5 TIME RESPONSE ANALYSIS, DESIGN SPECIFICATIONS AND PERFORMANCE INDICES ————— Control systems are generally called upon to perform both under transient (dynamic) and steady conditions. A feedback control systems has the inherent capability that its parameters anno Re can be adjusted to alter both its transient and steady-state behaviour. In order to analyze the transient and steady-state behaviour (these two together are referred to as time response) of control ystems, the fiset top awa is to obtain a mathematical model of the system, For any Speci inpal ¢ signal a Soaps time response expression can then be obtained through the Laplace transform inversion (or through convolution integral in case the input is such that we tain its Laplace transform). This expression yields the steady-state behaviour of the with the time tending to infinity. In case of simple deterministic signals, steady-state response can also be obtained directly without obtaining the time response expression by use of the final value theorem. Before proceeding with the time response analysis of a control system, it is necessary to test the stability of the system (the concept of stability shall be explained in Chapter 6). System stability can be tested through indirect tests without actually obtaining the transient response In case the system happens to be unstable, we need not proceed with its transient response analysis. Usually the input signals to control systems are not known fully ahead of time. In most cases these signals may be random in nature, e.g,, in a radar tracking system, the position and speed of the target to be tracked may vary in a random fashion. It is thus difficult to express the actual input signals mathematically by simple equations. The-characteristics of actual signals which severely strain a control system are a sudden shock, a sudden change, a constant velocity and a constant acceleration. System dynamic behaviour for analysis and design is therefore judged and compared under application of standard test signals—an impulse, a steP, a constant velocity (a ramp input) and constant acceleration (a parabolic input). 194RESPONSE TE ANALYSIS, DESIGN SPECIFICATIONS AND PERFORMANCE INDIGES ba The nature of the transient nature is dependent upon system = nse is revealed by any one of these test signals as this les and not upon the type of input. steady state can be quickly determined by the final value theorem as illustrated in Section 5.5 and will also elaborated in other sections of this chapter. | As explained above control systems are inherently time-domain systems subject to time- varying inputs and are to be tested, analyzed and designed by the time-domain test signals like step, ramp and parabolic. The time-domain command signals in a control can also be visualized as (through Fourier Transform) a band of sinusoidal signals of frequencies from de upwards (control systems are low-pass filters). So another important test signal for control systems is the sinusoidal signal which can be easily generated and its frequency varied. Steady- state sinusoidal response of a control system over a range of frequencies yields a great deal of information about the system; both about its time-domain response and its stability, as there is normally a good correlation between the frequency-domain response and time-domain response of a system. Chapters 8 and 9 will be wholly devoted to steady-state sinusoidal response of control systems. The time response performance of a control system is measured by computing several time response performance indices as well as steady-state accuracy for the standard input signals described above. These indices give a quantitative method to compare the performance of alternative system configurations or to adjust the parameters of a given system. As a given parameter is varied, various performance indices may change in a conflicting manner. The best parameter choice would thus be the best compromise solution. Certain of the performance indices may be specified as upper or lower bounds in a design. a Step Signal The step is a