the slope contributed by each term and the change in slope at the corner frequency. TABLE-1 Term Corner frequency | Slope Change in slope L rad/sec db/dec db/dec 5 = - ~20. - 20Chapter 1 Linear System Design e . 15, Choose a low frequency «, such that @,<@, and choose a high frequency «, such that 0,20, Let @,= 0.1 rad/see and «,=10 rad/sec Let A =|G(jo)] in db Ato=o,. A= 20104 5-20 og 5-24 db jo) r Ao=o,, A=20104.3 5 20 log—-=20 db 805 jo Ato=ay, A - [sone from @, to wy x t0e Asa ) ¢ ce . =-40x tog i 20 = -32 db Let the points a, b and c be the points corresponding to frequencies o,, @, and @, respectively on the magnitude plot. In a semilog graph sheet choose appropriate scales and fix the points a, b and c, Join the points by straight lines and mark the slope on the respective region. The magnitude plot is shown in fig 1.1.1. Phase Plot ‘The phase angle of G(ja) asa function of @ is given by $= £G(o)= 90° -tan"'20 ‘The phase angle of G(jc) are calculated for various values of « and listed in table- TABLE-2 o radisec | 9-! 05 [10 15 10 o deg -101) -135 | -153 | -174| -177 On the same semilog sheet take another y-axis, choose appropriate scale and draw phase plot as shown in fig 1.1.1.203 4 567891 2 3 4567891 li iF Fig [1-1 : Bode plot of G{ jw) =3/ jo (1+ j20)Step-3 : | Step-4: Step 5: Step-6 : Step-7 = Chapter 1 Linear System Design Determination of phase margin of uncompensated system. 17. Let, @,, = Phase of G(jo) at gain crossover frequency (0,.). and Phase margin of uncompensated system. From the bode plot of uncompensated system we get, ¢,, = -162°, Now, y = 180° + ee = 180°-162° = 18° ‘The system requires a phase margin of 40°, but the available phase margin is 18° and so lag compensation should be employed to improve the phase margin. Choose a suitable value for the phase margin of compensated system. The desired phase margin, y,= 40°. --Phase margin of compensated system, ¥,=1,+ € Let initial choice of e = 5° p= Yat © = 40° + 5° = 45° Determine new gain crossover frequency. Let a,,,= New gain erossover frequency and ¢,, = Phase of G(jia) at o,, Now, 7, = 180° +6... 694 = Yy~ 180 = 45° ~ 180° = =135° From the bode plot we found that, the frequency corresponding to a phase of 135° is 0.5 rad/sec. -New gain crossover frequency, «,,, = 0.5 radi/sec. Determine the parameter, B From the bode plot we found that, the db magnitude at Ox, 18 20 db. +. |GGo)| in db at (@ = ,,) Kew =20db Also, A... = 20 log B = 10Ae020 = 192020 — 19, Determine the transfer function of lag compensator. The zero of the compensator is placed at a frequency one-tenth of @,.. 1_®, --Zero of the lag compensator, 2, =~ =~ bAdvan ced Control Theory a Step-8 : Step-9: Lest yg (14206) ‘Transfer function of lagcompensator; G()=—,~P7yspT (1200s) + > Determine the open loop transfer function of compensated system. The block diagram of the compensated system is shown in fig 1.1.2. The gain of the compensator ified by introducing an attenuator in series with the compensator, as shown in fig 1.1.2. Fig 1.1.2 : Block diagram of lag compensated system Open loop transfer funetion| (1, 10 (14208), 3 of compensated system O10 (1+ 200s) * s(1+2s) ___5(1+208) (1 +200s) (1+2s) Determine the actual phase margin of compensated system. 5(1+j200) ituti =joi t, jo) = —— On substituting s =o in G6) we vet G.G0)= Tsay FBO) Let,@, = Phase of G,(j@) and 4,,,= Phase of G,(0) at ©= @,,, J, = tan 200 ~ 90° tan! 2000 ~ tan! 20 At 0 = 0,545 = Ogg = (AN 20 6,4 ~ 90° = tan! 2000,,,— tan"! 2, pq = tan" (20 x 0.5) 90° — tan (200 x 0.5) — tan“! (2 x 0.5) =-140°. Actual phase margin of compensated system, Y = 180°+jcq = 180°-140°= 40° CONCLUSION ‘The actual phase margin of the compensated system satisfies the requirement. Hence the design is acceptable. RESULT ‘Transfer function of lag compensator, G,(s) = LOU +208) __(s+0.05) | (142008) ~ (s+ 0.005) Open loop transfer function of compensated system, G,(s) = 5 (1+20s) s (1+ 200s) (1+2s) oti sia dcaChapter 1 Linear System Design EXAMPLE 1.2 lie ‘The open loop transfer function of certain unity feedback control system is given by G(s) = K/s(s+4) (s+80). It is desired to have the phase margin to be atleast 33° and the velocity error constant K, = 30 see“, Design a phase lag series compensator, SOLUTION Step-1 > Calculation of gain, K Given that, K, = 30 sec! By definition of velocity error constant, K, = Lt G(s) His) Since the system is unity feedback system, H(s) = 1. wri” Kt te y= Eh SG)" hs Seay Gr BO)” ax 80 i Ke oK= ni hen Tagg e30 K=30:80%4 = 9600, Step-2 : Bode plot of uncompensated system Go = _2600___ 9600/4 x80 30 (5+ 4) (5480) 54S) (q4S) - 9{1+0:25s) (14 0.01258) 4 30 Let s=ja, .G§a) (1+ 50.250) (1+ j0.01250) Magnitude plot The comer frequencies are, «, = 1/0.25=4 rad/sec and @.,= 1/0.0125 =80 radisee. The various terms of G(ja) are listed in table-1. Also'the table shows the slope contributed by each term and the change in slope at the corner frequency. TABLE-1 2 Term [Corner frequency | Slope | Changein slope iad/see db/dee db/dec 30 = - 20. : jo N ed 1+ 70250 th} eg T+yool2sa| °* ~ Qoras Choose a low frequeney 0, such that w,< ,, and choose a high frequency «@, such that ©)>0,9. Let @,= I rad/sec and «,=100 rad/sec.20 TABLE-2 o rad/sec 1 4 10 50 80 100 ate =104| -138]-165 ~-164 | -207 = -208 | -222 | -229 =-230 Advanced Control Theory Let A =|G(jo)] in db 0 log 32 = 295 ab =30db Ato=o,, A=20 log jo Ato= oq, A=20 log] =20 og 2-175 db =18 db Ato=0g, A -[nove from og, 10 2 * ie + Aaron) et 80 = -40 log=— +18 =-34 db Ato=o,. A = [sore from @,9 to 0 x tg 2 [+A 2 100 = -60 x log—— + (-34) =-40 db O85 +(-34) Let the points a, b, ¢ and d be the points corresponding to frequencies ®,, @,,and @,, respectively on the magnitude plot. Ina semilog graph sheet choose appropriate scales and fix the points a, b, ¢ and d. Join the points by straight Jines and mark the slope on the respective region. The magnitude plot is shown in fig 1.2.1. Phase Plot The phase angle of G(jo) as a function of @ is given by $= ZG(jo) =-90° - tan10.250 — tan“"0.01250 The phase angle of Gj) are calculated for various values of « and listed in table-2. On the same semilog sheet take another y-axis, choose appropriate scale and draw phase plot as shown in fig 1.2.1. Step-3: Determination of phase margin of uncompensated system. Let, 6,.