Mathematical Models of Systems Exercises E2.1 We have for the open-loop As and for the closed-loop e=r—y and y So,e=r—e ande?+e-r=0 FIGURE E2.1 Plot of open-loop versus closed-loop, For example, ifr = 1, then e” +¢— 1 = 0 implies that ¢ = 0.618. Thus, y = 0.382. A plot y versus r is shown in Figure B21E22 E28 Define and AR= f(T)- f(T), AT=T-Ty ‘Then, ar=s(r)— 10) = 2 Are a 0.1 Rye? To = 135, when Ry = 10,0000. Thus, the linear approximation is computed by considering only the first-order terms in the Taylor series expansion, and. is given by AR = -135AT The spring constant for the equilibrium point is found graphically by estimating the slope of a line tangent to the force versus displacement: curve at the point y = 0.5cm, see Figure E2.3. The slope of the line is Kx. Force 1s) os 8 08 ts 8 FIGURE €2.3 ‘Spring force as a function of displacementE24 Since we have __500(s + 100) YO = (ea 10st 50) * ‘The partial fraction expansion of ¥(s) is given by Arg aan teary YO= 5 + sy s+ 0 where 500(100) = 200000) 5 = 125 A= pcg) 100) Ae 112.5 and Ag = 125 Using the Laplace transform table, we find that y(t) = 100 — 112.5e" 1" + 12.575" ‘The final value is computed using the final value theorem Jim y(t) = lim s 100. ae 0 500(s + 100) [aera 1s + 500) E2.5 The circuit diagram is shown in Figure E2.5. FIGURE E2.5 Noninverting op-amp circuit With an ideal op-amp, we have U9 = A(vin — v-),B26 E27 where A is very large. We have the relationship ‘Therefore, Ry U9 = Alvin — Uo), \ Ry t+ Ry » and solving for vg yields Up = A Ui “Tg Since A> 1, it follows that 1+ ;h- © 78). Then the expression for 1. simplifies to _ Rit Ro = An Given J(x) = 2? and the operating point t=1/2, we have the linear approximation Ay=KAc where loa V2 ()—s Gys) 5 Ws IOs 10) Lf 609 a FIGURE €2.7 Block diagram modelStarting at the output we obtain I(s) = Gi(s)Gal(s) (3) But B(s) = R(s) — H(s)I(s), s0 1{s) = Ga(s)Go(s) [R(s) — H(s)1(s)]- Solving for I(s) yields the closed-loop transfer function Ls) Gals)Ga(s) Rs) 1+ Gils)Ga(s) HT £2.8 The block diagram is shown in Figure E28, ngs 5 Au) zi s Gs, 3) L is ts) —o ie fo a® [Guy Gs) L v9) His) FIGURE €2.8 Block diagram model Starting at the output we obtain ¥(s)= 241s) S 4G2(0)A(0) But A(s) = Gy(s) [-Ha(s)2(s) — Hg(s)A(s) + W(s)] and Z(s) = s¥(s), 80 ¥ (8) = ~Gils)Ga(s)Ha(s)¥(s) — Gi(s)Ha(s)¥(s) + 2G,(s)G2()W(). Substituting W(s) = KE(s) ~ Hi(s) (8) into the above equation yields ¥(9) = -Gx(s)Gals) MH (9)"(9) ~ (HY) + LGi(o)Gx(8) KEO)~ FZ]E29 and with E(s) = R(s) — ¥(s) and Z(s) = s¥(s) this reduces to Gi(s)Gols) (Ho(s) + Hh(s)) — Gi(s)Ha(s) = Lav(NGn(o)KIY(6) + ELGAR Solving for Y(s) yields the transfer function Y¥(s) =T(s)R(s), where noe KGi(s)Gals)/s 1+ Gi(s)Go(s) [(Ho(s) + Ai(s)] + Gi(s)Hs(s) + KG)(s)Ga(s)/s Fiom Figure F2.9, we observe that Fy(s) = Ga(s)U(s) and Fx(s ‘Then, solving for U(s) yields and it follows that, — Gals) F rls) = EQ) Again, considering the block diagram in Figure E2.9 we determine Fy(s) = Gi(s)Ga(s)[R(s) — Has) Fy(s) — Ho(s)Fe(s)] But, from the previous result, we substitute for F(s) resulting in F4(s) = Gi(s)Ga(s) R(s)~Gr(s)Ga(s) Ha(s) Fy (s)—G (s) Ha s)Ga(s)F (8) Solving for Fy(s) yields Lo) e Gr(s)Ga(s) n= [5 F Gino Hals) nA Gay a)£2.10 E211 tgs) eo {e} Las U6) | Gx) + Fx(s) RG) G6) Hx) FIGURE £2.9 Block diagram model ‘The shock absorber block diagram is shown in Figure E2.10. The closed- loop transfer function model is = — Gels) Gp(s)G(s) 1) = THe) Gls}Ggla)G) S| fem] ERS stn * oe on a i = ‘Piston travel Hs) FIGURE £2.10 Shock absorber block diagram Let f denote the spring force (n) and x denote the deflection (m). Then Af Kay. Computing the slope from the graph yields: (a) 9 = —O.1dm + K = Af/Ax = 10 n / 0.04 m = 250 n/m (b) t= 0m — K = Af/Ax = 10 n / 0.05 m = 200 nfm (0) % = 035m 4 K = Af/Ac =3n / 0.05 m = 60 n/mB2.12 The signal flow graph is shown in Fig, E212. Find ¥(s) when R(s) = 0. De) 1 x, Y! oe y Yo. 4 Ss