signal whose value changes from one level (usually zero) to another level A in zero time. The mathematical representation of the step function is vr) = Au) where u(t)=1;¢>0 ABD) =0;t<0 In the Laplace transform form, R(s) = A/s The graphical representation of a step signal is shown in Fig. 5.1(a). Vamp Signal The ramp is a signal which starts at a value of zero and increases linearly with time Mathematically, it) = At; t>0 =0;t<0 (5.2)ae CONTROL SYSTEMS ENGINEERING In the Laplace transform form, Ris) = Als? The graphical representation of a ramp signal is shown in Fig, 5.1(b). From eqns, (5.1) and (5.2), it is seen that a ramp signal is integral of a step signal. my 7) "0 A A A + ° to to @ ) © > t 0 A t (c) @ Fig. 5.1. Standard test signals. Parabolic Signal The mathematical representation of this signal is vt) = At/2 5 t>0 3t<0 =n(5.8) In the Laplace transform form, Rls) = Als? The graphical representation of a parabolic signal is shown in Fig. 5.1(c). From eqns. (6.2) and (6.3), itis seen that a parabolic signal is integral of a ramp signal. Signal Aunit-impulse is defined as a signal which has zero value everywhere except at ¢ = 0, where its magnitude Tes te. It is generally called the 6-function and has the following property: Vat) =0;t#0 +e J S(t)dt = 1 je where € tends to zero. a perfect impulse cannot be achieved in practice, it is usual. pulse of small width but unit area as shown in Fig. 6.1(@). Mathematically, an impulse function is the derivative of a step function ie., Vat) = wit) The Laplace transform of a unit-impulse is 18) =1= Res) The impulse response of a system with transfer function O(s/R(s) = Cs) = G(s) Ris) = Ga) or et) = LIG(s) = gt) lB) ‘wThus the impulse response of a system, indicated by g(t), is the inverse Laplace transform of its transfer function. This is sometimes referred to as weighting function of the syste G(s), is given byweighting function of a system can be used to find the system's response input r(¢) Y e .m’ to it Xt) though its Laplace transform cannot be found) by mears af the cevwolution integra a Thus at) = fee ~or(ade AB.) Let us consider the first-order system of Fig, 6.2 with unity fedback. As an example, this can be visualized to be the block diagram of the pneumatic system of Fig. 2.16. Whose dynamics is described by the Laplace transform of eqn. (2.36) wherein one let T = RC, R(s) = 4P,(s) and (5) = APs). _ _Ats) 1 | os) = APG. §.2. Block diagram of a first-order system ‘The transfer function of the system from Fig. 5.2 is co), 1 (5.6) \ Rs) Ts+1 In the following sections we shall analyze the system response to unit-step and unit- ramp inputs assuming zero initial conditions. ‘A sponse to the Unit-Step Input For the unit-step input [R(s) = V/s], from eqn. (5.6), the output response is given by 1 T 1 CW) = tse) “5 Teri Taking the inverse Laplace transform, we get Vet)= 1-2" which is plotted in Fig. 5.3. t o z Jiig. 5.3. Unit-step response ol a first-order system.CONTROL SYSTEMS: 198 ENGINES ing, Itis seen that the output rises exponentially from zero value to the final value of urgy The initial slope of curve at ¢ = 0 is given by de - ae) ake dt\|.o T where T : known as the time constant of the system. time constant is indicative of how fast the s; ystem tends to reach the final Value. The speed of response can be quantitatively defined as the time for the output to become a particular percentage of its final value. A large time constant corresponds to a sluggial system and 4 ‘small time constant corresponds to a fast response as shown in Fig. 5.4, Se time constant corresponds to a fast response 1 wr] i T 1-0 ‘WBIG. 5.4. Effect of time constant on system response. Consider now, the error response of the system which is given by \ et) = rt) = c(t) = et teady-state error is given by (e= lim e(#) =0 Thus this system tracks the unit-step input ) which is also plotted in Fig. 5.3. The s t with zero steady-state error. «Flesponse to the Unit-Ramp Input From eqn. (5.5), the output response for the unit-ramp input [R(s) = V/s2] is given 1 Tf Cs) = wee a "(Ts +1) ree Teri Taking the inverse Laplace