Solutions for Modern Control Systems 13th Us Edition by Dorf

Page 1


CHAPTER2

MathematicalModelsofSystems

Exercises

E2.1 Wehavefortheopen-loop y = r 2 andfortheclosed-loop e = r y and y = e 2

So, e = r e2 and e2 + e r =0 .

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREE2.1

Plotofopen-loopversusclosed-loop.

Forexample,if r =1,then e2 + e 1=0impliesthat e =0 618.Thus, y =0.382.Aplot y versus r isshowninFigureE2.1.

E2.2 Define

when R0 =10, 000Ω.Thus,thelinearapproximationiscomputedby consideringonlythefirst-ordertermsintheTaylorseriesexpansion,and isgivenby

R = 135∆T.

E2.3 Thespringconstantfortheequilibriumpointisfoundgraphicallyby estimatingtheslopeofalinetangenttotheforceversusdisplacement curveatthepoint y =0.5cm,seeFigureE2.3.Theslopeofthelineis K

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

y=Displacement (cm) For ce (n) Spring compresses Spring breaks

FIGUREE2.3

Springforceasafunctionofdisplacement.

CHAPTER2MathematicalModelsofSystems

E2.4 Since R(s)= 1 s wehave Y (s)= 6(s +50) s(s +30)(s +10)

Thepartialfractionexpansionof Y (s)isgivenby

where

UsingtheLaplacetransformtable,wefindthat

Thefinalvalueiscomputedusingthefinalvaluetheorem: lim t→∞ y(t)=lim s→0 s 6(s +50) s(s2 +40s +300) =1

E2.5 ThecircuitdiagramisshowninFigureE2.5.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREE2.5 Noninvertingop-ampcircuit.

Withanidealop-amp,wehave vo = A(vin v ),

Exercises 25 where A isverylarge.Wehavetherelationship

v = R1 R1 + R2 vo

Therefore, vo = A(vin R1 R1 + R2 vo), and solvingfor vo yields

vo = A 1+ AR1 R1+R2 vin

Since A ≫ 1,itfollowsthat1+ AR1 R1 +R2 ≈ AR1 R1 +R2 .Thentheexpressionfor vo simplifiesto

E2.6 Given

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

vo = R1 + R2 R1 vin

y = f (x)= ex andtheoperatingpoint xo =1,wehavethelinearapproximation y = f (x)= f (xo)+ ∂f ∂x x=xo (x xo)+ ·· · where f (xo)= e, df dx x=xo=1 = e, and x xo = x 1

Therefore,weobtainthelinearapproximation y = ex

E2.7 TheblockdiagramisshowninFigureE2.7.

FIGUREE2.7 Blockdiagrammodel.

CHAPTER2MathematicalModelsofSystems

Startingattheoutputweobtain I(s)= G1(s)G2(s)E(s).

But E(s)= R(s) H(s)I(s),so I(s)= G1(s)G2(s)[R(s) H(s)I(s)] .

Solvingfor I(s)yieldstheclosed-looptransferfunction

E2.8 TheblockdiagramisshowninFigureE2.8.

FIGUREE2.8

Blockdiagrammodel.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Startingattheoutputweobtain

Substituting

and with E(s)= R(s) Y (s)and Z(s)= sY (s)thisreducesto

Y (s)=[ G1(s)G2(s)(H2(s)+ H1(s)) G1(s)H3(s) 1 s G1(s)G2(s)K]Y (s)+ 1 s G1(s)G2 (s)KR(s)

Solvingfor Y (s)yieldsthetransferfunction

Y (s)= T (s)R(s), where

T (s)= KG1(s)G2 (s)/s 1+ G1(s)G2(s)[(H2(s) + H1(s)]+ G1(s)H3(s)+ KG1(s)G2(s)/s

E2.9 FromFigureE2.9,weobservethat

Ff (s)= G2(s)U (s) and

FR(s)= G3(s)U (s)

Then,solvingfor U (s)yields

U (s)= 1 G2(s) Ff (s) and itfollowsthat

FR(s

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Again,consideringtheblockdiagraminFigureE2.9wedetermine

But,fromthepreviousresult,wesubstitutefor FR(s)resultingin

Solvingfor Ff (s)yields

CHAPTER2MathematicalModelsofSystems

FIGUREE2.9

Blockdiagrammodel.

E2.10 TheshockabsorberblockdiagramisshowninFigureE2.10.Theclosedlooptransferfunctionmodelis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Piston travel measurement

FIGUREE2.10

Shockabsorberblockdiagram.

E2.11 Let f denotethespringforce(n)and x denotethedeflection(m).Then K = ∆f ∆x . Computingtheslopefromthegraphyields:

(a) xo = 0.14m → K =∆f/∆x =10n/0.04m=250n/m

(b) xo =0m → K =∆f/∆x =10n/0.05m=200n/m

(c) xo =0.35m → K =∆f/∆x =3n/0.05m=60n/m

E2.12 ThesignalflowgraphisshowninFig.E2.12.Find Y (s)when R(s)=0. Y(s) -1 K 2 G(s) -K 1 1 Td(s)

FIGUREE2.12

Signalflowgraph.

Thetransferfunctionfrom Td(s)to Y (s)is Y (s)= G(s)Td(s) K1K2G(s)Td(s) 1 ( K2G(s)) = G(s)(1 K1K2)Td(s) 1+ K2G(s)

Ifweset K1K2 =1 , then Y (s)=0forany Td(s).