= Phase of Gi) at gain crossover frequency (@,,). and y = Phase margin of uncompensated system.Chapter 1 Linear syswin Design Le yd pos PET Met 1 lease co % ItSvep4 Step $: Step-6 : Step-7: a From the fede pot af uncompensated systenn we Four that fy = 108 Now, y= 180" + 4, HO TORE © 12 1 of atheast 33°, but the available phase The system requires a phase margi : : ie tion should be employed to improve the margin is 12° and eo Ing compensa phase margin Choose a suitable value for the phase margin of compensated system The desired phase margin, 7, " Phase margin of compensated system, 7," %y* © Let initial choice of e = 5° 5 yy" yy + © = 33° + 8° = 38% Determine new gain crossover frequency. Leto, © New gnin crossover frequency and 4, " Phase of GGo) at ©, Now, 7," 180° + 6, og y= H8O? = 38° = 180° 142° From the bode plot we found that, the frequeney corresponding to a phase of ~142" is 4.7 rad/sec. ". New gain crossover frequency, «,., = 4.7 rad/sec. Determine the parameter, From the bode plot we found that, the db magnitude at, as 16 db, |G Gj) in db at (0 = 0.) = Ay, = 16 db Also, A= 20 log B 3. = 10° = 10167 = 6,3, Determine the transfer function of lag compensator. ‘The zero of the compensator is placed at a frequency one-tenth of a. en Zero of the lag compensator, % Now, T= 10 gen he ee ccs, AT" 63x23 * Tato Pole of the Ing compensator, Pe = 1 5 3 Transfer function Jovine List 4 cain s+ of lag compensator it eat” erericnChapter 1 Linear System Design ‘Step-8 : Determine the open loop transfer function of the compensated system. (23 The block diagram Papa a of the compensated $29] 116.3} 4} as bl airoasy Gr onras system is shown in fig 1.2.2. The gain of the compensator Fig 1.2.2 : Block diagram of lag compensated system is nullified by | introducing an attenuator in series with the compensator, as shown in fig 1.2.2. Open loop transfer function} ¢. 1 63 (142.138) 30 ‘of compensated system oe G3 (1# 134195)” {1+ 025s) (1+ 0.01258) - 30 (1+2.13s) 5 (1413.4195) (1-+0.25s) (1+ 0.01255) Step-9 : Determine the actual phase margin of compensated system. On substituting s = jw in G(s) we get, sccm 30 (1+ 2.130) oe) Fo (+ j13A190) (1+ 50.250) (+ j0.01250) Let 4, = Phase of G.(jo) and 6, = Phase of G,(ja) at @ = $, = tan! 2.130 ~ 90° — tan* 13.4190 = tan“! 0.25 — tan! 0.012500 At © = 0.49.5 eco = tan! 2.13 ,,, — 90° - tant 13.4190, — tan" 0.250, — tan"! 0.0125 Deen %eco = tan! (2.13 x 4.7) — 90° — tan“! (13.419 x 4.7) = tan"! (0.25 x 4.7) tans! (0.0125x4.7) = = 147°, Actual phase margin of compensated system, 7, = 180°)4., = 180°-147°= 33° CONCLUSION The actual phase margin of the com, Hence the design is acceptable. RESULT pensated system satisfies the requirement. Transfer function of lag compensator, G(s) = 22 (1 +2-13s) _ (s+0.469) (1¥134198) ~ (6+0074) ‘Transfer function of lag compensated system, 0) EA EESAdvanced Contro! Theory EXAMPLE 1.3 The forward path transfer function ofa certain eva iuasthe given by G(s) = K/s (s+2) (6 +8). Design a suitable lag compensator $9 Taf) system meets the following specifications. (i) Percentage overshoot $ 16% for unit step input, (ii) Steady state error $ 0.125 for unit ramp input. SOLUTION unity feedback control system is ‘Step-1 : Sketch the root locus of uncompensated system. To find poles of open loop system Given that, G(s) = K/s(s+2) (s+8) The poles of open loop transfer function are the roots of the equation, s(s+2) (s+8)=0. +. The poles are 0, 2, -8. To find root locus on real axis The segments of real axis between s=0 and s= 2 and the segment of real axis between s = -8 and s = ~ will be part of root locus. Because if we choose a test point in this segment then to the right of this point we have odd number of real poles and zeros. To find angles of asymptotes and centroid Since there are three poles, the number of root locus branches are three. There is no finite zero and so all the three root locus branches will meet the zeros at infinity. Hence the number of asymptotes required is three. £180°(2q +1) a-m Angles of asymptotes Heren =3 andm=0 +180° 3 = 260° «If q=0, Angles £180°(2+1) 3 .. Angles of asymptotes are +60°, — 60° and +180° Ifq=1, Angles 180° Sum of poles~Sum of zeros _ 0-2 Centroid = nom To find breakaway point K ion, (9). SS) 56268) .d loop transfer function, = SS) __s(s+2)(s+8) _ K Close Pp R@ TG)" |, OK "5(s72)@78) +k s(S*2)(s=8)Chapter 1 Linear System Design The characteristic equation is, s(s+2) (s+8) + a) 2K ==s(s#2) (#8) =-(s? + 108? + 168) On differentiating K with respect to s we get, dK/ds = -@s? + 20s +16) PutdK/ds=0, -.3s?+20s+16=0 WV. F -4x3x p= 204 N20" ~4%3%16 |g op 5:7, 2x3 0.9, K =-(s*+10s?+1 6s) = —((-0.9)' +10 (0.9)? +16(-0.9)) =7 5.7, K=-(s'+10s*+1 6s) = -((-5.7)' +10 (-5.7)? +16(—5.7)) = -48. Fors =-0.9, the value of K is positive and real and so it is actual breakaway point, To find crossing point on imaginary axis The characteristic equation is, s(s+2)(s+8)+K=0 s#+10s?+ 16s+ K =0 Put s=jo (j@)'+ 10Go)?+16G@)+K = 0 ~jo'~ 100+ j160+K = 0 On equating imaginary part to zero we get, Stas? -jo'+j160=0 -jo? = -jl6a a? = 16 o=4Vi6=44 , Hence the root locus crosses the imaginary axis at +j4 and —j4. The complete root locus sketch is shown in fig 1.3.] Step 2: Determine the dominant pole s,. Given that, %M, =16% efi ‘We know that, %M, =e nee =0.1 and BT=10 pr ane B Transfer function of ste lag compensator sol BT Step-6 : Determine the open Joop transfer function of compensated system. The block diagram of lag compensated system is shown in fig 1.4.2and in this, the lag compensator is connected in series with G(s). 240 5 (s+10)? Fig 1.4.2 : Block diagram of lag compensated system Open loop transfer function gy (Stl) 2402400841) of compensated system on (#01) S410)? 54 10)? (§+0.1) ‘Step-7 : Check K, of compensated system Velocity error constant’ ay Ky = Lt sG,(6) of compensated system| 0 . 