FIGURE £2.12 Signal low graph ‘The transfer function from D(s) to ¥(s) is G(s)D(s) = KiK2G(s)D(s) _ Gis)(l— Ki K2)D(s) 1 — (—KG{s)) 1+ K2G{s) Kika =1 then ¥(s) =0 for any D(s). £2.13 Since we want to compute the transfer function from Ro(s) to Yi(s), we can assume that R, = 0 (application of the principle of superposition) ‘Then, starting at the output ¥;(s) we obtain As) rw {0} -—410-| a BO Gu Ys) GAs) Gls) Gols) we Hs) FIGURE €2:13 Block diagram modelE214 E2.15 ¥a(s) = Gals) [-H(s)¥i(s) + Gols) Gals) W(s) + Go(a)(s)] or [1 + G(s) Ha(s)]¥i(s) = [Gs(s)G2(s)Ge(s)1/(s) + Ga(s)Go(s)] Ws) Considering the signal W(s) (see Figure E2.13), we determine that W (3) = Gols) [Ga(s)Ra(s) ~ Hals)W(3)}, or [+ G(s) Ha(s)] W(s) = Ga(s)Ga(s) Rals) Substituting the expression for W(s) into the above equation for ¥i(s) yields Nils) __G2ls)Gs(s)G(s)Gs(s)Ga(s) + Ga(s)Ga(s)Gs(s)Go(s) Ra(s) 1+ Gs(s)Hi(s) + Gs(s)Ha(s) + Go(s)Gs(s)Hi(s) Ha(s) For loop 1, we have di 1 Rain + Gh + Ga [uw — ig)dt + Ro(is — ig) = v(t) And for loop 2, we have a dig ie _ Gf ind + La + Ralin —in) + a [la-ayae=o The transfer function from R(s) to P(s) is P(s) 42 RO) 43 Fas 443 ‘The block diagram is shown in Figure E2.15a. The corresponding signal flow graph is shown in Figure E2.15b for 42 SHIP LIT AD” P(s)/Rls) =+P(s) fa) 1 wo a wo yy > > Po a (b) FIGURE £2.15 (2) Block diagram, (b) Signal flow graph 2.16 A linear approximation for f is given by Akai: Ake where x) =1, Af = f(x) — f(z), and Ac =2~ 25 £2.17 The linear approximation is given by Ay = mde where (a) When 24 = 1, we find that yp = 24, and yp = 13.2 when ary = 2 (b) The slope m is computed as follows: ay orl, 144.205 ‘Therefore, m= 5.2 at to = 1, and m = 18.8 at zo =2 E2.18 The output (with a step input) is 10(s +2) Yl) s+ 3)(s+5)‘The partial fraction expansion is Taking the inverse Laplace transform yields Wd) = 5436 St _ 3e-3t £2.19 The input-ontput relationship is Vo _ AUK 1) Vv 1+AK where a Zit Ze Assume A> 1, Then, where Ro Zy id Z2= =. 1 Rest O° = Rast 1 ‘Therefore, Vols) RoA(RiCis+1) _ _ 2(8+1) Vis) Ri(RoCos +1) +2 152.20 The equation of motion of the mass me is mek y + (ba + bs sip + katy = bain + Kain Taking the Laplace transform with zero initial conditions yields [rmes? + (ba + bs)8 + kalXp(s) bas + ku} Xin(s) So, the transfer funetion is Xp(s) bas + ke 0.7842 Xin(s) — mos + (bat beer ha FIRSTS £2.21 The rotational velocity is 25(8+2) 1 (a) = (F5)s+ 1p 4s£2.22 £2.23 £2.24 E2.25 £2.26 Expanding in a partial fraction expansion yields 0.0234 0.1562 _ 0.2734 0.25 wls s ‘Taking the inverse Laplace transform yields w(t) = 0.25 + 0.0234e- — 0,1562te* — 0.2734 ‘The closed-loop transfer func YS) _ rj) = ____ ike Rls) 82+ (Ki + KK + Ki Ka)s + Ki Kok; © ‘The closed-loop tranfser function is ¥(s) _ 10 Ro) 71 = Beas $10 Let 2 6 and y = 0.8. Then, with y = az’, we have 0.8 = a(0.6)" Solving for a yields a = 3.704. A linear approximation is ‘o) Y- Yo 3ar3( or y = 4x — 1.6, where yp = 0.8 and ro = 06. ‘The equations of motion are my + hen 2) =F mya + k(ay — 21) = 0 ‘Taking the Laplace transform (with zero initial conditions) and solving for Xo(s) yields k 20) = Gat hime PO) Then, with my = my =k = 1, we have 1 XAl9)/FO = aeaTH ‘The transfer funetion from D(s) to ¥(s) is Gols) 1+ AGH)” ¥(s)/D(s)E227 £2.28 £2.29 ‘The transfer function is RoRiC . | RaRa 0) ct Vis) Rs RRs = 2s +144 (a) If 1 = apis OM Me)=2415, G(s) then the closed-loop transfer function of Figure E2.28(a) and (b) (in Dorf & Bishop) are equivalent. (b) The closed-loop transfer function is 1 10> Sy iTe$ (a) ‘The closed-loop transfer function is rs) = 2) 19 where G(s) T¥G@) F 428420 (b) The output ¥(s) (when R(s) = 1/s) is 0.12 0.24 O12 8) = FIGURE £2.29 Step response(c) The plot of y(?) is shown in Figure E2.29. The output is given by y(t) = 0.12 ~ 0.2467" + 0.126 2 £2.30 The pay fraction expansion is, a b v eaarinaner where py =p} = 4+ 22.9). Then, the residues are a= b= 118;