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

E2.13 Thetransferfunctionfrom R(s), Td(s),and N (s)to Y (s)is Y (s)= K s2 +10s + K R(s)+ 1 s2 +10s + K Td(s) K s2 +10s + K N (s) Therefore,wefindthat Y (s)/Td(s)= 1 s2 +10s + K and Y (s)/N (s)= K s2 +10s + K

E2.14 Sincewewanttocomputethetransferfunctionfrom R2(s)to Y1(s),we canassumethat R1 =0(applicationoftheprincipleofsuperposition). Then,startingattheoutput Y1(s)weobtain

Consideringthesignal W (s)(seeFigureE2.14),wedeterminethat

CHAPTER2MathematicalModelsofSystems

FIGUREE2.14 Blockdiagrammodel. or

Substitutingtheexpressionfor W (s)intotheaboveequationfor Y1(s) yields

E2.15 Forloop1,wehave

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Andforloop2,wehave

E2.16 Thetransferfunctionfrom R(

)to P (

TheblockdiagramisshowninFigureE2.16a.Thecorrespondingsignal flowgraphisshowninFigureE2.16bfor

FIGUREE2.16

(a) Blockdiagram,(b)Signalflowgraph.

E2.17 Alinearapproximationfor f isgivenby

f = ∂f

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

= k

where xo =1/2,∆f = f (x) f (xo),and∆x = x xo

E2.18 Thelinearapproximationisgivenby ∆y = m∆x where m = ∂y ∂x x=xo

(a) When xo =1,wefindthat yo =2 4,and yo =13 2when xo =2.

(b)Theslope m iscomputedasfollows: m = ∂y ∂x x=xo =1+ 4 2x 2 o Therefore, m =5 2at xo =1,and m =18 8at xo =2.

CHAPTER2MathematicalModelsofSystems

E2.19 Theoutput(withastepinput)is Y (s)= 28(s +1) s(s +7)(s +2)

Thepartialfractionexpansionis Y (s)= 2 s 4 8 s +7 + 2 8 s +2 .

TakingtheinverseLaplacetransformyields

E2.20 Theinput-outputrelationshipis Vo V = A(K 1) 1+ AK where K = Z1 Z1 + Z2

Assume A ≫ 1.Then,

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Therefore,

E2.21 Theequationofmotionofthemass mc is

TakingtheLaplacetransformwithzeroinitialconditionsyields

So,thetransferfunctionis

E2.22 Therotationalvelocityis ω(s)= 2(s +4) (s +5)(s +1)2 1 s .

Expandinginapartialfractionexpansionyields

(

TakingtheinverseLaplacetransformyields

E2.23 Theclosed-looptransferfunctionis

(s) R(s) = T (s)= K1K2 s2 +(

.

E2.24 Let x =0.6and y =0.8.Then,with y = ax3,wehave 0.8= a(0.6)3 .

Solvingfor a yields a =3 704.Alinearapproximationis y yo =3ax 2 o(x xo) or y =4x 1 6,where yo =0 8and xo =0 6.

E2.25 Theclosed-looptransferfunctionis Y (s) R(s) = T

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

E2.26 Theequationsofmotionare

TakingtheLaplacetransform(withzeroinitialconditions)andsolving for X2(s)yields X2(s)= k (m

Then,with m1 = m2 = k =1,wehave

CHAPTER2MathematicalModelsofSystems

E2.27 Thetransferfunctionfrom Td(s)to Y (s)is

.

E2.28 Thetransferfunctionis

E2.29 (a)If

)= 1 s2 +15s +50 and H(s)=2s +15 , thentheclosed-looptransferfunctionofFigureE2.28(a)and(b)(in Dorf&Bishop)areequivalent.

(

(b)Theclosed-looptransferfunctionis T (s)= 1 s2 +17s +65

E2.30 (a)Theclosed-looptransferfunctionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREE2.30 Stepresponse.

(b) Theoutput Y (s)(when R(s)=1/s)is

Y (s)= 0 5 s + 0 25+ 0 1282j s +2 5 4 8734 + 0 25 0 1282j s +2 5+4 8734j or

Y (s)= 1 2 1 s s +5 s2 +5s +30

(c)Theplotof y(t)isshowninFigureE2.30.Theoutputisgivenby y(t)=0 5(1 1 1239e 2 5t sin(4 8734t +1 0968));

E2.31 Thepartialfractionexpansionis

V (s)= a s + p1 + b s + p2

where p1 =4 22j and p2 =4+22j.Then,theresiduesare a = 11 37jb =11 37j.

TheinverseLaplacetransformis v(t)= 11 37je( 4+22j)t +11 37je( 4 22j)t

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Problems

P2.1 Theintegrodifferentialequations,obtainedbyKirchoff’svoltagelawto eachloop,areasfollows:

P2.2 Thedifferentialequationsdescribingthesystemcanbeobtainedbyusing afree-bodydiagramanalysisofeachmass.Formass1and2wehave

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Usingaforce-currentanalogy,theanalagouselectriccircuitisshownin FigureP2.2,where Ci → Mi ,L1 → 1/k1 ,L12 → 1/k12 , and R → 1/b.

FIGUREP2.2

Analagouselectriccircuit.

P2.3 Thedifferentialequationsdescribingthesystemcanbeobtainedbyusing afree-bodydiagramanalysisofeachmass.Formass1and2wehave

Usingaforce-currentanalogy,theanalagouselectriccircuitisshownin FigureP2.3,where

FIGUREP2.3

Analagouselectriccircuit.

P2.4 (a)Thelinearapproximationaround vin =0is vo =0vin,seeFigureP2.4(a).

(b)Thelinearapproximationaround vin =1is vo =2vin 1,seeFigureP2.4(b).

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.4

Nonlinearfunctionsandapproximations.