240 +1) 240 “anon le Sar 290 s(5+10)" (FOI) 10? x0l CONCLUSION Since the velocity error constant of the compensated system is greater than the desired value, the design is accepted, RESULT ‘Transfer function of 6.8) (+1) ‘ 3) = 4) lag compensator © (s+ 01) 7 9. — 240 (s+1) s(s+10¥ +01) Open loop transfer function ro of lag compensated system a & 33 1 =iit of Theory he 1.3_LEAD COMPENSATOR A compensator having the characteristics of a lead network is called a lead sompensator. Ifa sinusoidal signal is applied to the lead network, then in steady state the output will have a phase lead with respect to the input. The lead compensation increases the bandwidth, which improves the speed of response and also reduces the amouint of overshoot. Lead compensation appreciably improves the transient response, whereas there is a small change in steady state accuracy. Generally, lead compensation is provided to make an unstable system as a stable system. ‘A lead compensator is basically a high pass filter and so it amplifies high frequency ignals. If the pole introduced by the compensator is not cancelled by a zero in the em, then lead compensation increases the order of the system by one. S-PLANE REPRESENTATION OF LEAD COMPENSATOR jo ‘The lead compensator has a zero at s = —1/T and a pole at s = —I/aT, The pole-zero plot of lead compensator is shown in fig 1.9. Here, o < 1, so the zero is closer to the origin than the pole. The general form of lead compensator transfer function is given by equ (1.19), Fig 1.9 : Pole-zero plot of lead compensator 9 +Pec (#5) soee (1.19) where T>0 and «<1 6.) 1 The zero of lead compensator, 2 = = ‘The pole of lead compensator, p, = a From equ (1.20) we get, T Pe From equ(1.21) we get, &=Chapter 1 Linear System Design REALISATION OF LEAD COMPENSATOR USING ELECTRICAL NETWORK 35 The lead compensator can be realised by the RC network shown in fig 1.10. Let, E, and E,(s Input voltage Output voltage In the network shown in fig 1.10, the input Voltage is applied to the series combination of (RIC) and R,. The output voltage is obtained across R,. Fig 1.10 : Electrical lead compensator Rt R+t (s x) Ry R, Bo()= BQ). a * 2) eR GDR hess eek R * Rosey R,Cs+1 ‘The transfer function of the electrical network is the ratio of output voltage to input voltage. R,C Roles (Ri +R) 1oR,fro Gaeta Transfer function of] Ey(s)___Ro(RyCs+1) E(s) [RiR,Cs+R, +Ry] [ra RC at | S*LRI7(R) Ry) Ry C The geheral form of lead compensator transfer function is electrical network eon(1.24) en 1.25)26) where, T=R,C and a= The transfer function of the RC network is similar to the general form of transfer function of lead compensator. FREQUENCY RESPONSE OF LEAD COMPENSATOR Consider the general form of lead compensator, oT, The sinusoidal transfer function of lead compensator is obtained by letting s=jo in equ(1.27). (i+ jor) (1+jooT) When © = 0, G.(jo) =a 2G, (jo) = From equ(1.29) we can say that the lead compensator provides an attenuation of (Here a. <1). If the attenuation of the compensator is not desirable then it can be eliminated bya suitable amplifier. Let us assume that the attenuation ais eliminated by a suitable amplifier network. Now, G,Ga) is given by ‘ (1+ joT) 1+(@T) Zan oT 6.00) 3 (1+ je Vis (ean? Zan oat cso) The sinusoidal transfer function shown in equ (1.30) has two comer frequencies ©, and o,.- 1 1 Here, ©, = T and > "oT Since. T>aT, 0 >1 and @aT <<1 “ A® 20 log Jon = 20log(@T) At very high frequencies ie., after @,,@T>>1 and @aT >> 1 Yor? i OD = 2010 oan? Ax 20log (oat? ‘The approximate magnitude plot of lead compensator is shown in fig 1.11. The magnitude plot of Bode plot of G,(jo) isa straight line through 0 db upto @,,, then it has a slope of +20 db/decade upto «,, and after «,, it isa straight line with a constant gain of 20 log (1/c:). Let $= 2G,(jo) . eebstantoTstantoeT a (1.32) Aso>0, $30 Asa>o, $30 Approximate __ magnitude plot +db 1G,Ge| indb 0 Fig 1. : Bode plot of lead compensator ease]. 26,0) o| ®, 0, 0, 1c? loge T Wa otAdvanced Control Theory __ 3] A\s @ is varied from 0 too, the phase angle increases from 0 to a maximum value of, at @ =, then decreases from this maximum value to 0. It can be shown that the frequency at which maximum phase lead occurs is the geometric mean of the two corner frequencies, Frequency of maximum phase lead, ©, = q-@g = oe . a at Ta The choice ofa. is governed by the inherent noise in control systems. From the Bode plot of the lead network, we observe that the high frequency noise signals are amplified by a factor 1/a., while the low frequency control signals undergo unit amplification. Thus the signal/noise ratio at the output of the lead compensator is poorer than at its input. To prevent the signal/noise ratio at the output from deteriorating excessively, it is recommended that the value of a should not be lessthan 0.07. A typical choice of c = 0.1. Also it is advisable to provide two cascaded lead networks when @,, required (i.e., phase lead required) is more than 60°. Determination of ©,, 4, and o& The frequency w,, can be determined by differentiating 6 with respect to w and equating d6/da to zero. From equ (1.32) we get, 132) Phase of G.Go), 6= 2G, (jo) = tan oT- tan! aoT On differentiating the equation (1.32) with respect to w and equating dg/de to wwe get the frequency corresponding to maximum phase lead as, «, = 1/T Jer - 1 Frequency corresponding to maximum phase lead, © == T= zero, ‘Also we can express @,, in terms of a and a. in terms of $,, as shown below. (Note : The equations (1.33), (1.34) and (1.35) can be derived by a similar analysis shown in section 1.2 for lag compensator affer replacing fi by’ a).PROCEDURE FOR f GN OF 1 AD COMPENSATOR USING BODE PLOT The following steps ma be followed to design a lead compensator using bode plot and to be connected in series with transfer function of uncompensated system, Gis). Step-1 : The open loop gain K of the given system is determined to satisfy the requirement of the error constant. Step-2 : The bode plot is drawn for the uncompensated s\ determined from the previous step. (Refer Appendiv: bode plot). stem using the value of K. -Lfor the procedure to sketch Step-3 : The phase margin of the uncompensated system is determined from the bode plot. ‘Step-4 : Determine the amount of phase angle to be contributed by the lead network by using the formula given below, m= Maximum phase lead angle of the lead compensator yg = Desired phase margin Y= Phase margin of the uncompensated system = Additional phase lead to compensate for shift in gain crossover frequency Choose an initial choice of € as 5° (Note : If 2, is more than 60° then realize the compensator as cascade of two lead compensator with each compensator contributing