Control Design Using Bode Plots

Topic #4 16.30/31 Feedback Control Systems Control Design using Bode Plots • Performance Issues • Synthesis • Lead/Lag examples Fall 2010 16.30/31 4–2 Bode’s Gain Phase Relationship • Control synthesis by classical means would be very hard if we had to consider both the magnitude and phase plots of the loop, but that is not the case. • Theorem: For any stable, minimum phase system with transfer func­ tion G(s), �G(jω) is uniquely related to the slope of |G(jω)|. • Relationship is that, on a log-log plot, if slope of the magnitude plot is constant over a decade in frequency, with slope n, then �G(jω) ≈ 90◦n • So in the crossover region, where L(jω) ≈ 1 if the magnitude plot is (locally): s0 slope of 0, so no crossover possible s−1 slope of -1, so about 90◦ PM s−2 slope of -2, so PM very small • Basic rule of classical control design: Select Gc(s) so that LTF crosses over with a slope of -1. September 19, 2010 Fall 2010 16.30/31 4–3 Performance Issues • Step error response ess = 1 1 + Gc(0)Gp(0) and we can determine Gc(0)Gp(0) from the low frequency Bode plot for a type 0 system. • For a type 1 system, the DC gain is infinite, but define Kv = lim sGc(s)Gp(s) ⇒ ess = 1/Kv s→0 • So can easily determine this from the low frequency slope of the Bode plot. September 19, 2010 Fall 2010 16.30/31 4–4 Performance Issues II • Classic question: how much phase margin do we need? Time response of a second order system gives: 1. Closed-loop pole damping ratio ζ ≈ P M/100, P M < 70◦ 2. Closed-loop resonant peak Mr = ωr = ωn √1 2ζ � 1−ζ 2 ≈ 1 2 sin(P M/2) , near 1 − 2ζ 2 3. Closed-loop bandwidth � � 2 ωBW = ωn 1 − 2ζ + 2 − 4ζ 2 + 4ζ 4 �� and ωc = ωn 1 + 4ζ 4 − 2ζ 2 20 10 Decibels Amplitude ratio 1 |y|/|r| 0.7 0.1 ωc Mr resonant peak 0 -3 ω, rad/s -20 Band width ωBW Image by MIT OpenCourseWare. Fig. 1: Frequency domain performance specifications. • So typically specify ωc, PM, and error constant as design goals September 19, 2010 Fall 2010 16.30/31 4–5 Fig. 2: Crossover frequency for second order system 1 2 3 4 5 6 7 8 9 10 figure(1)% z=[0:.01:1]'; zs=[0:.01:1/sqrt(2)−eps]'; wbw=sqrt(1−2*z.ˆ2+sqrt(2−4*z.ˆ2+4*z.ˆ4));% wc=sqrt(sqrt(1+4*z.ˆ4)−2*z.ˆ2);% wr=sqrt(1−2*zs.ˆ2);% set(gcf,'DefaultLineLineWidth',2) % plot(z,wbw,z,wc,zs,wr)% legend('\omega {BW}/\omega n','\omega c/\omega n','\omega r/\omega n') % xlabel('\zeta');ylabel('\omega')% print −dpng −r300 fresp.png • Other rules of thumb come from approximating the system as having a 2nd order dominant response: 10-90% rise time Settling time (5%) Time to peak amplitude Peak overshoot September 19, 2010 1 + 1.1ζ + 1.4ζ 2 tr = ωn ts = 3 ζωn tp = �π ωn 1 − ζ 2 M = e−ζωntp p Fall 2010 16.30/31 4–6 Frequency Domain Design • Looked at the building block Gc(s) = Kc s+z s+p • Question: how choose Gc(s) to modify the LTF L(s) = Gc(s)Gp(s) to get the desired bandwidth, phase margin, and error constants? • Lead Controller (|z| < |p|) • Zero at a lower frequency than the pole • Gain increases with frequency (slope +1) • Phase positive (i.e. this adds phase lead) dB 0 20 log |p/z| 20 0 0 ω Phase -90o θ 0o z ω p ω Image by MIT OpenCourseWare. Fig. 3: Lead: frequency domain. September 19, 2010 Fall 2010 16.30/31 4–7 Lead Mechanics • Maximum phase added 1−α 1+α where α = |z|/|p|, which implies that sin φmax = α= 1 − sin φmax 1 + sin φmax • Frequency of the maximum phase addition is ωmax � = |z| · |p| • Usually try to place this near ωc • High frequency gain increase is by 1/α • So there is a compromise between wanting to add a large amount of phase (α small) and the tendency to generate