CHAPTER2MathematicalModelsofSystems

P2.5 Given Q = K(P1 P2)1/2

Let δP = P1 P2 and δPo =operatingpoint.UsingaTaylorseries expansionof Q,wehave Q = Qo + ∂Q

Define∆Q = Q Qo and∆P = δP δPo.Then,droppinghigher-order termsintheTaylorseriesexpansionyields ∆Q = m∆P where m = K 2δP 1

P2.6 FromP2.1wehave

and

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

TakingtheLaplacetransformandusingthefactthattheinitialvoltage across C2 is10vyields

and

Rewritinginmatrixformwehave

Solvingfor I2 yields

P2.7 Considerthedifferentiatingop-ampcircuitinFigureP2.7.Foranideal op-amp,thevoltagegain(asafunctionoffrequency)is

and Z2 = R2 aretherespectivecircuitimpedances.Therefore,weobtain

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.7

Differentiatingop-ampcircuit.

CHAPTER2MathematicalModelsofSystems

P2.8 Let

Then,

Therefore,thetransferfunctionis

FIGUREP2.8 Pole-zeromap.

Pole-zero map (x:poles and o:zeros)

Using R1 =1.0, R2 =0.5,and C =0.5,wehave T (s)= s2 +4s +8 s

The pole-zeromapisshowninFigureP2.8.

P2.9 FromP2.3wehave

.

TakingtheLaplacetransformofbothequationsandwritingtheresultin matrixform,itfollowsthat

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.9 Pole-zeromap.

CHAPTER2MathematicalModelsofSystems or

where∆=(Ms2 + bs + k)(Ms2 +2k) k2 . So, G(s)= X1(s) F (s) =

.

When b/k =1, M =1, b2/Mk =0.04,wehave G(s)= s2 +0 04s +0 04 s4 +0.04s3 + 0.12s2 +0.0032s +0.0016

Thepole-zeromapisshowninFigureP2.9.

P2.10 FromP2.2wehave

TakingtheLaplacetransformofbothequationsandwritingtheresultin matrixform,itfollowsthat

or

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

where

So,when f (t)= a sin

,wehavethat

Formotionlessresponse(inthesteady-state),setthezeroofthetransfer functionsothat

P2.11

Thetransferfunctionsfrom Vc(s)to Vd(s)andfrom Vd(s)to θ(s)are:

and

Theblockdiagramfor θ(s)/Vc(s)isshowninFigureP2.11,where

(s)/Vc(s)= θ(s) Vd(s) V

(s) Vc(s) = K1K2K

∆(s) , where

FIGUREP2.11

Blockdiagram.

P2.12 Theopen-looptransferfunctionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

With R(s)=1/s,wehave Y (s)= K s(s +50) . Thepartialfractionexpansionis Y (s)= K 50 1 s 1 s +50 , andtheinverseLaplacetransformis y(t)= K 50 1 e 50t , As t → ∞,itfollowsthat y(t) → K/50.Sowechoose K =50sothat y(t)

CHAPTER2MathematicalModelsofSystems

approaches1.Alternativelywecanusethefinalvaluetheoremtoobtain

Itfollowsthatchoosing K =50leadsto y(t) → 1as t →∞.

P2.13 Themotortorqueisgivenby

Combiningtheaboveexpressionsyields

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

P2.14 Forafield-controlleddcelectricmotorwehave

Withastepinputof Vf (s)=80/s,thefinalvalueof ω(t)is

(t)t→∞ =lim s→0 sω(s)= 80Km Rf b =2.4or Km Rf b =0.03 .

Solvingfor ω(t)yields

At t =1/2, ω(t)=1,so ω(1/2)=2 4(1 e (b/J )t)=1implies b/J =1 08sec

Therefore, ω(s)/Vf (s)= 0 0324 s +1.08

P2.15 SummingtheforcesintheverticaldirectionandusingNewton’sSecond Lawweobtain x + k m x =0

Thesystemhasnodampingandnoexternalinputs.TakingtheLaplace transformyields

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

X(s)= x0s s2 + k/m ,

whereweusedthefactthat x(0)= x0 and˙x(0)=0.Thentakingthe inverseLaplacetransformyields x(t)= x0 cos k m t.

P2.16 (a)Formass1and2,wehave

(b)TakingtheLaplacetransformyields

(c)Let

FIGUREP2.16

Signalflowgraph.

CHAPTER2MathematicalModelsofSystems

G2(s)=1/p(s)

G3(s)=1/q(s)

G4(s)= sb1 , where

and

ThesignalflowgraphisshowninFigureP2.16.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(d)Thetransferfunctionfrom X3(s)to X1(s)is

P2.17 UsingCramer’srule,wehave

The signalflowgraphisshowninFigureP2.17.

FIGUREP2.17

Signalflowgraph.

P2.18 ThesignalflowgraphisshowninFigureP2.18.

FIGUREP2.18

Signalflowgraph.

Thetransferfunctionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

P2.19 (a)Assume Rg ≫ Rs and Rs ≫ R1.Then Rs = R1 + R2 ≈ R2,and vgs = vin vo , whereweneglect iin,since Rg ≫ Rs.AtnodeS,wehave vo Rs = gmvgs = gm(vin vo)or vo vin = gmRs 1+ gmRs . (b) With gmRs =20,wehave vo vin = 20 21 =0.95 .

CHAPTER2MathematicalModelsofSystems

(c)TheblockdiagramisshowninFigureP2.19. gmRsvin(s) vo(s)

FIGUREP2.19

Blockdiagrammodel.

P2.20 Fromthegeometrywefindthat

Theflowratebalanceyields A dy dt = p∆z whichimplies Y (s)= p∆Z(s) As

Bycombiningtheaboveresultsitfollowsthat

Therefore,thesignalflowgraphisshowninFigureP2.20.UsingMason’s

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.20

Signalflowgraph.

gainformulawefindthatthetransferfunctionisgivenby

where

P2.21 (a)Theequationsofmotionforthetwomassesare

FIGUREP2.21 (a) Blockdiagram.(b)Pole-zeromap.