half of the required angle) ‘Step-5 : Determine the transfer function of lead compensator - Calculate & using the equation, From the bode plot, determine the frequency at which the magnitude of Gio) is -20 log I/ Ja db. This frequency is « ‘alculate T from the relation, «44 =—y= aleulate T from the relation, = == aaa a(list) ‘Transfer function a 56 : (Least) lead compensatorAdvanced Control Theory Step-6 : Determine the open loop transfer function of compensated system. The lag compensator is connected in series with G(s) as shown in fig 1.12. When the lead network is inserted in series with the plant, the open loop gain of Fig Lia: Block diagram of lead the system is attenuated by the compensated system factora| <1), soan amplifier . with the gain of I/c has to be introduced in series with the compensator to nullify the attenuation caused by the lead compensator. Open loop wanster function 1 S44 aes of the overall system ve 1 a (L+sT) G a (esoty 0“ Wasaty Step-7: Verify the design. Finally the Bode plot of the compensated system is drawn and verify whether it satisfies the given specifications. If the phase margin of the compensated system is less than the required phase margin then repeat step 4 to 10 by taking € as 5°more than the previous design. PROCEDURE FOR DESIGN OF LEAD COMPENSATOR USING ROOT LOCUS The following steps may be followed to design a lead compensator using root locus and to be connected in series with transfer function of uncompensated system, G(s). Step-I : Determine the dominant pole, s, from the given specifications. Dominant pole, s4=-{ w, + jeqy1-C? . [Note : if 6 alone is specified and ©, is nat available, then draw the root locus and from the root locus find the dominant pole, Refer example 1.8). Step-2 + Mark the poles and zeros of open loop transfer function and the dominant pole on the s-plane, Let the dominant pole be point P. Step-3 : Determine the angle to be contributed by lead network. Let, 6= Angle to-be contributed by lead network to make the point P as a point on root locus. . Draw vectors from all open loop poles and zeros to point P. Measure the angle contributed by the vectors. (For the procedure to find angle contribution by vectors refer root locus in Appendix I].Sum of angles Now, 4°] contributed by potes | ~ ofuncompensated system) (of uncomper where nis an odd integer so that n1 80° is nearest to,the difference between 10 be contributed is more than 60 then realise the compensator ax cascade of tro dead compensators with each compensator contributing half of the required angle] Step-4 : Determine the pole and zero of the lead compensator, gin of s-planc and point P be the dominant pole. Draw ich that AP is parallel to x-axis as shown in fig 1.13. 7 APO [ZAPO] where the point C is on Let point © be the o straight fines OP and A\ Draw a ling PC soasto b the real axis, With line PC draw angles BPC and CPD such that cach equal to 9/2. F Dare located on the real axis. Now the point B is the location of the pole of the compensator (-[/cTyand the point Dis the location of the zero of the compensator (-1/T), From the values of point D and B compute T and ct, Step-5 : Determine the transfer function of lead compensator ‘Transfer function of lead compensator Fig 1.14 Block diagram of lead compensated system ‘Step-6 : Determine open loop transfer function of lead compensated system. ‘The lead compensator is connected in series with the plant as shown in fig 1.14. Open loop transfer function of compensated system, G,(s) = G,(s)G(s). ‘The open loop gain K is given by the value of gain of s= s,. The value of gain, K is determined from pole-zero plot of lead compensated system and by using the magnitude condition given below.42 Product of vector lengths from all poles to 8 = 54 Product of vector lengths from all zeros to s = Sy (Note : The length of vectors should be measured to scale. For details of magnitude ee rroot locus in appendix 1) Cheek whether the compensated system satisfies the error requirement. If the error requirement is satisfied then the design is accepted. Otherwise repeat the design by altering the location of poles and zeros by trial and error, without changing the value of 6. Note: If the open loop gain K is specified in the problem, then take the gain at ©, 48K, Find a parameter, A where A= K,/K. Now introduce an amplifier with gain, A in cascade with compensator to account for reduction in gain due to attenuation by parameter, a. Now, G,(s)-" A.G (s).G(s)- | ign a phase lead compensator for the system shown in fig 1.5.1 to satisfy the following specifications, (i) The phase margin of the system 245°. (ii) Steady state error for a unit ramp input Fig 15.1 | 1/15. (iii) The gain crossover frequency of the i: system must be lessthan 7.5 radésec. : SOLUTION Step-1 : Determine K | Given that, steady state error, e,, $ 1/15 for unit ramp input When the input is unit ramp, ¢, = 1/K,= 1/15... K,=15 By definition of velocity error constant, K, we get, Ky = Lt s.G(s) H(s) S50 K Lt s =K Here, G(s) = and H(s)=1, * 2 (5) ep 330 (s+ 1) Step-2 : Draw bode plot i K 15 | Given'that, GO)= S75 seth 15 Let s= joo, 000) * 55014 ja)Chapter 1 Linear System Design Magnitude plot 43 The comer frequency IS, ,, = | rad/sec. The various terms of Go) are listed in table-1. Also the table shows the slope contributed by each term and the change in slope at the comer frequency. TABLE- Term | Corner frequency | Slope | Change in slope rad/see db/dee db/dec 15 ry : . -20 ‘ el -20 Choose a low frequency «, such that ,<¢9,, and choose a high frequency ‘©, such, thata,>e,,. Let «,=0.1 rad/sec and 0,=10 rad/see =|G(jo) in dd Letus calculate A at@,,0_,ando,. Let A [13] At@=@; =O rd/sec, A=201o; 20 log AS =435 db = 44 db ‘jel ol At@=0@, =Irad/sec, A=20 wel At@=@y =10 rad/sec, xs from @4 10 ©, wots] Ao j Oy _ 23.5 db = 24 db =—40 x logins 16 db Let the points a, band ¢ be the points corresponding to frequencies ©), 0,,.and ©, respectively on the magnitude plot. In a semilog graph sheet choose appropriate seales. and fix the points a, b and c. Join the points by straight lines and mark the slope on the respective region. The magnitude plot is shown in fig 1.5.2.ave eee OY 1 TReOn S40 phase Plot The phase angle of Gerd asa timiction ofr is given by’ 8% AGG) = 80" — tates cote