Scribd Logo
Upload

Welcome to Scribd!

  • 86823-Chapter 13

    Uploaded by

    qqaqq
    14 views54 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    14 views54 pages

    86823-Chapter 13

    Uploaded by

    qqaqq

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 54
    CHAPTER 13 Control theory: analysis Chapter objectives ‘When you have finished this chapter you should be able to: = understand poles and zeros as valuable tools for analysis and design of control systems i= find the step response of a first-order control system; i= find the step response of a second-order control system; = use the Routh-Hurwitz criterion to determine the stability of control systems; = determine steady-state errors for a unity feedback system; '= determine steady-state errors for a non-unity feedback system. 13.1 Introduction This chapter demonstrates the applications of the system representation by evalu- ating the transient response from the system model. The concept of poles and zeros, which is a valuable analysis and design tool is first discussed, then we show how to analyze our models in order to find the step response of first- and second-order control systems. 13.2 System response Nowadays, it is possible to implement continuous control systems directly using computers. If the computer is fast enough (a close enough approximation 449 450 Mechatronics Computer CO). > Ceo — Figure 13.1 A computerconroled motor spend control em to continuous), then the system generally works as predicted by continuous theory. Consider for example, a motor speed control system: normally, a dc. motor's speed will not be constant if there are load variations, and feedback from a tachometer can be used to reduce speed fluctuations as shown in Figure 13.1 However, as the sampling speed varies the system's performance changes. For example, at a high sampling rate, we might get a response such as the one shown in Figure 13.2(a). As the sampling rate is reduced, we might get the response shown in Figure 13.2(b). Finally, at a rate slow enough, the system might go unstable and the response would be depicted as shown in Figure 13.2(c). 13.2.1 Poles and zeros of a transfer function The poles of a transform function are the values of the Laplace transform variable, s, that cause the transfer function to be infinite; they are the roots of the denominator of the function that are common to the roots of the numerator. The zeros of a transform function are the values of the Laplace transform variable, s, that cause the transfer function to be zero; they are the roots of the numerator of the function that are common to the roots of the denominator. ‘The poles and zeros have certain characteristics: = A pole of the input function generates the form of the forced response. WA pole of the transfer function determines the form of the natural response, = A pole on the real axis generates an exponential response of the form e~*, where a is the pole location on the real axis. = The poles and zeros generate the amplitudes for both the forced and natural responses. Control theory: analysis _451 Speed Speed “Actual PRA AeA. Desi Actual ~ Desired Time @ Time Speed tual Desired Time © Figure 13.2 Variance bowween actual and devred response with diferent sampling rates (a) Nii (b) reduced; (2) ‘exea-zeduced showing instability 13.3 Dynamic characteristics of a control system Most dynamic systems can be modeled as linear ordinary differential equations with constant coefficients, The general way of expressing the dynamics of linear systems is: a3.) where Xoue is the output variable, Xjq is the input variable, 4, and Byy are constant coefficients of the system's behavior; and N and M independently define the order of the system. It should be mentioned that many mechatronic systems exhibit non-linear behavior and therefore, cannot be accurately modeled as linear systems, 452 Mechatronics Fortunately, a non-linear system may often exhibit linear behavior over a specified range of inputs and hence can be adequately approximated using a linear model over this range; a process known as linearization in dynamic system response. By varying the values of V and M, itis possible to examine in some detail different orders of dynamic systems. 