CHAPTER2MathematicalModelsofSystems

(b) Define a = g/L + k/4M and b = k/4M .Then

1(s)

(c) Theblockdiagramandpole-zeromapareshowninFigureP2.21.

P2.22 Foranoninvertingop-ampcircuit,depictedinFigureP2.22a,thevoltage gain(asafunctionoffrequency)is

Vo(s)= Z1(s)+ Z2(s) Z1(s) Vin(s), where Z1(s)and Z2(s)aretheimpedancesoftherespectivecircuits.In

FIGUREP2.22 (a) Noninvertingop-ampcircuit.(b)Voltagefollowercircuit.

thecaseofthevoltagefollowercircuit,showninFigureP2.22b,wehave Z1 = ∞ (opencircuit)and Z2 =0.Therefore,thetransferfunctionis

P2.23 Theinput-outputratio, Vce/Vin,isfoundtobe

P2.24 (a)Thevoltagegainisgivenby

(b) Thecurrentgainisfoundtobe

(c) Theinputimpedanceis

+ R2 , and when β1β2 isverylarge,wehavetheapproximation

P2.25 Thetransferfunctionfrom R(s)and Td(s)to Y (s)isgivenby

Thus,

(s)/R(s)= G(s) . Also,wehavethat Y (s)=0 when R(s)=0.Therefore,theeffectofthedisturbance, Td(s),iseliminated.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

P2.26 Theequationsofmotionforthetwomassmodeloftherobotare

x + b(˙x y)+ k(x y)= F (t) m ¨ y + b(˙ y ˙ x)+ k(y x)=0 .

TakingtheLaplacetransformandwritingtheresultinmatrixformyields

Solvingfor Y (s)wefindthat

CHAPTER2MathematicalModelsofSystems

P2.27 Thedescribingequationofmotionis

Defining

(

)=

leadsto ¨ z = f (z,i) .

Theequilibriumconditionfor io and zo,foundbysolvingtheequationof motionwhen

is

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

WelinearizetheequationofmotionusingaTaylorseriesapproximation. Withthedefinitions ∆z = z zo and∆i = i io , wehave ˙ ∆z =˙ z and ¨ ∆z =¨ z.Therefore,

z = f (z,i)= f (zo,io)+

But f (zo,io)=0,andneglectinghigher-ordertermsintheexpansion yields

Usingtheequilibriumconditionwhichrelates zo to io,wedeterminethat

TakingtheLaplacetransformyieldsthetransferfunction(validaround theequilibriumpoint)

P2.28 ThesignalflowgraphisshowninFigureP2.28.

FIGUREP2.28 Signalflowgraph.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning.

(a)ThePGBDPloopgainisequalto-abcd.Thisisa negative transmissionsincethepopulationproducesgarbagewhichincreasesbacteria andleadstodiseases,thusreducingthepopulation.

(b)ThePMCPloopgainisequalto+efg.Thisisa positive transmissionsincethepopulationleadstomodernizationwhichencourages immigration,thusincreasingthepopulation.

orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(c)ThePMSDPloopgainisequalto+ehkd.Thisisa positive transmissionsincethepopulationleadstomodernizationandanincrease insanitationfacilitieswhichreducesdiseases,thusreducingtherate ofdecreasingpopulation.

(d)ThePMSBDPloopgainisequalto+ehmcd.Thisisa positive transmissionbysimilarargumentasin(3).

P2.29 Assumethemotortorqueisproportionaltotheinputcurrent Tm = ki.

Then,theequationofmotionofthebeamis

¨ φ = ki,

where J isthemomentofinertiaofthebeamandshaft(neglectingthe inertiaoftheball).Weassumethatforcesactingontheballaredueto gravityandfriction.Hence,themotionoftheballisdescribedby mx = mgφ bx

CHAPTER2MathematicalModelsofSystems

where m isthemassoftheball, b isthecoefficientoffriction,andwe haveassumedsmallangles,sothatsin φ ≈ φ.TakingtheLaplacetransfor ofbothequationsofmotionandsolvingfor X(s)yields

X(s)/I(s)= gk/J s2(s2 + b/m) .

P2.30 Given H(s)= k τs +1

where τ =5µs =4 × 10 6 secondsand0 999 ≤ k< 1 001.Thestep responseis

Y (s)= k τs +1 · 1 s = k s k s +1/τ

TakingtheinverseLaplacetransformyields

y(t)= k ke t/τ = k(1 e t/τ ) .

Thefinalvalueis k.Thetimeittakestoreach98%ofthefinalvalueis t =19 57µs independentof k

P2.31 Fromtheblockdiagramwehave

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Therefore, Y1(s)=

And,computing E2(s)(with R2(s)=0)wefind

Substituting E2(s)intoequationfor Y1(s)yields

Finally,solvingfor Y1(s)yields

Similarly,for Y2(s)weobtain

P2.32

Thesignalflowgraphshowsthreeloops:

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Thetransferfunction Y2/R1 isfoundtobe

, whereforpath1

∆1 =1

andforpath2

∆2 =1 L1

Sincewewant Y2 tobeindependentof R1,weneed Y2/R1 =0.Therefore, werequire

CHAPTER2MathematicalModelsofSystems

P2.33 Theclosed-looptransferfunctionis

(

) =

P2.34 Theequationsofmotionare

TakingtheLaplacetransformyields

Therefore,aftersolvingfor Y1(s)/X(s),wehave

P2.35 (a)WecanredrawtheblockdiagramasshowninFigureP2.35.Then,

(b)Thesignalflowgraphrevealstwoloops(bothtouching):

Therefore,

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(c)Wewanttochoose K1 and K2 suchthat

+10)2

Therefore, K1 =100and1+ K2K1 =20or K2 =0 19.