an! liste in tables? arionis valties The phasoangle of Gijon) are calculated for 10 =1s3 | -160 | =174 xis, choose appropriate scale and ‘On the same semilog sheet take another y draw: phase plot as shown in fig LS.2. rgin of uncompensated system, Step-3 : Determine the phase m Let, 8. = Phase of G(jeo) at gain crossover frequency (@,.). and y= se margin of uncompensated system, From the bode plot of uncompensated system we get, 6, = “167°, 45°, but the available phase margt Id be employed to improve the phase margin. The system requires a phase 13° and so lead compensation shou! Step-4 + Find @ The desired phase margin, y, 2 45° Let additional phase lead required, € Maximum lead angle, 6, = y,— y+ € 45° 13° + 5° = 37° Step-5 Determine the transfer function of lead compensator. losing _ 1sin37* _ g24g6 = 025 sing, 1+sim. ‘The db magnitude corresponding £0 ©,Step-6: Step-7 : Chapter 1 Linear System Design From the bode plot of uncompensated system the frequency «, corresponding toa db gain of -6 db is found to be 5.6 rad/sec, <0, = 5.6 rad/see. Now, T= a= = 0357 = 036 ©, wee ane _ 1+sT) _ p9,(1+0365) s+ (I+saT) (14 009s) ot 1 Transfer function of es the lead compensator °° Open loop transfer function of compensated system. The block diagram of the lead compensated system is shown in fig 1.5.3. “The compensator will provide an attenuation of @, To compensate for that, an amplifier of gain I/a is introduced in series with compensator. Fig 1.5.3: Block diagram of lead compensated system Open loup transfer function oe _1_ 025 (140363) 15 of compensated system 0.25 (140.098) s (+1) 15 (1+0.36s) ~'$(1¥0.095)(1+s) Draw the bode plot of compensated system to verify the design. 15 (1+ j0.360) jo (1+ j0.090) (1+ jo) Puts=jo in G(s), -. Go(jo) Magnitude plot ‘The comer frequencies are @,,,@,, and ©,. = | rad /see ; Og = Tage 8 mad see; Og = Gpp =! rad/ seeAdvanced Control Theory “OF + DOs = (OND fo s01d apog + eS Mt¢ japter 1 Linear System Design ‘The various terms of G, (ja) are listed in table-3, Also the table shows the slope |47 contributed by each term and the change in slope at the corner frequency. rer frequency rad/see Slope Change in slope dbidec db/dec 5 jw 1 1 jo 140,360 1 14 j0.090 Chooue a low frequency co, such that ©,<.0,, and choose a high frequency such that ,>0,,. Let, 0.1 rad/see and «50 rad/see Let Ay =|Go(jo)} in ab Is Mom, #04 rud/sec, Ay 2a 15 = 20 log 57 = 435 db~ 44 db ALO > Wey 1 rad / see, y= 20 = 23.5 db ~24 db Mow «24 rad (sec, Ay =f fiom 0,4 10 ¢9 * ieee ].(@ st ) Og] \O=Oy} 2, = 407 log? 4.24 =6 db © ain at ALO MM rad sec, Ag © [sn fiom 04 a9 tg2a](8 } o=o, <2 ey i = 207 lop——+6=-6 *loeret db ALo* ey, » 50 1ad/s0¢, Ay ~| slope from «5 t0 0, x log 22- | (280 Oey O = Oe3 =-40% ton“? +(-6)= 32 dbAdvanced Control Theory —48] Let the points a, b, d, e and fbe the points corresponding to frequencies 0, and &, respectively on the magnitude plot of ‘compensated system. The magnitude plot of compensated system is drawn on the same semilog graph sheet by using the same scales as shown in fig 1.5.2. Phase Plot The phase angle of G,(ja) asa function of « is given by 6, = 2G,(ja) = tan“10.360 ~90° — tan“!0,090 — tanto. The phase angle of G,(ja) are calculated for various values of. and listed in table-4, TABLE-4 radisee | 9! | os | 1 2 5 10 % ; deg 94 | 109] -120 | -128 | -132 } -142 In the same semilog sheet and by using the same scales, the phase plot of compensated system is sketched as shown in fig 1.5.2. Let, 4,9 = Phase of G,(iw) at new gain crossover frequency. and y, = Phase margin of compensated system. 134°, From the bode plot of compensated system we get, 4... = Now, r= 180° +64. = 180° -1342 = 46° CONCLUSION The phase margin of the compensated system is satisfactory. Hence the design is acceptable, RESULT i _ 025 (140.368) _ (s+2.78) The transfer function of lead compensator, G(s) = T0085) +0096) 7 ‘en shy Open loop transfer function]. _15(1-+036s) of lead compensated system{ °” 5 (1-+0.09s) (+s)Chapter 1 Linear System Design AMPLE 1.6 Design a lead compensator fora unity feedback system with open loop transfer function, G(s) = K/s (s+1)(5+5) to satisfy the following specifications (i) Velocity error constant, K, > 50 (ji) Phase margin is > 20°, SOLUTION Step-1 : DetermineK Given that, K, 250, Let K,=50 By definition of velocity error constant, K, we get, K K G(s)= Lt s—*>__=4 Esse) 590 SFI) GF) 5 J K=SxK, =5x50=250, ‘Step-2 : Draw bode plot ‘ , K 250 hat, G(s) = ——K___250_ Stren thes: ae) S(6+1) #5) S(t) (45) _ 250 _ 50 s(I+s)x5x (148/95) s(1+5) (140. 50 Let s =ja, * G(jo) “jo (iF jo) (¥J020) Magnitude plot ‘The comer frequency are, w,,= 1 rad/sec and @,= 1/0.2= 5 radisec, The various terms of G(ja) are listed in table-1. Also the table shows the slope contributed by each term and the change in slope at the corner frequencies. TABLE-1 ‘Term | Corner frequeney | Slope] Change in slope rad/see db/dee db/dec 20. = SY 40. 21 10 ~ 4020 =~6 {40Advanced Control Theory Choose a low frequency «, such that @,<,, and choose a high frequency, such that o>0,). Let @,= 0.5 rad/sec and o,=10 rad/see Let A =|G(jo)| in db At@=o,, A=20 log) 0 log 22 = 40 ab [5 At@=0,,, A=20log>2| =20 log =34 db al At@=05, A -[ from @¢) to @,9 x log 28 |e haan j s @ey - =-40x log? +34 =6 db At@=Oy, A = nove from 0.9 to ©) x 162] Aan ) Me * =~ 60% log i246 = 12 db Let the points a, b, c and d be the points corresponding to frequencies OO 49 4, and @, respectively on the magnitude plot. In a semilog graph sheet choose appropriate scales and fix the points a, b, ¢ and d. Join the points by straight lines and mark the slope on the respective region, The magnitude plot is shown in fig 1.6.2. Phase Plot The phase angle of G(ja) as a function of w is given by $= ZG(ja) = -90° - tan“ o — tan'0.20 ‘The phase angle of G(jo) are calculated for various values of w and listed in table-2. TABLE-2 adisee |! | 05 | 10} 5 | a0 ‘ if -96 | -122 | -146 | -214 | -238 On the same semilog sheet take another y-axis, choose appropriate scale and draw phase plot as shown in fig 1.6.2.