transform of the above equation, we get Welt) =t- 11-67) The error signal is et) = rit) ~ et) = 11 — ev) and the steady-state error is given by eq = lim e(t) = 7 z Thus the first-order system under conside ich i ration will track the unit-ramp input with # steady-state error T, wl is equal to the time coi tof the zoit-ram inp‘TE RESPONSE ANALYSIS, DESIGN SPECIFICATIONS AND PERFORMANCE INDICES patente sree time constant therefore not only improves its speed of response but also ces its stea i -atate error to ons me WHE 5.5. Unit-ramp response of a first-order system If we examine the derivative of c(t), i.e., et) = 1-e*? we find that it is identical to the system response to the unit-step input. The transient response to the ramp input signal thus yields no additional information about the speed of response. of the system. We therefore need examine only the steady-state error to the ramp input which can be obtained directly by the final value theorem as given below: V5 lime(t)= lims E(s) = lim R(s)— C(s)] Be seer er a : tig o(Ts+ sl r ‘This avoids the need to obtain the inverse Laplace transform resulting in a considerable labour saving in higher-order systems. : Consider the servomechanism shown in Fig. 5.6, which controls the position of a ical load in dance with the position of the reference shaft. The two potentiometers connical input and output positions into proportional electrical signals, which are in turn compared and an error signal equal to the difference of the two appears at the leads coming from potentiometer wiper arms. ‘The error signal (voltage) is 0, = K(r-0) where r = reference shaft position in rad; c = output shaft position i sumtihettytn : put shaft position in rad; and K, = potentiometer ‘The error signal is amplified by a factor K, by the am . a apa e Plifier and is appli circuit of the d.c. motor whose field winding is excited with a consta linen as fe obana ic,'S AND PERFORMANCE INDICES Krk, where f= fo + Re The block diagram can now be simplified to the form of Fig. 5.7(b), where n K= KKKrR ‘The forward transfer function can be written in the time constant form as Gs) = Ky *sGs+D where K, = Kif, t= Jif. Fig. 5.7(b) is in fact a standard second-order system involving one forward path integration. As shall be explained later in this chapter, this configuration belongs to a general dass of systems called type-1. \ Response of Second-order System to the Unit-step The analysis, now onwards, is not restricted to the particular system used as an example above but is valid for a general second-order type-1 system. From Fig. 5.7(b), the overall transfer function of the system is Cs) K, K,/t == 2. =— > 5.7) RG)” w+s+K, 24154 Ko te It can also be written in the standard form 2 Ee) a BO) (6.8) RG) s*+20,s+@2 (8) A 3 al i = dampi ieee where {= damping factor (or damping ratio) = Tees = oT and @, = undamped natural frequency = {K,/t = \(K/J)- ee ee The time response of any system is characterized by the roots of the denominator polynomial q(s), which in fact are the poles of the transfer function. The denominator polynomial (s) is therefore called the characteristic polynomial and ~aqs)=0 -(5.9) is called the characteristic equation. The characteristic equation of the system under consideration is 8? + 20,8 + 0,2 =0 ‘The roots of this characteristic equation are given by 6? + 20,8 + 0,2) = (88 Xs —5,) (5.10) For (<1, M8 =~ $0, #40, [1-27 =- Co, + joyCONTROL SYSTEMS ENGINEERING Ree ae where @, = @, (1 £2), is called the damped natural frequency. ‘ i trol systems are designed wit} Most control systems with the exception of robotic cont th damping factor {< 1 to achieve high response speed consistent with other performance measure, empiained later in thi ‘ is chapter). eintans From eqn. (5.7), for the unit-step input, the output response is given by 2 Cs) = oe A511) ef 2 als+ lo, ~ jo, 1-5) Ils +50, + JO, V(1- 67) ‘The Laplace inverse* of eqn. (5.11) is obtained, by the method of residues (refer Appendix Das oF c(t) = ——»_ __| 8? +2&0,8+ 02 Is=0 oF o-Weat gin V-Le sls +0, — jO,VI-O7,_ go, — jo, fa la-¢?) $ The steady-state