13.4 Zero-order systems A zero-order dynamic system is obtained by setting N=_M=0 in Equation 13.1 This leads to the following equation: AoXou = BoXine (13.2) leading to B Nout =" Nig = Hin, 133) xz (13.3) where as we have already discussed in amplifiers, H is the gain or sensitivity of the system, A potentiometer used to measure displacement is 2 typical example of a zero-order dynamic system, where the output is related to the input as follows: Ry = (4), = You = Bes (x0 HX, (3.4) where Rp is the maximum resistance of the potentiometer, Ry is the current resistance of the wiper position, and L is the maximum amount of wiper travel 13.5 First-order systems A first-order dynamic system is obtained by setting N=1 and M=0 in Equation 13.1. This leads to the following equation: Ay S88 Ay Xa = Bo Xing (13.5) ad leading to ArdXou By a Xou =F Ay dt a again, as we have already discussed in amplifiers, H is the gain or sensitivity of the system, in = HXiny (13.6) Control theory: analysis 453 ‘The ratio on the right-hand side of Equation 13.6 is known as the time constant: A . (37) Consequently the first-order equation becomes Aon py rapt Xow = AXin (13.8) ‘There are several input models such as step, impulse, and sinusoidal functions. Let us apply the step function, although any of the other types could bbe considered. The mathematical expression for a step function is 0 1<0 "lay 20° we with the initial condition being Xoxl0) =O. (13.10) ‘The standard approach to solving linear differential equations is to assume a solution of the form ous = Cel aay It is an easy matter to show that the characteristic equation (considering only the left-hand side of Equation 13.8) is given as +10. (3.12) 2): Noun, = CO" (3.13) ‘We also now have the particular or steady state solution (considering only the left-hand side of Equation 13.8) to be: Kou And the homogenous or transient solution is (since 2= Haig 3.14) ‘The general solution (considering both Equation 13.13 and Equation 13.14) becomes: out) = Kou, + Xoaty = COM + Any (3.15) which when we apply the initial condition of Equation 13.15 becomes Xous(0) = C+ HA (3.16) C- HAs, agin), 454 Mechatronics leading to Xou(0) = HAyg(l — e*, 03.18) A monostable vibrator (one-shot) containing @ capacitor, used to generate 4 one-shot pulse is a typical example of a first-order dynamic system. Indeed, most RC circuits are first-order dynamic systems. Let us approach the first-order systems from the point of view of transfer functions. For the first order system, G(s)=aj(s-+a), and considering the input of a unit step, where (3) = I/a, the output becomes (9) = ROG) 13.19) 1) = ROG) = (13.19) ‘The inverse Laplace transform for the step response gives c(t) = at) + en) = 1 =e (13.20) Figure 13.3 shows the plot of Equation 1 the origin generated the forced response, (0) the natural response ¢4(t)= =<" When = La, 20. Notice that the input pole at = 1, and the system pole at —a, 037 (03.21) eta] 0.37 = 0.63 (03.22) Figure 133 Fintorder gsiem response (0 @ unit sep input Control theory: analysis _455 Equations 13.20, 13.21, and 13.22 form the basis for the definition of the three transient response performance specifications: time constant, rise time, and settling time. = The time constant, I/a, is the time for e~“ to decay to 37% of its initial value or the time it takes the step response {0 rise to 63% of its final value. = The rise time, T,, is the time for the waveform to go from 0.1 to 0.9 of its final value, This is the same as letting ()=0.9 and 0.1, respectively, and solving for time, t, which gives 231 01 T= (13.23) = The settling time, 7;, is the time for the response to reach, and stay within, 2% of its final value. This is the same as letting e(1)=0.98 and solving for time, 1, which gives n=4 (13.24) 13.6 Second-order systems este ‘or A second-order dynamic system is obtained by setting N=2 and M=0 in ‘eshienal material Equation 13.1. There are several responses that we need to consider: overdamped, eee. te 9 underdamped, undamped, and eritially-damped. 13.6.1 An overdamped response Let us consider for this system that the input is a step function R(s)=/s and that, 5 Is 60) = SETH (13.240) Concentrating on the denominator and finding the roots, we have Tet) 2 =A TAS) 2 (13.248) = -754£642=-1.077 or ~ 13.922 456_ Mechatronics leading to w= Is 1s 5 CO=SE FISHIN AGT LOTS FISID) 1525) The pole at the origin comes from the unit step input and two real poles come from the system. The input pole results in a constant forced response, while cach of the (wo system poles on the real axis generate an exponential natural response, the frequency of which is equal to the pole location, Hence, cht) = Ky + Kye Kye (13.26) 13.6.2 An underdamped response Let us consider for this system that the input is a step function RG)= l/s and that, 15 Fray Concentrating on the denominator and finding the roots, we have Gs) (13.264) = ($4 VSE=TRTB)2 = FAO 5.20685 (15.268) leading to 15 15 CO FETS IS) — NE TIS = LICH FES + DIR (32 (0 = Ki +o (Ky cos 2.958 + Ky sin 2.9581) (13.28) 13.6.3 An undamped response Let us consider for this system that the input is a step function R(s)=1/s and that, 15, wy Concentrating on the denominator and finding the roots, we have v1, (13.288) Gs) = (13.284) Control theory: analysis _457 leading to 15, x 60 = 5 (13.280) (= 1- cos VTS (1329) 13.6.4 A critically-damped response Let us consider for this system that the input is a step function R(9)=1/s and that =a 1329) 9 = Fis +25) (3.298 Concentrating on the denominator and finding the roots, we have s = (vice), leading to cs) = (13.298) e+ e(t) = 1 = Se (13,30) Figure 13.4 shows undamped, underdamped, and overdamped second-order systems. 