___ Chapter 1_Linear System Design © margin of G(jw) at gain crossover frequency. and y = Phase margin of uncompensated system. From the bode plot of uncompensated system we get, ¢,, =—224°. Now, y= 180° +g), = 180° — 224° = 44° ise Margi of the sys The pha m is negative and so the system is unstable. Hence lead compensation is required to make the system stable and to have a phase margin of 20°, Find ¢,, ‘The desired phase margin, 7,2 20° Let additional phase lead required, ¢ = 5° Maximum lead angle, 6, yt @ = 20° - (44°) + 5° = 69" Since the lead angle required is greaterthan 60°, we have to realise the lead compensator as cascade of two lead compensators with each compensator providing half of the required phase lead angle. bm & oS a34s° > Step-5 : Determine the transfer function of lead compensator © 148i, 1 1 ‘The db magnitude corresponding f0 0, =~20 log = =-20 los = From the bode plot of uncompensated system the frequency, «corresponding toa db gain of -5.5 db is found to be 7.8 rad/sec. 8 rad/sec. tranefer fanetion of Transfer function | geek the lead compensator} +0.28 x 0.248) 51Advanced Control Theory $21" Step-6 : Open loop transfer function of compensated system. The block diagram of the fon Gael _ $0 Froos? |x #9) F025 compensated system is shown in fig 1.6.1. Fig 1.6.1 : Block diagram of lead compensated system The attenuation provided by the compensator can be retained to reduce the large value of open . 5 loop gain, so that the unstable system can be easily brought to stable region, Open Toop transfer function] a7e4(1+0245)° | 50 of compensated system 000678)?” s (#8) (1025) . 4 (1+0.24s)? “S09 (1402s) (140.0678? ‘Step-7 : Draw the bode plot of compensated system to verify the design. ; 2 Puts jo in Go(S), ::Go(jo)=—__4 (1+ 0.24a)* _ - Jo (1+ ja) (1+ j0.20) (1+ j0.0670) Magnitude plot The comer frequencies are@,,, 04,0, anda, 1 . 1 1 @y=13 OQ= 77 =42 3 Wy= Hs, g- a 1 <2 = O04 3 O59 33 Oy 10g] 15 The various terms of G,(je) are listed in tal ible-3. Also the table shows the slope contributed by each term and the change in slope at the comer frequency. Bi TABLE-3 Term (Corner frequency | Slope Change i i rad/see db/dec dhides Aho : -20 ‘ 1 ine (| @a=l ~20 1 (140.240)? | = ta 72 1 1+ j020 _ (1+ j00670)—____Chapter 1 Linear Syste Choose a low frequene: ©, stich that «, < @ such that o>«, »,, and choose a high frequ rad/see and @,= 30 rad/sec Let Ay = [Go(io)) in db 4 4 = 20 log —=18 db a ° 05 I+ | = 20 tog =12 ab io Hi ALO=0), Ag =20 logl Ato=a, + Ao =20 log| At@=0Q, Ag -| ire from 0-4 0 Wp x wate], at (o = 0) et = 40x tog? +12 =-13 db Dea - ALO=0,5, Ag =| spe from 2 to 3 x (ee +Ag at (0 = 0.) = 0x log 5+ (-13)=-13 ab Pca - At@=O4, Ao -| sive from ©, 10 04 woe Ag at(o=0,5) = -20 tog “2 + (-13) =-225 db ~ -23 db Oy “ —* | + Ag at (= 0,4) Ato=Oy, A=[sore fom ou too xue2*] 0 30 = 60x logs e+ (-23)=-41 db iand j f ding to frequencies 0, i dj be the points correspon: } Let the pins © ee respectively on the magnitude plot of compensated oe nitude plot of compensated system is desi onthe same rie ee by using the same scales as shown in fig 1.6.2. semilog gr O43 Phase Plot i is given by C j@) as a function of w is given by areal — tanr!0,20 ~ 2tan-"0.0670. «oy = Dtanel 90° — tan by ZGGjo) = 2tan"0.240 ious listed in table - 4. £G,() are calculated for various values of and liste ‘The phase angle of Gywe fos = (ODD fo 10rd apog = e°9°1 Ht 2 aes == =| a 7 2 “gt = ae ee a SSI = | = = “Ss 4 Chapter 1 Linear System Design TABLES 55 o rad/see JON} 05 | 1.0} 20 | 5 10 15 46,() deg 1-94 | -112. -150 | -170 In the same semilog sheet and by using the same scales, the phase plot of compensated system is sketched as shown in fig 1.6.2. Let, 6.9 = Phase of G,(jo) at new gain crossover frequency (o,.,). and y, = Phase margin of compensated system. From the bode plot of compensated system we get, 6, =~ 140°, Now, yp = 180° + %sc0 = 180° -140° = 40° CONCLUSION ‘The phase margin of the compensated system is satisfactory, Hence the design is acceptable, RESULT . 0.0784 (1+0.24s)?__(s#4.17) (s+14.92)? * The transfer function) of lead compensator J * 4 (1+0.24s)? © 8 (1s) (1+0.2s) (1+ 0.0678)" Open loop transfer function of lead compensated system| EXAMPLE 1.7 Consider a unity feedback system with open loop transfer function, G(s) = K/ s(s#8). Design a lead compensator to meet the following specifications. (i) Percentage peak overshoot = 9.5%. (ii) Natural frequency of oscillation, «, = 12 rad/sec. (iii) Velocity error constant, K, 2 10, ©, Le SOLUTION ¥ 7 v Step-1 : Determine the dominant pole, s,. Dominant pole, 54 = fio, +jtyyl—@ » Given that, o, = 12 rad/see and %M, = 9.5%. MIE 199 2 go v BR We know that, %M,,Advanced Control Theory 56 On taking natural log we get, -Gr/J1-¢? = dn (0.095) Cn? 2: On squaring we get, + oo (in 0,098) = 5.54, C2n? = 554-5540? On? +5540? = 554 O(n? +554) =554 Sq = —(0.6 x12) + jl2x ¥1-0.6? = ‘Step-2 : Draw the pole-zero plot The pole-zero plot of open loop transfer function is shown fig 1.7.1. Poles represented by the symbol "x", The pole at point P is the dominant pole, ‘Step-3 : To find the angle to be contributed by lead network, Let $= Angle to be contributed by lead network to make point, P as a point on root locus. sum ofangles sum of angles Now, 6=| contributed by poles |-| contributed by zeros [+ n180° of uncompensated system) | of uncompensated system From fig 1.7.1, we get, ‘Sum of angles contributed = 0, +05 =127°485°= 212° brpokset mcomeeaned seen} 8) 405 =127YES= 21 Since there is no finite zero in uncompensated system, there is no angle contribution by zeros. = 212° +n 180° Let n=1, 4g =212°- 180° = 32°, Step-4 : To find the pole and zero of the compensator Draw a line AP parallel to x-axis as shown i 1.7.1. The bisector PC is drawn to bisect the angle APO, The angles CPI) and BPC are constructed as shown in fig 1.7.1. Here CPD = ZBPC = d/2 = 32°/2 = 16°.Chapter 1 Linear System Design ig 1.7.1. : Pole ze| 38) ‘Step-6 : Determine the open loop Advanced Control! Theory From fig | Pole of the compen: Zero 0} We know that, We know tha Transie er fietion of @= lead compensator | 7S" transfer function of lead compensated system. The block diagram of k compensated system shown in fig 1 Open loop transfer function) | of lead compensated system) © Here the value of K is given by the value of gain at the dominant poles, onthe root locus. From magnitude condition K K = Product of vector lengths from all poles to s Product of vector lengths from all zeros to's From fig 1.7.1. we get. \Step-7 : Check for error requirement Forthe compensated system. the velocity error constant is given by 150(8+9.1) = Lt sGy(@9= Ly s SEA) “so §(6+8) (8+16.25) 10s2 Chapter 1_Linear System Design ince the velocity error constant of the compensated system, satisfies the requirement, the design is acceptable. SULT lead compensator ‘s+1625) "(1 +0.06s) Transfer function “| G.