value of c(t) is given as ye = Him eft) =1 The time response of an underdamped ({ < 1) second-order system is plotted from eqn. (5.11) in Fig. 5.8. It is damped sinusoidal. The response reaches a steady-state value of c, = 1, ie, the steady-state error of this system approaches zero. The time response for various values ' ETE we 1 all 7) t+ tan a ;t>0 (5.12) of plotted against normalized time o,¢ is shown in Fig. 5.9. The system breaks into continuous oscillations for ¢ = 0, as can be seen from eqn. (5.12). The mathematical expression for the time response in this case is given by Velt) = 1-cos wt t et. Unit-step response of underdamped second-order system sit can also be written down directly from Table 12, item 17 of Appendix 1TIME RESPONSE ANALYSIS, DESIGN SPECIFICATIONS AND PERFORMANCE INDICES As ¢ is increased, the res] till it becomes cxitically = 1 and becomes overdamped for { > 1, Robotic (5.27) fa 0 5? + 5°G(s) 290 8G) Ky where K, = tm sG(s) is defined as the acceleration error constant. yf of Feedback Control System F The open-loop transfer function of a unity feedback system can be written in two standard forms—the time-constant form and the pole-zero form. In these two, forms, G(s) is as given time-constant form and Se pe below.Ge) = KEas + DT + De ime-constant form) "(7,18 + W(T,28 + D ++(5.28) = Kiet avery). (pole-zero form) (529) "(8 + pis + De ‘The gains in the two forms are related by Tz; K=K AB Tp; 30) With the gain relation of eqn. (6.30) for the two forms of Gis), it is sufficient to obtain steady-state errors in terms of the gains of any one of the forms. We shall use the time constant form in the discussions below. Equation (5.28) involves the term s" in the denominator which corresponds to number of integrations in the system. As s tends to zero, this term dominates in determining the steady. state error. Control systems are therefore classified in accordance with the number of integrations in the open-loop transfer function G(s) as described below: \ACType-0 System Ifn = 0, the steady-state errors to various standard inputs, obtained from ens. (5.25), (6.26), (5.27), and (5.28) are (Position) = ——— = —— «+ _ (5.81) eta tone ¢a( velocity) = lim TS = (acceleration) = lim ¥G@ 7” Thus a system with n = 0 or no integration in G(s) has a constant position error, infinite velocity and acceleration errors. The position error constant is given by the open-loop gain of the transfer function in the time-constant form. \2Type-1 System 5 Ifn = 1, the steady-state errors to various standard inputs are ro. 1 1 ¢q(Position) = ay = ys =0 ay etm. 11 slim—t_ L121 32) calves) = lin 2 = ae AB 1 a( acceleration) = lim = = « Thus a system n = 1 or with one integration in G(s) has zero position error, a constant velocity error and an infinite acceleration error at steady-state. t aT Fype 2 System If n = 2, the steady-state errors to various standard inputs are[ TIME RESPONSE ANALYSIS, DEGION SPECIFICATIONS ANO PERFORMANCE INDICES 213 (position) = —+— #0 aE): a (velocity) = lim a" . 1 1 ¢,(aceeleration) = lim -— "Fan "CE (6.33) ; Thus a system with n = 2 or two integrations in G(s) has a zero position error, zero velocity error and a constant acceleration error at steady-state. Steady-state errors for various inputs and systems are summarized in Table 5.1. «Fable 5.1. Steady-State Errors for Various Inputs and Systems Types Type of input , Steady-state error Type-0 system Type-1 system Type-2 system Unit-step V1 +K,) 0 0 Unit-ramp i vk, ‘5 Unit-parabolic - é UK, : : Sng x, = lim Gos) K, = lim 0G(s) K, = lim sG(s) For nonunity feedback systems (Fig, 5.18) the c difference between the input signal R(e) and feedback —~”. eo signal B(s) is the actuating error signal E,(s) which is given by [refer eqn. (5.20)] 24s) wa 1 a ae” E,©)= T Gee) omen FG. 5:18 Therefore, the steady-state actuating error is sR(s) ou = ies 1+ G(s)H(s) +-(5.34(b)) The error constants for nonunity feedback systems may be obtained by replacing G(s) by G(s)H(s) in Table 5.1. error constants K,, K, and K, describe the ability of a system to reduce or eliminate tate errors. As