13.7 General second-order transfer function ‘The general second order transfer function takes the form (3.31) ‘The roots of the denominator are saa (tan VRP STR) /2=—to,toy/F=T (1532 -veral damping cases arise from Equation 13.32 and these are summarized in Table 13.1 458 Mechatronics = 1 >| Poses Step Gain overdamped |] 5 Fissne underdamped Scope ze @ o Figure 13.4 The undamped, undewlampad, and overdamped responses fora step Input. EXAMPLE 13.1 EXAMPLE 13.2 Control theory: analysis _459 Table 13.1 Damping cases foe second-order sjstem case Expression Description ro to, Undamped Underdamped dest tio TE ™ CCritcaly-damped Overdamped For the transfer function given, determine the values of ¢ and a: 64 FFer8 (13.324) Gs) = Solution Comparing the given transfer function and Equation 13.31, we find that wo? = 64 _ (13.328) = VG Also, 2» = 6, 80 6 pyr 0ts Categorize the following transfer functions as underdamped, critically damped, or overdamped. 16 FF 10s+16 (13.320) © G6) O60 =a HG Solution (@) Comparing the given transfer function and Equation 13.31, we find thats a? =25, $0 w_ = 5. Also, 2f0, = 10, so ¢=1, implying critical damping, 460 Mechatronics (b) Comparing the given transfer function and Equation 13.31, we find that «,? = 36, s0 wp = 6. Also, 2¢, = 10, so ¢ = 0.833, implying under- damping, (© Comparing the given transfer function and Equation 13.31, we find that o,? = 16, 50.) =4. Also, 2g = 10, so ¢= 1.25, implying over- damping, 13.7.1 Underdamped second-order systems Underdamping is often encountered in physical problems, so let us consider the step response, R(s)= Ijs, for the general second-order system. Cs) = ROG) = 13.33) (= ROG) Mons tor) (33) leading to a) (13.34) For the underdamped case, where the poles are at tiovT=F, (1335) using partial fraction method, we have (+ bon) + Goo <4 ——_ 7 (13.354) 5 (P+ kon) ten%(1 — 2°) Taking the Laplace transform gives the following result, oe ~e( os con T= Ft gigi /T=F) (13.36) = 1 phget'( cosmovI=Fr— 8), where (13.364) 137.41 13742 Control theory: analysis _461 We now define four other parameters associated with the underdamped response for the second-order system, These parameters are rise time, peak time, percentage overshoot, and settling time. Rise time The rise time, T;, is the time for the waveform to go from 0.1 10 0.9 of its final value. This is the same as letting e(1)=0.9 and 0.1, respectively, and solving for apf. Subtracting the 1Wo values of of gives us the normalized rise time, wT, for a given damping ratio, ¢. Solving Equation 13,36 for these conditions is ‘difficult and hence the use of a computer is advisable, Table 13.2 shows the normalized rise time for values of damping ratios. We shall see later how to use these values to determine rise time. Peak time ‘The peak time, Tp, is the time required to reach the first, or maximum, peak, Itis obtained by differentiating Equation 13.36 with respect to time and equating the differential to zero, then solving for the time. a= HH sin cig T= Ft = 0 (13.37) (13.38) Table 13.2 Normalized rise time f T 01 ro 02 203 03 ma a4 4653 os a8 06 est 0 2108 08 2467 09 2883 462 Mechatronics yielding (for (13.39) 13.7.1.3 Percentage overshoot The percentage overshoot, %0S, is the amount that the waveform overshoots the steady-state, or final, value at the peak time, expressed as a percentage of the steady-state value. 0 = Sst — Sos 5 199, 03.40) Cha ‘The maximum value, dpa (when 1=7;) is substituted in Equation 13.36, Since the maximum value of cosine is 1, then te le/ Y=) 3.41) Cmax = (Tp We take cfu = 1 for a unit step, and then obtain the expression for pereentage overshoot as mos = (Y=) 5 100 (03.42) Since the percentage overshoot is a function of the damping factor, it is possible to express the damping factor in terms of the percentage overshoot from Equation 13.42. Rearranging, we got eles] VR) _ 408 0 (13.43) ‘Taking the log of both sides of this equation yields tn /V1=@ = In), 03.44) Squaring both sides of the equation gives fi (13.45) (13.46) Notice that we have included a negative sign to cater for 2, which is normally less than 1. Control theory: analysis 463 13.7.1.4 Settling time ‘The settling time, 7,, is the time for the response to reach, and stay within, 2% of its final value. This is the same as letting c(t) =0.98 and solving for time, ¢ In other words, (1349) (13.48) For the range of 0.1¢<09, we can approximate the numerator of Equation 13.48 as 4 since it varies between 3.92 and 4.74, hence a4 1349) Fay, ‘ Z EXAMPLE For the transfer function 133 4 150 G9 = (9.494) determine: (a) the peak time; (b) the percentage overshoot; (c) the settling time; and (d) the rise time. Solution oy = VT3O = 12.2474 +o 20 (13.498) Foe TaD TATE = 08165 (underdamped), (a) The peak time is a 13.49¢ onJl= 122474VT— 0816S (BPO) (b) The percentage overshoot is mos =e (lv 7) x 100 = exp( SAS %100.