(s)= (s+9.) (1+0.11s) 150 (89.1) 105 (1+6.1 1s) 8) = 5 (48) (416.25) s(1¥0.1258) (1+ 0.065) Transfer function of lead compensated system of MPLE, Design a lead compensator for a unity feedback system with open loop transfer function G(s)= K/s (+4) (s +7) to meet the following specifications. (i) % Peak overshoot = 12.63%. (ii) Natural frequency of oscillation, «, = 8 rad/see. (ii) Velocity error constant, K,2.5. ; SOLUTION Step-1: Determine the dominant pole, 5, Dominant pole, sy =-C®,tj0,y1-C > Given that, @, =8 rad/see and %M, = 12.63%. We know that, %M, =e NE x 100 Sali-S 0.1263 On taking natural log we get, —Cr/y1 -C? = In (0.1263) 2,2 On squaring we get, a = (in 0.1263)? =428. On? 2428-4280 or 44.280? = 4.28 ew +4,28) =4.28 (055 x8) + j8x V1- 0557 j6.68 SION 59Advanced Control Theory eee!) Step-2: Step-3 : Step-4 : Step-5* Draw the pole-zero plot ‘The pole-zero plot of open loop transfer function is shown fig 1.8.1. Poles are represented by the symbol he pole at point P is the dominant pole, s,, To find the angle to be contributed by lead network. Let = Angle to be contributed by lead network to make point, P as a point on root locus. sum of angles sum of angles Now, $=] contributed by poles |-| contributed by zeros | n180° of uncompensated system) | of uncompensated system From fig 1.8.1, we get, Sum of angles contributed } =6)+0) 48) aera aEe by poles of uncompensated system Since there is no finite zero in uncompensated system, there is no angle contribution by zeros. 9 = 285° + n180° 1.9 = 285° — 180° = 105°, Since the angle contribution is more than 60°, the lead compensator is realised as cascade of two compensators with each compensator, contributing half of the required angle. 1.9 = 105/2 = 52,5° = 52° To find the poles and zeros of the compensator Draw aline AP parallel to x-axis as shown in fig 1.8.1. The bisector PC is drawn to bisect the angle APO. The angles CPD and BPC are constructed as shown in fig 1.8.1. Here ZCPD = ZBPC = /2 = 52°/2 = 26°, From fig 1.8.1. Pole of the compensator, p, = -13.55 Zero of the compensator, z, = —4.65 We know that,z,=-I/T T= 1/4.65=0.215 We know that, p.=-VorT aT = 1/13.55 (or) = WT x 13.55 = 0.343, Determine the transfer function of lead compensator (s+4.65)? (s+1355)? Transfer function “| Get lead compensatorChapter 1 Linear System Design zero plot of lead compesated systerm.Advanced Control Theory — Sepe6 : Determine the open loop transfer function of lead compensated system, (44. 05)" (9 13.55) The block diagram of lead 2: Black dlagram of lead compensated sysrent Open loop trans 14.65)? “ K 155)? s(s-h4) (S47) of lend compensated s Here the value of K is given by the value of root locus. From magnitude condition K is given by, in at the dominant pole, s, on the From fig 1.8.1, we get, xb KN xy BIXGK*73%2* ILS 2 : 2675 = 685 685 (8:4-4.65)" s(s44) (84-7) (6+13,55)° Step-7 + Check for error requirement Gols) For the compensated system, the velocity error constant is given by, K,= Ll sG(s)= Ls —— GES 44.65)" ___ 6854.65" _ agp sO s(8+4) (S47) (8413.55) dx 7x 1355" GONGLUSION ince the velocity ¢rror const requirement, the design is nt of the compensated aye, satisfies the ‘eptable, RESULT ‘Transfer fiction “| . = 0.1178 (140.2155)? lead compensator Gus 3557" Pe canaapt .07 385) ‘Transfer function of Gols) = 885 (644,65)? lead compensated system, s BH) Grae “Taaaaeee tats? _, (10.255) (140.1435) ¢140.07385)"L Chapter 1 Linear System Design 1.4 LAG-LEAD COMPENSATOR A compensator ha he characteristics of lag: compensator. Ina lag-lead network when sinusoidal signal is phase lead occurs in the output, bul in different frequency revions. Phase lay, occurs in the low frequency region and phase lead occurs in the high frequency region (i.€) the phase angle varies from fay to lead as the frequency is increased from zero to infinity. ipplied, both phase lag and A lead compensator basically increases bandwidth and speets up the response and decreases the maximum overshoot in the step response. Lag compensation increases the Jow frequency gain and thus improves the steady state accuracy of the system, butreduces the speed of responses due to reduced bandwidth, Iimprovements in both transient and steady state response are desired, then both a Jead compensator and lagcompensator may be used simultancously, rather than introducing both a lead and lag compensator as separate elements. However it is economical to use a single lag-lead compensator. A lag-lead compensation combines the advantages of lag and lead compensations. Lag-lead compensator possess two poles and two zeros and so such a compensation increases the order of the system by two, unless cancellation of poles and zeros occurs in the compensated system, jo §-PLANE REPRESENTATION OF LAG-LEAD COMPENSATOR The s-plane representation of lag-lead compensator is shown in fig 1.15. The lag section has one real pole and one real zero with pole to the right of zero. The lead section also has one real pole and one real zero but the zero is tothe right of the pole. Fig 1.15 : Pole-zero plot of tag-lead compensator G+YT) (6+) sv WPT) (8+ 1/aT,) ee) ees, Tagecetion lead section Transfer function of Jag - lead compensator where B> Land 01,7, =R,C,, T,=R,C, and FG FREQUENCY RESPONSE OF LAG-LEAD COMPENSATOR Consider the transfer function of lag-lead compensator. (s#1/T)) +1/1 SG em, ‘The sinusoidal transfer function of lag-lead compensator is obtained by letting s=jo in equ (1.46). (1+ jo) (1+ jo) + G40) ="B 767) (1+ joa) wn( 147) Fora single lag-lead compensator, a8 = 1. Hence from equation (1.47) we can say that the lag-lead compensator providesa dc gain of unity. (1+ joT)) C+ jot) “GG = 7 joBT,) = ioaT) 48) 65jal tra ‘on shown in equ (1.48) has four corner frequencies and they are ©, ©. ©,, and @,,, where 0,6 0,505 O° 1 3 Os Fa and O.4 a Here = al, BT, By an analysis similar to that of lag compensator the bode plot of lag-lead compensator is sketched as shown in fig 1.17. (6,Go] in