the type of system becomes higher (ie., increasing number of steady. ‘ations), progressively more steady-state errors are eliminated. Although it appears that integrations), progressively more Steacy State Crrors are etiminate’ there is no limit to the number of integrations, types-0, -1 and -2 are the most commonly employed systems in practice, Systems of type higher than 2, i, with more than two te wns are not employed in practice because of two reasons. be discussed in detail in Chapter 9). ;) The dynamic errors for such systems tend to be larger than those for types-O, -1 and -2, although their steady-state performance is desirable.i basic types-step, ramp and para : steady-state error when inputs are other than the three | s 2 Seothor ulty is t the error constants fail to indicate the exact manner in which error function change with time. The dynamic error may be evaluated using the dynamic coefficients—the concept generalized to include inputs of almost any arbitrary function of time (refer Problem 5.13). The expressions for time response specifications established in Section 5.4 are valid only for the second-order closed-loop transfer function of the form given in eqn. (5.8). Let a zero at 8 =~z be added to this transfer function. Then we have Cs) __(s+zXw5/ 2) CO) Sete ng! 2) -(5.35(a)) RG) 5? +2f0,8+ 02 = Note that multiplication term in the numerator of above expression has been adjusted 80 that steady-state gain 0) of the system is unity. This gives the steady-state value of output c,, = 1 when input is unit step. Thus the system will track the step input with zero steady-state error. From eqn. (5.35(a)) we have Cs) o 8 iia) 2 saree cate :( } (5.3506) Let c,(t) be the response of the system with a zero at s =—z. Then, from eqn. (5.35(5)), we have ld Ve = alt) + Fe) where c(t) is the response given by eqn. (5.12). e effect of added derivative term may be seen by examining Fig. 5.19 (a) where a case for a typical value of { (ess than one) is considered. We see from this that the effect of the zero is to contribute a pronounced early peak to the system's response overshoot may increase appreciably. From eqn. (5.35(c)) and Fig. 5.19 (b) it is seen that the ler the value of z, i, the closer the zero to origin, the more pronounced is the henomenon. On account of this fact, the zeros on the real axis néur the oniei voided in design. However, in a sluggish system the artful introduction of a zero --(5.35(e)) are at the proper position can improve the transient ‘We further observe from eqn. (5.35(c)) that as z increases, ie, the zero moves farther into the left half of s-plane, its effect becomes less pronounced. For sufficiently large values of z, the effect of zero on transient response may become negligible. For the closed-loop transfer function of eqn. (5.35(a)), peak percentage over-shoot for 8 unit step input can be read from the log-log graphs of Fig. 5.19(6) as function of (z/S«,) for various values of ¢ <1. vJAG 5.19(@). Effect of closed-oop zero on unt-step response of a second-order system. 10; 5.0} 20 ls os 0.4 1 2 5 10 20 50 100 200 500 1000 Percent overshoot Fig. §.19(b). Peak overshoot to unit step input of a second-order system with closed-loop zero. accuracy (opecified in terms of permissible error ,,), damping factor ¢ (or peak overshoot step input, M,) and settling time ¢,, Ifthe rie time ¢, i alvo specified jt should be constant with the specification of ¢, as both these depend upon ¢ and @,. Steady-state accuracy Fequirement is met by a suitable choice of K,, K, or K, depending upon the type of the system. ‘As explained earlier, the damping factor sufficiently leas than ones preferred in most controlRn o> 6 Ws) Tus) us) Ouls) Aes) -1 Fig. 5.31 The governing differential equation can be written from the transfer function as JO yft) + fa yt) + KpK Kp Oy(t) = KpK Kp O_(t) The system characteristic equation is 82+ (Js + KpK,Kz/J = 0 KpK Ky 2. —PRATT a3 = 06x Ky x2 et totais sl or 10? = 04 (©) With tachogenerator (switch ‘S’ closed) _ KpK4Ky ds? +(f+KaK,Kr)s+ KpK,Kp 2le, = (f+ K,KKVI ; o, = ((KpK Kyl) a LLL K AK