= 1.176, (13.490) VT 08165 464 Mechatronics Fasingt Figure 13.5 A woond: (©) The settling time 4 4 Yon O8T6S x 12.2474 (@)_ The rise time is obtained using ¢=0.8165 in Table 13.2, for which the normalized rise time is T-=2.5s (by interpolation) and the actual rise time 4s (13.49) r i 25 ay T2284 204s. (13.49) Consequently, for the underdamped second-order system, we can obtain peak time, percentage overshoot, settling time, and rise time without plots. AS an illustration of the second-order dynamic system, let us consider Figure 13.5, which is a mechanical mass-spring-damper system, This is a forced damped vibration system since the applied force is not natural but it has an amplitude and follows a sinusoidal waveform, The governing differential equation is: mi = ~kx — fic+ Fr sing. (13.50) We define s2eo=F ad m=z (13.51) We now rewrite Equation 13.50 as + 2408+ ox = FA sn gt (13.52) Control theory: analysis 465 leading to E+ ax-+ ote =o? yq singe (13.53) Let us define poofimt (13.54) Then for this class of problem, we choose a complementary function (transient solution) of the form xg = Csingt + Evos qt (13.55) Hence applying Equation 13.55 to Equation 13.53 we have —P Csingt — @ Ecos gt + 2gay Ceosgt 03.56 — 2aglisin gt + a*Csin gt + a? Ecos gt = oq sin gt We can now equate terms for sine and cosine accordingly: sin: PC 2ogh + C= 07ym «3s cos : PE + wg + a (13.58) Hence, = PC ~ Boge = ym 3.59) ogC +(e? —P)E=0. (13.60) From Equation 13.6, we have =2tage 161 ea wey Substituting in Equation 13.59, we now have (0? C4 2g TE = oy (13.62) Consequently, from Equation 13.62 the expressions for C and E are obtained as (13.63) 466 Mechatronics | (3.6 ‘The general solution PI, obtained by substituting Equations 13.63 and 13.64 into Equation 13.55 becomes (-24.9)0?yq cost ; lie a+ 74] ee @)sin gt — 2¢004 60s qi, (13.66) J sinat—v)} (13.67) Finally, we have yn sin(gt — PI (13.68) (13.69) and then the phasor: =H We now plot the frequency response of the second-order system as a function of the damping ratio ¢ as shown in Figure 13.6, When the input frequency, 4g, is equal to the natural frequency « (jc. g/@=1) resonance occurs. We note from Equation 13.69 that when ¢=0 (no damping), the amplitude ratio is Control theory: analysis _467 Increasing amplitude ratio Figure 13.6 Second-order sysem amplitude response 13.7.1.5. The maximum amplitude ratio The maximum amplitude ratio is obtained by taking the squares of both sides of Equation 13,52, differentiating the denominator with respect to w and equating to zero: é afl-€ wom leading to 4c = oyT=2E_ at resonance (13.2) But paoyT=F, hence o> p> a asa) We can now obtain the maximum value of the amplitude ratio from Equations 13.72 and 13.69 as 1 1 2WO-P) Hesax = (13.73) ti == 2¢5)7+ 4220 — 2¢2)} 468 Mechatronics This maximum value together with the frequency ratio, g/o=1, give 1/2 and help us to draw the graph of the amplitude ratio, 4, against the frequency ratio, gla. Although we have successfully determined the frequency response, the method is unwieldy. Not only that, it is difficult to know what form of solution to assume, We now present the Operator D-Method, which is very user friendly, 13.7.2 Operator-D method In this method, we use the following notations f=D'y, $= Dx; singr= te (13.73) Let us restart from Equation 13.53, which is reproduced below: £ + Bot + utr =r yp sing. (13.738) Applying the Operator-D notations, we have (D? + 2gwD + 07x =o Yule (13.74) We can immediately write the general solution PI as: Pl=x Teh (13.75) “DF yeds (13.76) ) — 2togie* am + 4F are] P) — 2gagi)(cos gt + Tart 13.78) — ay aoe] ‘ Vial (w — g*) sin gt — 2¢a9 cos gi] (13.79) [oe -@F + aera] Control theory: analysis _469 and ya sin(gt = V) Yo sin(gt = ¥) [or =F +4e0@?] Plox= (13.80) rove |S] easy which is exactly the same solution as with the previous method. As demon- strated using this example, the Operator-D method is quite powerful and takes ‘a much shorter time to analysis the response equation, EXAMPLE A mass of 25g is suspended from a spring with stiffness $00 000 N m=! 