large gains at high frequency • So try to keep 1/α ≤ 10, which means |p| ≤ 10|z| and φmax ≤ 60◦ • If more phase lead is needed, use multiple lead controllers Gc(s) = kc (s + z1) (s + z2) (s + p1) (s + p2) • Select of the overall gain is problem specific. • One approach is to force the desired crossover frequency so that |L(jωc)| = 1 • Use Lead to add phase ⇒ increases PM ⇒ improves transient re­ sponse. September 19, 2010 Fall 2010 16.30/31 4–8 Lead Mechanics II • Adding a lead to the LTF changes both the magnitude and phase, so it is difficult to predict the new crossover point (which is where we should be adding the extra phase). dB 0 z ω p Phase 0o ω -90o o -180 Phase margin, compensated Phase margin, uncompensated Image by MIT OpenCourseWare. Fig. 4: Lead example • The process is slightly simpler if we target the lead compensator design at a particular desired ωc 1. Find the φmax required • Note that φrequired = P M − (180◦ + �G(jωc)) 2. Put φmax at the crossover frequency, so that ωc2 = |p| · |z| 3. Select Kc so that crossover is at ωc September 19, 2010 Fall 2010 16.30/31 4–9 Design Example • Design a compensator for the system G(s) = 1 s(s+1) • Want ωc = 10rad/sec and P M ≈ 40◦ • Note that at 10rad/sec, the slope of |G| is -2, corresponding to a plant phase of approximately 180◦ • So need to add a lead compensator ⇒ adds a slope of +1 (locally), changing the LTF slope to -1, and thus increasing the phase margin -1 |.| |G| z 1 0.1 1 p 10 -2 |GGc| New |.| = 1 Line so ωc = 10 Image by MIT OpenCourseWare. Fig. 5: Lead example • Design steps: � � � �G(jω ) ≈ −180◦ � c φm � zp = 1−sin , ⇒ φm = 40◦ ◦ 1+sin φm P M = 40 z ⇒ p = 0.22 � ωc2 = z · p = 102 ⇒ z = 4.7, p = 21.4 � Pick kc so that |GcG(jωc)| = 1 September 19, 2010 Fall 2010 Code: Lead Example 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16.30/31 4–10 Fig. 6: Lead example % % Lead Example 2009 % close all set(0,'DefaultLineLineWidth',2) set(0,'DefaultlineMarkerSize',10) set(0,'DefaultlineMarkerFace','b') set(0, 'DefaultAxesFontSize', 14, 'DefaultAxesFontWeight','demi') set(0, 'DefaultTextFontSize', 14, 'DefaultTextFontWeight','demi') % % g=1/s/(s+1) figure(1);clf; wc=10; PM=40*pi/180; 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 G=tf(1,conv([1 0],[1 1])); phi G=phase(evalfr(G,j*wc))*180/pi; % phi m=PM−(180+phi G); % zdp=(1−sin(phi m))/(1+sin(phi m)); z=sqrt(wcˆ2*zdp); p=z/zdp; % Gc=tf([1 z],[1 p]); k c=1/abs(evalfr(G*Gc,j*wc)); Gc=Gc* k c; w=logspace(−2,2,300); f G=freqresp(G,w);f Gc=freqresp(Gc,w);f L=freqresp(G*Gc,w); f G=squeeze(f G);f Gc=squeeze(f Gc);f L=squeeze(f L); loglog(w,abs(f G),w,abs(f Gc),w,abs(f L)) xlabel('Freq');ylabel('Magnitude');grid on legend('Plant G','Comp G c','LTF L','Location','SouthWest') print −dpng −r300 lead examp2.png September 19, 2010 Fall 2010 16.30/31 4–11 Lag Mechanics • If pole at a lower frequency than zero: • Gain decreases at high frequency – typically scale lag up so that HF gain is 1, and thus the low frequency gain is higher. • Add negative phase (i.e., adds lag) dB 0 0 p ω z -20 0 20 log |p/z| Phase o 0 ωm ω θm -90 o ω Image by MIT OpenCourseWare. Fig. 7: Lag: frequency domain Glag = kc s/z+1 s/p+1 • Typically use a lag to add 20 log α to the low frequency gain with (hopefully) a small impact to the PM (at crossover) • Pick the desired gain reduction at high frequency 20 log(1/α), where α = |z|/|p| • Pick |Glag |s=0 = kc to give the desired low frequency gain increase for the LTF (shift the magnitude plot up) • Heuristic: want to limit frequency of the zero (and pole) so that there is a minimal impact of the phase