Kr VOKpK,Kr) 24+5x2xK, 2 f0.4 06x65 x2) Example 5.4 : The system illustrated in Fig. 5.32 is a unity feedback control system with 4 minor feedback loop (output derivative feedback). (a) In the absence of derivative feedback (a = 0), determine the damping factor and natural fequency. Also determine the steady-state error resulting from a unit-ramp input. (6) Determine the derivative feedback constant of which will increase the damping factor the system to 0.7. What is the steady-state error to unit-ramp input with this setting of the derivative feedback constant ? ; (c) Ilustrate how the steady-state error of the system with derivative feedback to un ramp input can be reduced to the same value as in part (a), while the damping factor maintained at 0.7. or K, = 33.3 amp/volt on On (s) IH p9} or 1s or —_K, = 0.11 Wrad/sec.Solution. (a) With a = 0, the characteristic equation is (8 +2)+8=0 or 824294820 ©, = 8 = 2V2 rad/sec 2a, = 2 t= 0.388 22 oma k, 5278 e,, (to unit-ramp) = 1/4 = 0.25 ® With derivative feedback, the characteristic equation is 8 3(s + 2) a a 1+Gs)=1 * s+) sor s+ (2+ 8a)s+8=0 2a, = 2 + 8a 2x 0.7 x 2/2 =2+8a System K, « 6A2 + 8a) = (2 + 8aV/8 = 0.495 Go) Lae ths gain inte rar op bel tste wa higher vein The new characteristic equation is s?+(2+aK,)s+K,=0 26@, = 2+ aK, 2x0.7/K, =2+aK, (5.59) System K, = K,K2 + aK,) A560) , = (2 + aK,VK, = 0.25 Solving eqns. (6.55) and (5.56) we obtain K, = 31.36; a = 0.186 Alternative solution is obtained by adding an amplifier of gain K, between the two samming blocks. The characteristic equation now becomes B/a(e+2) 1+ G)= 14K, 15 gay / aia) 7° or #94 (2 + Bas + BK, =0 D 2a, = 2+ 8a 2% 0.7 x22 (Ky, =2+80 (5.61)ES memo oa, System K, = 8K,/2 + 8a) e eg = (2 + 8aV8K, = 0.25 Solving eqns. (5.57) and (6.58) we obtain K, = 3.9250 = 0.73 Note. The second solution requires a smaller gain but a separate amplification stage, Example 5.5 : Consider the liquid-level system shown in Fig. 5.33. The pump controls the liquid head A by supplying liquid at a rate Q@ m*/sec to the tank of cross-sectional area | m? ¥ shall assume that the flow rate Q is proportional to the error in liquid level (desired leva), actual level). Under these assumptions, the system equations are: - , His)_ 1 @MQ=hor Qs) ~s Gi) Q = Ke; e = error in liquid level, K = gain constant The block diagram representation of the system is given in Fig. 5.34. Let us pose the problem of computing the value of K that minimizes the ISE for unit-step input. (5.62) Desired Actual J+ s —__4 Fig. 5.33. A liquid-level system. Controller Process Fig. 5.34. Block diagram representation of Fig. 5.33. From Fig. 5.34 we have Es) __s R(s) ~ s+KExample 6.12: A unity feedback (negative) system has open-loop transfer function Kk GS) = 542) ue of gain K so that the closed-loop system has a steady-state, . ramp err Ode sae ae necorresponding damping factor and percentage perk, oma to unit step input. () Ths ane is now modified to include a forward path zero at s = - 6. What is the ney value of K for the same steady-state error as in part (a). Also calculate the damping factor nj find the percentage peak overshoot corresponding to this damping factor and that found fron the curves of Fig. 5.19 (b). Solution. (a) Steady-state error to unit ramp input 1 _y, (+2) = lim ——=lim f= 0 G6) 190K 2 ” or Cet Td (given) or K=20 Characteristic equation s(s+2)+K=0 or s?+2s+K=0 or s? + 2s +20=0 2 =— = = 0.2: 5 2/20 26 eM NI-F 0.486 or 486% (®) Zero added at 6 Modified Ge) = KE+9) s(s +2) Unit ramp input —e,, = lim —1_-_2_ gy = a0 sG(@) 6K =O which gives K=108 Notice that gain of 6 is contained in (s + 6) ; so the system’s K, remains unchanged. Characteristic equation s(s +2) +K(s +6) =0 K= 10/3 gives s* + (16/3)s + 20 =0 20, €= 16 ©, = —— =06 2x3 J20 Ignoring the effect of zero M, = ee at £=06 gives M, = 9.4%, considerable reductionLet us now take into account the effect of ze1 ig. 5.19 () ae Zero,2z=6, a, = 0.6 20 z 6 fo, 06 V20 72 $= 06 We obtain from this figure M, ~ 16% We find that the effect of zero is to increase M, from 9.4% to 16%. Example 5.13. A certain system is described by the differential equation j+by+4er Determine the value of b to satisfy the following specifications. 