13.4 ‘The damping is negligible, The mass is initially resting in its equilibrium position when a fluctuating force of amplitude 10000N and frequency 100 rads” is suddenly applied to it. Determine: (a) the amplitude of the steady state vibration; and (b) the arbitrary constants. Sketch the displacement-time curve for the first few cycles of the oscillation. Solution ‘The system is represented in Figure 13.7. ‘The governing differential equation is given as: mi = kx + Fy sin gt . (381A) (&+0")x = 0'yp sings V_ | F1t)= 10000 singt Figure 13.7 Sytem for Frample 1 470 Mechatronics Applying the Operator-D notation, we have (D’+.0%)x=ayalne™ (1381B) Equating the left-hand side to zero and solving for D, we obtain the complementary function: (D'+0) =0 D=+io. (13.810) Hence xy = A sin wt + Boos. We can immediately write the general solution PI as: Pax = rah x= Ip Bolt PEA _ tart ralo? — @o [=] Jae? Yalu? — @ (cos gt + j singe) (13.81D) ya sin gt Hence, x= Asinot + Beoswr +28 S0a CF ]e = 1412 rad Control theory: analysis 471 This value is at steady-state condition, Using the equation for the final equation for displacement, at time 10, Xo=0, hence B=0. $= 0d cos of —oB sin ot + PE OF 7]0) Oat ¢ =O since the system is resting at equilibrium position initially on O04 TK Ba ala -vi/i ; A= Gaga = Tae = Ve (1381E) Resulting in swsingt (sgn ot0or _-VBsin or 2as24t_,(asington—VEsinst2x)—(a81F where yn =0.02 (as already calculated) This equation describes the response of the dynamic system being considered in this example; the equation is used to draw the graph of x asa function of sin 100r and sin 141.2r with the appropriate amplitudes, 13.8 Systems modeling and interdisciplinary analogies An ordinary linear differential equation is known to model any linear system, relating the output response of the system to the input, whether electrical, ‘mechanical, hydraulic, or thermal. The analogies between these several engineer- ing disciplines have been developed over time. The mechanical analogy used is the mobility analogy in which the physical analogies are sacrificed in favor of creating equivalent mathematical relationships, which hold in network analysis Table 13.3 summarizes the interdisciplinary analogies that exist between different systems. Figure 13.8 shows a simple analog mechanical system and a simple electrical system. Their respective governing differential equations are given in Equations 13.82 and 13.83, The analogy between these systems is of practical importance because cach of them can be independently analyzed and substituted for each system, for example, for a mechanical system, it is now possible to use the mesh circuit theory to write the governing differential equations. wep =F Gedep sovedgsep °F2) (a) ¥iTI=Y_—_(qwauodu0o asndsip) Pg} SNeOSIA—AOUESON, ow ywuonsedorg uid= oy “ous PML vird=w 0 dtd 30) 83 9] puoass .u0189N wid) =o 0 mipr=z Baw wouodwos fej) ‘souruoey ‘epiou jo suauow jog epsauy 9ysseyy——souRFIMPL wow 8 ase so La) doup suns. [wz ‘onbiog, Isla eu0g—[A]4‘simoa, Biqeurs jeusiog A-2umoy ——waupsedap sauy wouroexdr (Patra suouaoeidsic, [.-sp2) “p09 sepstoy aun jroupevoay suet jeune e re3Ueypari seriojeus Samuydospiway EEL o1geL Di wo ln= naa “ona “apes wot ‘doo=99 ‘sounds pay ora use nk = 4 suosts [eweysoat vone03 jo am puoaag, f= 8 sey surues aupds ruopio1= py 2 ed wae wee 04 st sons 10 Sytuav0y ons sxdound sunquary.P waS=x suowsoeysie (soups) wesu09 duude—onps fy m0] 3400 fi a Woibi= a ‘ious renwore Ihur="9 s8iou9 onoure iwlara 0d) somo 0= (enu08) 1961 doo 8019 oy asta 0= Wo OHH ssn] apoujuonsuny bmi = sam Sayan 1910 "apumovde 474 Mechatronics 4 le an. aa Figure 13.8 Analog mechanieal and cecrcal gsm (m3) + kx) + foxy) am = 0 (13.82) (msiz + axa) — ham = 0 (Lah + 1/Cin + Cami) — 1/C2g2 = 0 (13.83) (Lain + Cag) — 1/Caqi = 0. 13.9 Stability Stability is the most important system specification. An unstable system cannot be designed for a specific transient response or steady-state error requirement. There are many definitions for stability, depending upon the type of system or the point of view. In this section we limit ourselves to linear, time-invariant systems, We have already discussed that we can control the output of system if the steady-state response consists of only the forced response. But the total response of a system is the sum of the forced and natural responses, or (0) = crores) + Camu. (13.84) 13.9.1 13.9.2 13.9.3 Control theory: analysis _ 475 ‘We now present the following definitions of stability, instability, and marginal stability Stable systems Let us focus on the natural response definitions of stability. Recall from our study of system poles that poles in the left half-plane (Ihp) yield either pure exponential decay or damped