(d)ThestepresponseisshowninFigureP2.35.

FIGUREP2.35

The equivalentblockdiagramandthesystemstepresponse.

P2.36 (a)Given R(s)=1/s2,thepartialfractionexpansionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Therefore,usingtheLaplacetransformtable,wedeterminethatthe rampresponsefor t ≥ 0is

(b) Fortherampinput, y(t) ≈ 0 25at t =1second(seeFigureP2.36a).

(c)Given R(s)=1,thepartialfractionexpansionis

(s)= 30 (

Therefore,usingtheLaplacetransformtable,wedeterminethatthe

CHAPTER2MathematicalModelsofSystems

impulseresponsefor t ≥ 0is

(d) Fortheimpulseinput, y(t) ≈ 0 73at t =1seconds(seeFigureP2.36b).

(a) Ramp input

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.36 (a) Rampinputresponse.(b)Impulseinputresponse.

P2.37 Theequationsofmotionare

When

(b) Impulse input

P2.38 Theequationofmotionforthesystemis J d2θ dt2 + b dθ dt + kθ =0 , where k istherotationalspringconstantand b istheviscousfriction coefficient.Theinitialconditionsare θ(0)= θo and θ(0)=0.Takingthe Laplacetransformyields

Therefore,

(s)= (s + b J θo) (s2 + b J s + K J ) = (s +2ζωn)θo s2 +2ζωns + ω2 n .

Neglectingthemassoftherod,themomentofinertiaisdeteminedtobe

Also,

Solvingfor θ(t),wefindthat

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

+ φ) , wheretan φ = 1 ζ 2/ζ).Therefore,theenvelopedecayis θe = θo 1 ζ 2 e ζωnt .

So,with ζωn =2 × 10 4 , θo =4000o and θf =10o,theelapsedtimeis computedas

P2.39 When t< 0,wehavethesteady-stateconditions i1(0)= 6 7 A,va(0)= 12 7 V and vc(0)= 36 7 V, where vc(0)isassociatedwiththe0.75Fcapacitor.After t ≥ 0,wehave

CHAPTER2MathematicalModelsofSystems and

TakingtheLaplacetransform(usingtheinitialconditions)yields

Solvingfor I2(s)yields I2(s)= 4s(27s2 +216s +604) 7(s +2)(60s2 +203s +14)

Then, Vo(s)=10I2(s).

P2.40 Theequationsofmotionare

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

TakingtheLaplacetransformyields (J1s 2 + bs + K)θ1(s) bsθ2(s)=

and(

),wefind that

)) ∆(s) , where∆(

P2.41 Assumethattheonlyexternaltorquesactingontherocketarecontrol torques, Tc anddisturbancetorques, Td,andassumesmallangles, θ(t). Usingthesmallangleapproximation,wehave ˙ h = Vθ

Jθ = Tc + Td , where J isthemomentofinertiaoftherocketand V istherocketvelocity (assumedconstant).Now,supposethatthecontroltorqueisproportional tothelateraldisplacement,as

Tc(s)= KH(s) , wherethenegativesigndenotesanegativefeedbacksystem.ThecorrespondingblockdiagramisshowninFigureP2.41.

FIGUREP2.41 Blockdiagram.

P2.42 (a)Theequationofmotionofthemotoris

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

J dω dt = Tm bω, where J =0 1, b =0 06,and Tm isthemotorinputtorque. (b)Given Tm(s)=1/s,and ω(0)=0 7,wetaketheLaplacetransform oftheequationofmotionyielding

sω(s) ω(0)+0.6ω(s)=10Tm or

ω(s)= 0 7s +10 s(s +0.6) Then,computingthepartialfractionexpansion,wefindthat ω(s)= A s + B s +0.6 = 16 67 s 15 97 s +0.6 Thestepresponse,determinedbytakingtheinverseLaplacetransform,is ω(t)=16.67 15.97e 0 6t ,t ≥ 0 .

CHAPTER2MathematicalModelsofSystems

P2.43 Theworkdonebyeachgearisequaltothatoftheother,therefore

Also,thetraveldistanceisthesameforeachgear,so

Thenumberofteethoneachgearisproportionaltotheradius,or

So,

and

Finally,

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

P2.44 Theinertiaoftheloadis

Also,fromthedynamicswehave

and

So,

since ω2 = nω1 .

Therefore,thetorqueatthemotorshaftis

T = T1 + Tm = n 2(JL ˙ ω1 + bLω1)+ Jm ˙ ω1 + bmω1 .

P2.45 Let U (s)denotethehumaninputand F (s)theloadinput.Thetransfer functionis

P (s)= G(s)+ KG1(s) ∆(s) U (s)+ Gc(s)+ KG1(s) ∆(s) F (s) , where

∆=1+ GH(s)+ G1KBH(s)+ GcE(s)+ G1KE(s)

P2.46 ConsidertheapplicationofNewton’slaw( F = m ¨ x).Fromthemass mv weobtain

mv ¨ x1 = F k1(x1 x2) b1(˙x1 ˙ x2)

TakingtheLaplacetransform,andsolvingfor X1(s)yields

X1(s)= 1 ∆1(s) F (s)+ b1s + k1 ∆1(s) X2(s), where

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Fromthemass mt weobtain

1(˙x1 x2).

TakingtheLaplacetransform,andsolvingfor X2(s)yields

X2(s)= b1s + k1 ∆2(s) X1(s), where ∆2 := mts 2 +(b1 + b2)s + k1 + k2.