1. M, to be as small as possible but no greater than 15%. 2. Rise time ¢, to be as small as possible but no greater than 1.2 sec. If both specifications can not be met simultaneously at least (1) should be met and (2) to be met as closely as possible. Solution. Since specification (1) must be met, we choose M, = 15% M, = eela-F) Take log natural on both sides -né| fa-2) =-19 ng = 19 \(1- €?) Squaring we get We = (1.97 (1- &) 9.872 = 3.61 (1- 2) or E=0.52 From the given differential equation o, = 4 =2rad/s =0.15 Rise time or 1,=1.22 sec. 9 = 1.036 rad We find that ¢, condition is met. In this example the two specifications are such that both are met with one parameter é (corresponding to 6). This is usually not the case. Example 3.14. The block diagram of a robot joint control is drawn in Fig. 5.47. Various system parameters are : J=10 kg/m, f= 20 Nm/rad/sec. ‘K,, (motot torque constant) = 2 Nm/V K, (velocity feedback constant) = 1 V/rad/sec K (error amplification) in V/degree, to be determined. (a) Calculate the value of K, for the closed-loop systemto have § = 1.[ -T(s) (8) Fig. 5.48 is . 1™)=2 As = 5 6,88) = lim 50,(s) = 2° im, Kn 10 8. = = - (88) = Egy = 0-826 degree © Ts) = 0, 6, (8) = 3 The steady-state error will be zero as there is one integration in the forward path Thus oss) = 0 Example 3.15: The block diagram of a unity feedback control system has K OS) = +8) Find the value of K such that ITAE performance index is minimized. For this value of K determine the peak overshoot to unit step input. Solution. From Table 5.2 optimum closed-loop characteristic equation is s4+140,8+0,2=0 Ai) For the given system C(s)__G(s)_ K wii) RG) 1+G(s) s?+8s+K Comparing the denominator of Eq. (ii) with Eq. (@), we have o, = JK 140, =8 cs 14/K =8 = K=327 8 _ 8 5° o ym 2Ja27 or &=0.7 . M,= -BNO-F) 9.046 or 4.6% It is noticed that IATE-based system has rather high damping. Example 3.16. A unity feedback control system has an open-loop transfer function of 2(s+8) G9) = (542) :NEERING | (a) Write the closed-loop transfer function C(s¥R(s) and find there from systems damping factor and natural frequency and unit peak overshoot. (6) Calculate c(t) for unit step input. () Calculate e,, for unit ramp input. (d) Remove the zero s = — 10 and increase the gain to 16. Find now € and @,. (e) Calculate c(t) for unit step input. Cls)__ Gs) Solution. (a) 8)= FO =F Gq Cls)__2(8 +8) Rs) s*+48+16 fi) 4 o,= VIG =4, §= 557 =05, M,=0.163 or 163% or @) RG) = 3, then 5 2(s +8) s(s? +48 +16) - =——2_,__16_ . 8445416 s(s?+4+ 16) Looking up Appendix I Table 1.2 items 15, 17 are reproduced below. Cs) = ai) 2 On RY ele oes 3€<1 8? +2£0,8+@2 on Bt gai say oi sinlw, \-F*)t+0): 36? +260,s+a0!) Ja-e° We can express C(s) as 46 1 46 Oe) = Fa 4s416) 8 sa 45016 0,=4, £=05, fo, =2 Va-8) = 0.866, @, (a-) = 3.464, 9-607, ~~ -402 (a-&) ‘Taking inverse Laplace transform c(t) = 1— 1.155e™ sin (3.464¢ + 60°) + 0.155e sin 3.464¢ we) It can be shown by plotting e(¢) that the third term in Eq. (iv) caused M, to be higher than that given by E = 0.5 because of the presence of zero (s = 8) , 1 (c) Unit ramp input, Ris) = > s’(@ Zero 8 =- 8 removed, then Se s(s+2) characteristic equation s?+2s+16=0 and baa = 0.25, M,=0.444 or 44.4% It is observed that & reduces from 0.5 to 0.25 by removing the zero while M, increases from 16.3% to 44.4% ) Rs) = = C(s)_ 16 R(s) s?+2s+16 16 s(s” + 2s + 16) Its Laplace inverse gives C(t) = 1 - 1.033e~ sin (3.87¢ + 75.5°) wv) As there is no open-loop zero M, = 44.4% will be confirmed by § = 0.25. Example 5.17: Reconsider the position control system of Fig. 2.54. Determine the transfer function 6,(s)/6,{s). Also find the time response of the system for a unit step position input i.e. @{s) = Us. Given : Cs) = J=4x10%kgm? K,=60 f= 2x 10 Nm/rad/sec. K, = 0.15 Virad Kee K,=05Nm/A — n= 1/10 =52. Solution. We will a by drawing the signal flow graph of the block diagram of Fig. 2.54 and then use Masons gain formula 96) 1 Ky KF wis’) 3 Oh (8) Forward path gain P, = “4422