sinusoidal natural responses. These natural responses decay to zero as time approaches infinity, Thus, if the closed-loop system poles are in the left half of the s-plane and hence have a negative real part, the system is stable. That is, stable systems have closed-loop transfer functions with poles only in the left half-plane. Here are some definitions of stable systems: = A linear, time-invariant system is stable if the natural response approaches zero as time approaches infinity = A system is stable if every bounded input yields a bounded output Unstable systems Poles in the right half-plane (rhp) yield either pure exponentially increasing or exponentially increasing sinusoidal natural responses. These natural responses approach infinity as time approaches infinity. Thus, if the closed-loop system poles are in the right half of the s-plane and hence have a positive real part, the system is unstable. Also, poles of multiplicity greater than one on the imagi- nary axis lead to the sum of responses of the form AP" cos(w! +9), where n=1 2,..., which also approaches infinity as time approaches infinity. Thus, unstable systems have closed-loop transfer funetions with at least one pole in the right half-plane andjor poles of multiplicity greater than one on the imaginary axis, Here are some definitions of unstable systems = A linear, time-invariant system is unstable if the natural response grows without bound as time approaches infinity. =A system is unstable if any bounded input yields an unbounded output, Marginally stable systems A system that has imaginary axis poles of multiplicity 1 yields pure sinusoidal oscillations as a natural response. These responses neither inerease nor decrease in amplitude. Thus, marginally stable systems have closed-loop transfer functions Mechatronics with only imaginary axis poles of multiplicity | and poles in the left half-plane. Here is a definition of « marginally stable system: ar, time-invariant system is marginally stable if the natural response neither decays nor grows but remains constant or oscillates as time approaches infinity 13.10 The Routh-Hurwitz stability criterion The Routh-Hurwitz stability criterion for stability is a method that yields stabi- lity information without the need to solve for the closed-loop system poles. Using this method, we can determine how many closed-loop system poles are in the left half-plane, in the right half-plane and on the jo-axis. An impor- tant observation is that we say how many, not where, We can determine the number of poles in each section of the s-plane, but we cannot find their exact coordinates, The method requires two steps: 1. Generate a data table called a Routh-rable. 2. Interpret the Routh table to tell how many closed-loop system poles are in the left half-plane, the right half-plane, and on the jo-axis. 13.10.1 Generating a Routh table Referring to the equivalent closed-loop transfer function shown in Figure 13.9, ‘we focus our attention on the denominator since we are interested in the system poles, First create the Routh table shown in Table 13.4, We commence by label- ing the rows with powers of s from the highest power of the denominator of the closed-loop transfer function to s°. Next start with the coefficient of the highest power of s in the denominator and list, horizontally in the first row, every other coefficient. In the second row, list horizontally, starting the next highest power of s, every coefficient that was skipped in the first row. The remaining entries are filled in as follows. Each entry is a negative determinant of entries in the previous Ms) os) ays + agst + ays? + aps? + ays + ay Figure 13.9 A closed loop transfer fetion, EXAMPLE 135 Control theory: analysis _477 Table 13.4 Initial template for Routh table Table 135 Completed Rovth table ‘two rows divided by the entry in the first column directly above the calculated row. The left-hand column of the determinant is always the first column of the previous swo rows, and the right-hand column is the elements of the column above and to the right. The table is complete when all of the rows are completed down to s°, Table 13.5 is the completed Routh table for Figure 13.9. Generate the Routh table for the system shown in Figure 13.10, Solution ‘The first step is to find the equivalent closed-loop system because we want to test the denominator of this function, not the given forward transfer function, for pole location. Using the Feedback formula, we obtain the equivalent system as Figure 13.10(b), We will apply the Routh-Hurwitz criterion to the denominator, (649s? + 26s-+81). First label the rows with powers of s from s* down to sin 478 Mechatronics 13.10.2 Re) + Els ois) 1) ) 7 (s+ 2)s+3(5+4) @ Als) —>| 57 Ly» ow ry Bae + 265481 Figure 13.10 Closed Joop transfer function for Example 133, Table 13.6 Completed Routh table for Example 13.8 ? v 2 or Ls a vertical column, as shown in Table 13.6. Next form the first row of the table, using the coefficients of the denominator of the closed-loop transfer function, We commence with the coefficient of the highest power and skip every other power of s. We now form the second row with the coefficients of the denominator skipped in the previous step. Subsequent rows are formed with determinants as shown, For convenience any row of the Routh table can be multiplied by a positive constant without changing the values of the rows below. This can be proved by examining the expressions for the entries and verifying that any multiplicative constant from a previous row cancels out. In the second row of Table 13.6, for example, the row was multiplied by 1/9. We see later that care must be taken not to multiply the row by a negative constant Interpreting a Routh table The basic Routh table applies to systems with poles in the left and right half- planes, Systems with imaginary poles (and the kind of Routh table that results) Control theory: analysis _479 will be discussed in the next section. Simply stated, the Routh-Hurwitz criterion declares that the number of roots of the polynomial that are in the right half-plane is equal to the number of sign changes in the first column. If the closed-loop transfer function has all poles in the left half of the s-plane, the system is stable. Thus, a system is stable if there are no sign changes in the first column of the Routh table, For example, Table 13.6 has two sign changes in the first column, Thus, the system of Figure 13.10 is stable since no poles exist in the right half-plane Now that we have described how to generate and interpret a Routh table, let us ook at two special cases that can arise. 13.10.2.1 Zero only in the first column If the first element of a row is zero, division by zero would be required to form the next row. Two methods are normally used: (a) the epsilon method; and (b) the reciprocal-roots method. In the first method, to avoid this zero-row phenomenon, an epsilon, e, is assigned to replace the zero in the first column. The value of ¢ is then allowed to approach zero from cither the positive or the negative side, after which the signs of the entries in the first column can be determined. In the second method, we show that the polynomial we are looking for, the ‘one with the reciprocal roots, as simply the original polynomial with its coefficients written in reverse order, For example, Sb gas + 4 ays-+dg = 0 (13.85) If s is replaced by 1/d, then 4 will have roots which are the reciprocal of s. Making this substitution gives, (13.86) Factoring out (6) [eG (Q) [lt ajaid-t.. tad"! + ayd"] =0. 13.87 (13.864) 480_ Mechatronics Table 13.7 Pastial Routh table for Fxample 13.6 * a aos @ Thus, the polynomial with reciprocal roots is a polynomial with the coefficient written in reverse order. EXAMPLE Determine the stability of the closed-loop transfer function 13.6 5 See webste for dow Ts) = (13.88) Fie Tiss lO OFS ‘oacabie VATLAB code te sae this prebiem Solution Filling the first two rows of the Routh table gives Table 13.7, Here the pivot for the third row contains a zero. So we form a polynomial with its coefficients written in the reverse order, First write a polynomial that is the reciprocal roots of the denominator of Equation (13.88). This polynomial is formed by writing the denominator in reverse order. Hence, PO = TEER EO SPELT (13.89) We form the Routh table as shown in Table 13.8 using Equation 13.89. Since there are (wo sign changes, the system is unstable and has two right-half plane poles. Notice that Table 13.8 does not have a zero in the first column, 13,10,2.2. Entire row is zero We now look at the second special case. Sometimes while making a Routh table, we find that an entire row consists of zeros because there is an even polynomial that is a factor of the original polynomial. This case must be handled differently from the case of a zero in only the first column of a row. Let us look at an example that demonstrates how to construct and interpret the Routh table when an entire row of zeros is present. EXAMPLE Determine the number of right-half plane poles in the closed-loop transfer 137 function 30. 