Substituting X2(s)aboveintotherelationshipfpr X1(s)yieldsthetransferfunction

X1(s) F (s) = ∆2(s) ∆1(s)∆2(s) (b1s + k1)2

CHAPTER2MathematicalModelsofSystems

P2.47 Usingthefollowingrelationships

h(t)= (1 6θ(t) h(t))dt

ω(t)= ˙ θ(t)

J ˙ ω(t)= Kmia(t)

va(t)=50vi(t)=10ia(t)+ vb(t)

˙ θ = Kvb

wefindthedifferentialequationis

P2.48 (a)Thetransferfunctionis

V2(s)

V1(s) = (1+ sR1C1)(1+ sR2C2) R1C2s

(b) When R1 =250 kΩ, R2 =200 kΩ, C1 =2 µF and C2 =0 1 µF ,we have

V2(s) V1(s) = 0.4s2 +20.8s + 40 s .

(c) Thepartialfractionexpansionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

V2(s) V1(s) =20 8 + 40 s +0 4s.

P2.49 (a)Theclosed-looptransferfunctionis T (s)= G(s) 1+ G(s) = 5000 s3 +20s2 +1000s +5000 .

(b)Thepolesof T (s)are s1 = 5.43and s2,3 = 7.28 ± j29.46.

(c)Thepartialfractionexpansion(withastepinput)is Y (s)= 1 s 1.06 s +5 43 + 0.0285+0.0904j s +7 28 j29 46 + 0.0285 0.0904j s +7 28+ j29 46 , and y(t)=1 1 06e 5 43t +0 06e 7 28t (cos29 46t 3 17sin29 46t);

(d)ThestepresponseisshowninFigureP2.49.Therealandcomplex rootsareclosetogetherandbylookingatthepolesinthes-planewe

P2.50

FIGUREP2.49 Stepresponse.

havedifficultydecidingwhichisdominant.However,theresidueat therealpoleismuchlargerandthusdominatestheresponse.

(a)Theclosed-looptransferfunctionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(b)Thepolesof T (s)are s1 = 5and s2,3 = 20 ± j50

.

(c)Thepartialfractionexpansion(withastepinput)is

.

(d)ThestepresponseisshowninFigureP2.50.Therealrootdominates theresponse.

(e)Thefinalvalueof y(t)is yss =lim s→0 sY (s)=0.9655 .

CHAPTER2MathematicalModelsofSystems

FIGUREP2.50 Stepresponse.

P2.51 ConsiderthefreebodydiagraminFigureP2.51.UsingNewton’sLaw andsummingtheforcesonthetwomassesyields

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREP2.51 Freebodydiagram.

AdvancedProblems

AP2.1

Thetransferfunctionfrom V (s)to ω(s)hastheform

ω(s) V (s) = Km τms +1

In thesteady-state,

ss =lim s→0 s Km τms +1 5 s =5Km So, Km =70/5=14 .

Also,

where V (s)= Vm/s.Solvingfor τm yields

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(t)= VmKm(1 e t/τm )

m = t ln(1 ω(t)/ωss)

When t =2,wehave τm = 2 ln(1 30/70) = 3 57

Therefore,thetransferfunctionis ω(s) V (s) = 14 3.57s +1

AP2.2 Theclosed-looptransferfunctionform

Ifweselect G5(s)= G2G3G

thenthenumeratoriszero,and Y2(s)/R1(s)=0.Thesystemisnow decoupled.

CHAPTER2MathematicalModelsofSystems

AP2.3 (a)Computingtheclosed-looptransferfunction:

Y (s)= G(s)Gc(s) 1+ Gc(s)G(s)H(s) R(s)

Then,with E(s)= R(s) Y (s)weobtain E(s)= 1+ Gc(s)G(s)(H(s) 1) 1+ Gc(s)G(s)H(s) R(s)

Ifwerequirethat E(s) ≡ 0foranyinput,weneed1+ Gc(s)G(s)(H(s) 1)=0or

H(s)= Gc(s)G(s) 1 Gc(s)G(s) = n(s) d(s)

Sincewerequire H(s)tobeacausalsystem,theorderofthenumerator polynomial, n(s),mustbelessthanorequaltotheorderofthedenominatorpolynomial, d(s).Thiswillbetrue,ingeneral,onlyifboth Gc(s) and G(s)areproperrationalfunctions(thatis,thenumeratoranddenominatorpolynomialshavethesameorder).Therefore,making E ≡ 0 foranyinput R(s)ispossibleonlyincertaincircumstances.

(b)Thetransferfunctionfrom Td(s)to Y (s)is

Y (s)= Gd(s)G(s) 1+ Gc(s)G(s)H(s) Td(s) .

With H(s)asinpart(a)wehave

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

(s)= Gd(s) Gc(s) Td(s) .

(c) No.Since

theonlywaytohave Y (s) ≡ 0forany Td(s)isforthetransferfunction T (s) ≡ 0whichisnotpossibleingeneral(since G(s) =0).

AP2.4 (a)With q(s)=1/s weobtain

Then,itfollowsthat

TakingtheinverseLaplacetransformyields

Qs+1/R (c)Toincreasethespeedofresponse,youwanttochoose Ct, Q, S and R suchthat

AP2.5 Consideringthemotionofeachmass,wehave

Inmatrixformthethreeequationscanbewrittenas

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

AP2.6 ConsideringthecartmassandusingNewton’sLawweobtain M x = u bx F sin ϕ where F isthereactionforcebetweenthecartandthependulum.Consideringthependulumweobtain m d2(x + L sin ϕ) dt2 = F

CHAPTER2MathematicalModelsofSystems

m d2(L cos ϕ) dt2 = F cos ϕ + mg

Eliminatingthereactionforce F yieldsthetwoequations

(m + M )¨ x + b ˙ x + mL ¨ ϕ cos ϕ mL ˙ ϕ 2 sin ϕ = u mL2 ϕ + mgL sin ϕ + mLx cos ϕ =0