lag at ωc ⇒ z = ωc/10 September 19, 2010 Fall 2010 16.30/31 4–12 Lag Compensation • Assumption is that we need to modify (increase) the DC gain of the LTF to reduce the tracking error. • Two ways to get desired low frequency gain: � Using just a gain increase, which increases ωc and decrease P M � Add Lag dynamics that increase increase the gain at low frequency and leave the gain near and above ωc unchanged dB |KcGp| |GcGp| 0 p z wc ω 20 log |p/z| Original ωc Phase Gp(jω) -90 o Gc(jω)Gp(jω) -180 ω o Phase margin, uncompensated Phase margin, compensated Image by MIT OpenCourseWare. September 19, 2010 Fall 2010 16.30/31 4–13 Lag Example Lag example with G(s) = 3/((s + 1)2(s/2 + 1)) and on right Gc(s) = 1, left Gc(s) = (s + 0.1)/(s + 0.01), which should reduce steady state step error from 0.25 to 0.032 September 19, 2010 Fall 2010 Code: Lag Example 1 2 3 4 5 6 7 8 9 % % Lag Example 2009 % close all set(0,'DefaultLineLineWidth',2) set(0,'DefaultlineMarkerSize',10) set(0,'DefaultlineMarkerFace','b') set(0, 'DefaultAxesFontSize', 14, 'DefaultAxesFontWeight','demi') set(0, 'DefaultTextFontSize', 14, 'DefaultTextFontWeight','demi') 10 11 12 13 G=tf([3],conv(conv([1/1 1],[1/2 1]),[1 1]));% wc=1;zl=wc/10; pl=zl/10; close all; 14 15 Gc=zl/pl*tf([1/(zl) 1],[1/(pl) 1]);% 16 17 18 19 20 21 22 23 24 figure(1);clf; % set(gcf,'DefaultLineLineWidth',2) L=Gc*G;Gcl=(L)/(1+L);% [y,t]=step(Gcl,20);plot(t,y);grid on; axis([0 20 0 1.3])% hold on;plot([min(t) max(t)],[y(end) y(end)],'r−−');% xlabel('Time (sec)');ylabel('y') % ess=1/(1+evalfr(L,0*j)) print −dpng −r300 lag examp1−1.png % 25 26 27 28 29 30 31 32 33 34 35 w=logspace(−2,2,300)'; figure(2);clf;% set(gcf,'DefaultLineLineWidth',2) [mag,ph]=bode(L,w);grid on; % subplot(211);loglog(w,squeeze(mag));grid on;% axis([1e−2 1e2 .001 100]) % xlabel('Freq');ylabel('Mag')% subplot(212);semilogx(w,squeeze(ph));% xlabel('Freq');ylabel('Phase');grid on% print −dpng −r300 lag examp1−2.png 36 37 38 39 figure(3);rlocus(L);rr=rlocus(L,1);hold on; % plot(rr+j*eps,'rs','MarkerSize',12,'MarkerFace','r');hold off % print −dpng −r300 lag examp1−3.png 40 41 42 Gc=1; Gc2=2; 43 44 45 46 47 48 49 50 51 52 53 54 55 figure(4); set(gcf,'DefaultLineLineWidth',2) % L=Gc*G;Gcl=(L)/(1+L);L2=Gc2*G;Gcl2=(L2)/(1+L2);% [y,t]=step(Gcl,20);% [y2,t]=step(Gcl2,t);% hold on; % plot(t,y,'b−',t,y2,'m−.');grid on; axis([0 20 0 1.2])% plot([min(t) max(t)],[y(end) y(end)],'r−−');% plot([min(t) max(t)],[y2(end) y2(end)],'g−−');% xlabel('Time (sec)');ylabel('y') % ess=1/(1+evalfr(L,0*j)) % hold off print −dpng −r300 lag examp1−4.png % 56 57 58 59 60 61 62 63 64 65 66 w=logspace(−2,2,300)'; figure(5);clf;set(gcf,'DefaultLineLineWidth',2) [mag,ph]=bode(L,w); [mag2,ph2]=bode(L2,w); subplot(211);loglog(w,squeeze(mag),'b−',w,squeeze(mag2),'m−.');grid on;% axis([1e−2 1e2 .001 100]) % xlabel('Freq');ylabel('Mag')% subplot(212);semilogx(w,squeeze(ph),'b−',w,squeeze(ph2),'m−.');% xlabel('Freq');ylabel('Phase');grid on% print −dpng −r300 lag examp1−5.png 67 68 69 70 71 figure(6);rlocus(L);rr=rlocus(L,Gc);rr2=rlocus(L,Gc2);hold on; % plot(rr+j*eps,'rs','MarkerSize',12,'MarkerFace','r'); plot(rr2+j*eps,'md','MarkerSize',12,'MarkerFace','m');hold off % print −dpng −r300 lag examp1−6.png September 19, 2010 16.30/31 4–14 Fall 2010 16.30/31 4–15 Summary • The design summary online is old, but gives an excellent summary of the steps involved. • Typically find that this process is highly iterative because the final performance doesn’t match the specifications (second order assump­ tions) • Practice is the only way to absorb these approaches. September 19, 2010