1) = (13.90) Fos 30 Control theory: analysis _481 Table 13.8 Completed Rovth table for Example 13.6 ° @2 2 * 6 1 6s 333066 ce ’ 333 06 sat a s 3811 ozs of ‘0218 Table 13.9 Partisl Routh table for Example 13.7 = 1 5 6 s 6 5s 60)6 ° ° 1s 16 log 1 élag Solution Start by forming the Routh table for the denominator of Equation 13.90 (see Table 13.9). In the second-row we multiply through by 1/5 for convenience. We stop at the third row, since the entire row consists of zeros, and use the following procedure. First we return to the row immediately above the row of zeros and form an auxiliary polynomial, using the entries in that row as coefficients, The polynomial will start with the power of s in the label column and continue by skipping every other power of s. Thus, the polynomial formed for this example is +584 Ps) (391) 482_ Mechatronics 13.10.3 EXAMPLE 138 See webste for dow Table 13.10 Completed Routh table for Fxample 13.7 r 1s - eI =25 * 25 6 ae 78 [Next we differentiate the polynomial with respect to s and obtain Ps) =4s' + 10540. (13.92) Finally, we use the coefficients of Equation 13.92 to replace the row of zeros, Again, for convenience the third row is multiplied by 1/4 afler replacing the The remainder of the table is formed in a straightforward manner by follow- ing the standard form shown in Table 13.10 which shows that all entries in the first column are positive, Hence, there are no right-half plane poles Stability design using the Routh-Hurwitz criterion Find the gain range, K (assume K>0) for the system shown in Figure 13.11 to cause the system to be: (a) stable; (b) unstable; and (©) marginally stable. acacia VATLAB code 1a sae this prebiem Solution The first step is to find the equivalent closed-loop system because we want (o apply the Routh-Hurwitz criterion to the denominator (s° +158 +50s-+K) of this function, not the given forward transfer function, for pole location, First label the rows with powers of s from s° down to sin a vertical column, as Control theory: analysis 483 Fis) +A ls) K a) S(5+5){5+ 10) Figure 13.41 A gem to be amined lor sabi fn Example 1h Table 13.11 Completed Routh tabi for example 13.8 ° fi 30 : 1 50] ° isk] 30: shown in Table 13.11, Next form the first row of the table, using the coefficients of the denominator of the closed-loop transfer function, T= ss + 15s? + 3084 K 13.93) ‘We consider the three possible cases. = K> 750: All clements in first column are positive except the row s! ‘There are two sign changes showing that the system is unstable, = K<0: All elements in first column are positive and there is no sign change showing that the system is stable, with three poles in the lefl- halfplane. = K=750: We have an entire row of zeros, which signify jo poles. Let us explore the last case further. First we return to the row immedi= ately above the row of zeros and form an auxiliary polynomial, using the entries in that row as coefficients, The polynomial will start with the power of s in the label column and continue by skipping every other power of s. Thus, the polynomial formed for this example is Ps) = 158 +750, (1394) Next we different te the polynomial with respect to s and obtain Pls) = 305-+0 (13.95) 484 Mechatronics 13.11 13.111 atta Table 13.12 Completed Routh table for Fxample 138 with K=750 1 0 sy (750) 50 0 ° ee sa ° Finally, we use the coefficients of Equation 13.95 to replace the row of zeros. For convenience we multiply the second row entries by 1/15. The remainder of the table (Table 13.12) is formed in a straightforward manner by following the standard form. Since there is no sign change from the even polynomial, s°, down to the bottom of the table, the even polynomial has its two roots on the ja-axis of unit multiplicity. The system is therefore marginally stable. Steady-state errors The steady-state error is the difference between the input and the output for a prescribed test input as ¢> 0c. Since we are concerned with steady-state ertor afier the steady state has been reached our discussions are limited to stable systems, where the natural response approaches zero as 1» 0. The control engineer must first check that the system is stable while performing steady-state error analysis and design, Steady-state error for unity feedback systems The steady-state error can be calculated from either a system's close-loop transfer function, 7(s), or the open-loop transfer function, G(s), for a unity feedback system. Let us deal with these two cases, ‘Steady-state error in terms of T(s) Consider the feedback control system shown in Figure 13.12(a) K(s) = R(9) - CO). (13.96) Control theory: analysis 485 Aisle FS) as) Gs) ©) Figure 13.42 Steady-state error.) in tenns of (9B) in terms oF OG. But CW) = REIT), (13.97) hence Els) = R66) — R()T3) = RSI = TEs). (13.98) From the final value theorem, e(00) = lim e(#) = lim sE(0. (13.984) EXAMPLE For a unit step input, and closed-loop transfer function, 139 1) =>-4 (13.988) ore determine the steady-state error of the system, Solution os 4th 6s44 RO) =H = Sere HOES 1H)

    You might also like