Ifweassumethattheangle ϕ ≈ 0,thenwehavethelinearmodel

(m + M )¨ x + bx + mLϕ = u

mL2 ¨ ϕ + mgLϕ = mL¨ x

AP2.7 Thetransferfunctionfromthedisturbanceinputtotheoutputis

Y (s)= 1 s +40+ K Td(s)

When Td(s)=1,weobtain

y(t)= e (40+K )t

Solvingfor t when y(t) < 0.1yields t> 2.3 40+ K

When t =0.05and y(0.05)=0.1,wefind K =6.05.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

AP2.8 Theclosed-looptransferfunctionis

T (s)= 200K(0.25s +1) (0 25s +1)(s + 1)(s +8)+200K

Thefinalvalueduetoastepinputof R(s)= A/s is

v(t) → A 200K 200K +8 .

We needtoselect K sothat v(t) → 50.However,tokeepthepercent overshoottolessthan10%,weneedtolimitthemagnitudeof K.FigureAP2.8ashowsthepercentovershootasafunctionof K.Let K =0.06 andselectthemagnitudeoftheinputtobe A =83 3.TheinverseLaplace transformoftheclosed-loopresponsewith R(s)=83.3/s is

v(t)=50+9 85e 9 15t e 1 93t (59 85cos(2 24t)+11 27sin(2 24t))

The resultis P.O. =9 74%andthesteady-statevalueoftheoutputis approximately50m/s,asshowninFigureAP2.8b.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGUREAP2.8 (a) Percentovershootversusthegain K.(b)Stepresponse.

AP2.9 Thetransferfunctionis

CHAPTER2MathematicalModelsofSystems

Thenwecanwrite

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

DesignProblems 73

DesignProblems

CDP2.1 closelythearmature-controlleddcmotormodeldepictedinFigure2.18 inDorfandBishop.Thetransferfunctionis

Themodelofthetractiondrive,capstanroller,andlinearslidefollows

DP2.1 Theclosed-looptransferfunctionis

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

When

DP2.2 Atthelowernodewehave

Also,wehave v =24and i2 = Gv. So

CHAPTER2MathematicalModelsofSystems and

DP2.3 TakingtheLaplacetransformof

yields

Similarly,takingtheLaplacetransformoftherampinputyields

Therefore

DP2.4 Foranidealop-amp,atnodeawehave

andatnodeb

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

fromitfollowsthat

Also,foranidealop-amp, Vb Va =0.Thensolvingfor Vb intheabove equationandsubstitutingtheresultintothenodeaequationfor Va yields

or

DesignProblems

For vin(t) = At,wehave Vin(s)= A/s2,therefore vo(t)= A 2 β e βt + t 2 β

where β =1/R2C

DP2.5 Theequationofmotiondescribingthemotionoftheinvertedpendulum (assumingsmallangles)is

ϕ + g L ϕ =0

Assumingasolutionoftheform ϕ = k cos ϕ,takingtheappropriate derivativesandsubstitutingtheresultintotheequationofmotionyields therelationship

ϕ = g L .

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Iftheperiodis T =2seconds,wecompute˙ϕ =2π/T .Thensolvingfor L yields L =0 99meterswhen g =9 81m/s2.So,tofitthependuluminto thegrandfatherclock,thedimensionsaregenerallyabout1.5metersor more.

CHAPTER2MathematicalModelsofSystems

ComputerProblems

CP2.1 Them-filescriptisshowninFigureCP2.1.

p=[1 8 12]; q=[1 2];

% Part (a)

pq=conv(p,q)

% Part (b)

P=roots(p), Z=roots(q)

% Part (c)

value=polyval(p,-1)

FIGURECP2.1

Scriptforvariouspolynomialevaluations.

CP2.2 Them-filescriptandstepresponseisshowninFigureCP2.2.

numc = [1]; denc = [1 1]; sysc = tf(numc,denc) numg = [1 2]; deng = [1 3]; sysg = tf(numg,deng)

% part (a)

sys_s = series(sysc,sysg); sys_cl = feedback(sys_s,[1])

% part (b) step(sys_cl); grid on

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGURECP2.2 Stepresponse.

CP2.3 Given

y +6˙ y +5y = u

with y(0)=˙ y =0and U (s)=1/s,weobtain(viaLaplacetransform) Y (s)= 1 s(s2 +6s +5) = 1 s(s +5)(s +1)

Expandinginapartialfractionexpansionyields

.

TakingtheinverseLaplacetransformweobtainthesolution

(t)=0.2+0.

. Them-filescriptandstepresponseisshowninFigureCP2.3.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

n=[1]; d=[1 6 5]; sys = tf(n,d); t=[0:0.1:5]; y = step(sys,t); ya=0.2+0.05*exp(-5*t)-0.25*exp(-t); plot(t,y,t,ya); grid; title('Step Response'); xlabel('Time (s)'); ylabel('Amplitude');

FIGURECP2.3 Stepresponse.

CP2.4 Themass-spring-dampersystemisrepresentedby mx + b ˙ x + kx = f.

TakingtheLaplacetransform(withzeroinitialconditions)yieldsthe transferfunction

X(s)/F (s)= 1/m s2 + bs/m + k/m

Them-filescriptandstepresponseisshowninFigureCP2.4.

m=10; k=1; b=0.5; num=[1/m]; den=[1 b/m k/m]; sys = tf(num,den); t=[0:0.1:150]; step(sys,t)

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGURECP2.4 Stepresponse.