db Fig 1.17: Bode plot of lag-lead compensator wet, WT, Wt, Wer, PROCEDURE FOR DESIGN OF LAG-LEAD COMPENSATOR USING BODE PLOT ‘The lag-lead compensator is employed only when a large error constant and a large bandwidth are requited. First design a lag section and then take « = L/B and design a lead seotion. The step by step procedure forthe design of lag-lead compensator is given below, Step-1 : Determine the open loop gain K of the uncompensated system to satisfy the specified error requirement, Step-2 : Draw the bode plot of uncompensated systen ‘Step-3 : From the bode plot determine the gain margin of the uncompensated system hase of Gia) at gain crossover frequency, hase margin of uncompensated system, Now, y= 180° + @. If the gain margin isnot satisfactory then compensation is required. Step-4 : Choose a new phase margin Let, y,= Desired phase margin Now, new phase margin, +, Choose an initial value of ¢ = 5°,er Chapter Linear Syst Step-5 : From the bode plot, determine the new gain crossover frequency, which is the |? frequency corresponding to a phase margin of Let. @ = New gain crossover frequency and 6, = Phase of G(jo) at o,,, ty = 180° +4, (01) },= /,— 180° In the phase plot of uncompensated system, the frequency corresponding to a phase of }, , is the new gain crossover frequency ©, Choose the gain crossover frequency of the lag compensator, @,.,, somewhat greaterthan @,., (i.c., choose @,., such that @, et gen) Step-6 : Calculate B of lag compensator. Let, Aw = |G(jo)jin db at o= Oy From the bode plot find A,,, Now, A,,=20logB (ot) parole’ 20) Step-7 : Determine the transfer function of lag section “The zero of the lag compensator is placed at a frequency one-tenth of o,,,. ¢.Zero of lag compensator, z,, = 1/T, = @,,,/10 Now, T, = 10/0,., Pole of lag compensator, p,, = !/BT, ‘Transfer —— G\)= (s+/T) _ 9 (+sT) of lag section (s+1/BT) © (1+sBT) Step-8 : Determine the transfer function of lead section Take, 0. = 1/8 From the bode plot find @,, which is the frequeney at which the db gain is -201og(1/ Ja.) Now T= Dm ‘Transfer function). (81) '2(8) = = — =O (s+ lark section \_Advanced Control Theory lead compensator. 88! Step-9 : Determine the transfer function of la Transfer function of lag- lead compensator, G(s) = G,(s) x G(s) =p Cts), M+sh) (l+sBT) — (+ser (+s) (l+sTy) Se) = GBT) (Isa) ‘Step-10 : Determine the open loop transfer function of compensated system. The lag-lead compensator is connected in series with G(s) as shown in fig 1:18. (l+sT)) (1+sT, (1+sBT)) (1¥saiT,) Fig 1.18 : Block diagram of lag-lead compensated system, Open loop transfer function + Ip P fi hou (1+sT)) (1+sT) xG(9) of compensated system (1+sBT,) (Isa) ‘Step-L : Draw the bode plot of compensated system and verify whether the specifications are satisfied or not. Ifthe specifications are not satisfied then choose another choice ofa. such that, o < 1/B and repeat the steps 8 to 11. PROCEDURE FOR DESIGN OF LAG-LEAD COMPENSATOR USING ROOT Locus ‘The lag-lead compensation is employed to improve both the transient and steady state responses of a system. First design a lead section to realize the required and «, for error constant has to be increased then design a lag section, The step-by-step procedure for the design of lag-lead compensator is given below, Step-1 : Determine the dominant pole, 84 51 =-Co, £0, 0-1 where, ¢ = Damping ratio @, = Natural frequeney of oscillation, rad/sec,Chapter 1 Linear System Design ‘Step-2 : Mark the poles and zeres of open loop transfer function and the dominant pole on the s-plane, Let the dominant pole be point P. Step-3 : Find the angle to be contributed by lead network to make the point P as. point on root locus. Let, ¢ = Angle to be contributed by lead network to make point, Pasa point on root locus Draw vectors from all open loop poles and zeros to point P. Measure the angle Contributed by the vectors. (For the procedure to find angle contribution by vectors refer root locus in Appendix Il]. sum of angles sum of angles Now, ¢=| contributed by poles -| contributed by zeros £n180° of uncompensated system) (of uncompensated system. where n is an odd integer, so that n180° is nearest to the difference between angles contributed by poles and zeros. Step-4 : Determine the pole and zero of the lead section. Let point © be the origin of s-plane and point P be the dominant pole. Draw straight lines OP and AP such that AP is parallel to x-axis as shown in fig.1.19. Draw a line PC so as to bisect the angle APO [ZAPO] where the point C is on the real axis. With line PC as reference, draw angles BPC and CPD such that each equal to ¢/2. Here the points B and D are located on the real axis. ‘Now the point B is the location of the pole of the compensator (~1/aT,) and the point D is the location of the zero of the ‘ZAOP. AN compensator (-1/T,). Compute T, and a 4 from the values of point D and B. Step-5 : Determine the transfer function of lead section. seh 1 é Transfer function) ¢ of lead section ( Step-6 : Determine the open loop gain, K. The open loop gain K is the value of gain at s=s,. The value of gain, K is determined from pole-zero plot of lead compensated system and by using the magnitude condition given below.70 Advanced Contro! Theory Product of vector lengths from all poles to Product of vector lengths from all zeros to (Note : The length of vectors should be measured to scale. For details of mognituds condition refer root locus in Appendix 11). Open loop transfer function Goals) = G2(5) « G(s) of lead compensated system! Step-7 : Determine the velocity error constant of lead compensated system. 1 sGa Velocity error constant of] 0 lead compensated system IfK,, satisfies the requirement then only lead compensation is sufficient but if K,, is lessthan the desired value then provide lag compensation. ‘Step-8 : Determine the parameter, B of lag section. Let, K_,= Desired velocity error constant. A= The factor by which K, is increased. Now, A=K JK, select B, such that B> A. [i.e.,B=(1.1 to 1.2)xA] ‘Step-9 : Determine the transfer function oflag section, Choose the zero of lag section as 10% of the second pole of uncompensated system. +-Zerovf lag section, z,, = 0.1 x second pole of G(s) Also, Z<1 i 1 Pole of lag section, pe = —~ BT, 1 stat Trahsfer function of lag section, G(s) ( z]