CP2.5 ThespacecraftsimulationsareshowninFigureCP2.5.Weseethatas J isdecreased,thetimetosettledowndecreases.Also,theoverhootfrom 10o decreasesas J decreases.Thus,theperformanceseemstogetbetter (insomesense)as J decreases.

%Part (a)

a=1; b=8; k=10.8e+08; J=10.8e+08; num=k*[1 a]; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]);

%

% Part (b) and (c) t=[0:0.1:100];

%

% Nominal case f=10*pi/180; sysf=sys_cl*f; y=step(sysf,t);

%

% O -nominal case 80%

J=10.8e+08*0.8; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f; y1=step(sysf,t);

%

% O -nominal case 50%

J=10.8e+08*0.5; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f; y2=step(sysf,t);

% plot(t,y*180/pi,t,y1*180/pi,'--',t,y2*180/pi,':'),grid xlabel('Time (sec)')

ylabel('Spacecraft attitude (deg)')

title('Nominal (solid); O -nominal 80% (dashed); O -nominal 50% (dotted)')

FIGURECP2.5

Stepresponsesforthenominalandoff-nominalspacecraftparameters.

CP2.6 Theclosed-looptransferfunctionis

num1=[4]; den1=[1]; sys1 = tf(num1,den1); num2=[1]; den2=[1 1]; sys2 = tf(num2,den2); num3=[1 0]; den3=[1 0 2]; sys3 = tf(num3,den3); num4=[1]; den4=[1 0 0]; sys4 = tf(num4,den4); num5=[4 2]; den5=[1 2 1]; sys5 = tf(num5,den5); num6=[50]; den6=[1]; sys6 = tf(num6,den6); num7=[1 0 2]; den7=[1 0 0 14]; sys7 = tf(num7,den7); sysa = feedback(sys4,sys6,+1); sysb = series(sys2,sys3); sysc = feedback(sysb,sys5); sysd = series(sysc,sysa); syse = feedback(sysd,sys7); sys = series(sys1,syse)

% pzmap(sys) % p=pole(sys) z=zero(sys)

7.0709 -7.0713 1.2051 + 2.0863i 1.2051 - 2.0863i 0.1219 + 1.8374i 0.1219 - 1.8374i -2.3933 -2.3333 -0.4635 + 0.1997i -0.4635 - 0.1997i

z = 0 1.2051 + 2.0872i 1.2051 - 2.0872i -2.4101 -1.0000 + 0.0000i -1.0000 - 0.0000i

where

s 2

s 1400 .

CP2.7 Them-filescriptandplotofthependulumangleisshowninFigureCP2.7. Withtheinitialconditions,theLaplacetransformofthelinearsystemis

θ(s)= θ0s s2 + g/L

To usethestepfunctionwiththem-file,wecanmultiplythetransfer functionasfollows: θ(s)= s2 s2 + g/L θ0 s , whichisequivalenttotheoriginaltransferfunctionexceptthatwecan usethestepfunctioninputwithmagnitude θ0.Thenonlinearresponse isshownasthesolidlineandthelinearresponseisshownasthedashed line.Thedifferencebetweenthetworesponsesisnotgreatsincetheinitial conditionof θ0 =30◦ isnotthatlarge.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

L=0.5; m=1; g=9.8; theta0=30; % Linear simulation sys=tf([1 0 0],[1 0 g/L]); [y,t]=step(theta0*sys,[0:0.01:10]); % Nonlinear simulation [t,ynl]=ode45(@pend,t,[theta0*pi/180 0]); plot(t,ynl(:,1)*180/pi,t,y,' '); xlabel('Time (s)') ylabel('\theta (deg)')

function [yd]=pend(t,y) L=0.5; g=9.8; yd(1)=y(2); yd(2)=-(g/L)*sin(y(1)); yd=yd';

Plotof θ versus xt when θ0 =30◦

FIGURECP2.7

CHAPTER2MathematicalModelsofSystems

CP2.8 Thesystemstepresponsesfor z =5, 10,and15areshowninFigureCP2.8.

FIGURECP2.8 The systemresponse.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

CP2.9 (a,b)Computingtheclosed-looptransferfunctionyields

T (s)= G(s) 1+ G(s)H(s) = s2 +2s +1 s2 +4s +3 .

Thepolesare s = 3, 1andthezerosare s = 1, 1. (c)Yes,thereisonepole-zerocancellation.Thetransferfunction(after pole-zerocancellation)is

T (s)= s +1 s +3

ng=[1 1]; dg=[1 2]; sysg = tf(ng,dg); nh=[1]; dh=[1 1]; sysh = tf(nh,dh); sys=feedback(sysg,sysh)

% pzmap(sys)

% pole(sys) zero(sys)

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

FIGURECP2.9 Pole-zeromap.

CP2.10 FigureCP2.10showsthesteady-stateresponsetoaunitstepinputanda unitstepdisturbance.Weseethat K =1leadstothesamesteady-state response.

K=[0.1:0.1:10]; sysg=tf([1],[1 20 20]); for i=1:length(K) nc=K(i); dc=[1];sysc=tf(nc,dc); syscl=feedback(sysc*sysg,1); systd=feedback(sysg,sysc); y1=step(syscl); Tf1(i)=y1(end); y2=step(systd); Tf2(i)=y2(end); end plot(K,Tf1,K,Tf2,'--') xlabel('K') ylabel('Steady-state response')

FIGURECP2.10

Gain K versussteady-statevalue.

ThisworkisprotectedbyUnitedStatescopyrightlaws andisprovidedsolelyfortheuseof instructorsinteachingtheircoursesandassessingstudentlearning. Dissemination orsaleofanypartofthiswork(includingontheWorldWideWeb) willdestroytheintegrityoftheworkandisnotpermitted.

Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.
Solutions for Modern Control Systems 13th Us Edition by Dorf by 6alsm - Issuu