CHAPTER 4 7 CONCEPTS OF STABILITY AND ROOT LOCUS 4.1__IMPULSE RESPONSE AND STABILITY DEFINITIONS OF STABILITY The term stability refers to the stable working condition of a control system. Every working system is designed to be stable. In a stable system, the response or output is predictable, finite and stable for a given input (or for any changes in input or for any changes in system parameters). ‘The different definitions of the stability are the following 1. A system is stable, if its output is bounded (finite) for any bounded (finite) input. 2. A system is asymptotically stable, if in the absence of the input, the output tends towards zero (or to the equilibrium state) irrespective of initial conditions. 3. A system is stable if for a bounded disturbing input signal the output vanishes ultimately as t approaches infinity. 4. A system is unstable if for a bounded disturbing input signal the output is of infinite amplitude or oscillatory. 5. Fora bounded input signal, if the output has constant amplitude oscillations then the system may be stable or unstable under some limited constraints. Such a system is called limitedly stable. 6. Ifasystem outputs stable for all variations of its parameters, then the system is called absolutely stable system. 7. Ifa system output is stable for a limited range of variations of its parameters, then the system is called conditionally stable system. IMPULSE RESPONSE OF A SYSTEM Let, M(s) = Closed loop transfer function of a system. C(s) = Output / Response in s-domain, R(S) = Input in s-domain Now, Mis) = £2. RG) -. Response or Output in s-domain, C(s) = M(s) Ris) Now, Response in time domain, (ft) = £*{C(s)} Input in time domain, Tt) =L{R@)} For an impulse input, r(t)= 8(f) ; -. R(s) = £[8(t)] =14.2 1EM(s) R(s)}. = £7 {M(s)} = m(t) woe 4.1) Hence, impulse response of a system is the inverse Laplace transform of system transfer function. .. Impulse response The importance of impulse response is that, the output of a system for any arbitrary input can be obtained by convolution of input and impulse response. ie., Response, o(t)=m(t) « r(t) where * is the symbol for convolution. Mathematically the convolution operation is defined as, o(t)= fre r(t=1) dt see 4.2) where t_is the dummy yariable used for integration. BOUNDED - INPUT BOUNDED - OUTPUT (BIBO) STABILITY A linear relaxed system is said to have BIBO stability if every bounded (finite) input results in a bounded (finite) output. A condition for BIBO stability can be obtained from convolution operation defined by equation (4.2). For a relaxed system the equation (4.2) can be written as, Response, t= f m(z) r(t—t) dt (43) ° Note : A relaxed system is one in which the initial conditions are zero. Hence the limits. of integration is from 0 to «. If the input r() is bounded then there exists a constant A,, such that jr(1)| < A,< ©. The condition for bounded output for this bounded input condition can be derived as follows. On taking the absolute value on both sides of equation (4.3), we get, le(t)i= weld) f m(t) (t=) dr ° Since the absolute value of an integral is not greater than the integral of the absolute value of the integrand the equation (4.4) can be written as, ‘ lett] < f foots) e—ayee => foto] < f imcwint—wide] => fete < F mcay Aer 3 a 0 ‘ For bounded input, a constant | = feo] < a jm(t)) de P ; such that, |r(t—1)| s Ay. Tf the output c(t) is bounded then there exists a constant A, such that |c(#)| < A, < 2. Ay Jimco) dtc A, 44 TABLE-4.1 Transfer function, MS) and location of roots on s-plane Impulse response, m(t) Root on negative real axis mO) » marl Ala sew A (staJ t Impulse response is exponentially decaying. Stable system. M(s) m0 M(s)= off bs t Impulse response is exponentially increasing, Unstable system. A M(s)= A. Or aarp Seb 4 +bx— = 3 jb» Complex conjugate roots on left half of s-plane mt) = £3—A_ a mee statjb s+a—jp = Ae OPH a Ate in(bt +90") 2Ae™ cosbt = 2Ae™ "Impulse response is damped sinusoidal (ie, Damped oscillatory). Stable system i} t Complex conjugate roots on right half of s-plarie m(t)= 4A z S-a+ jb s-a—jb (at poe oy =Ae +A‘e =2Ac™ cosbt = 2Aet sin (bt +90") Impulse response is exponentially increasing sinusoidal] (ie, Amplitude of oscillations exponentially increases with time). Unstable system.45 A aA’ io + = s+jb s-jb M(s) = ‘Single pair of roots on imaginary axis 5's 5| s+ jb s—jb] = Ae y ate ht = 2A cos bt =2A sin (bt +90°) m(t) -« Impulse response is oscillatory Marginally stable Double pair of roots on imaginary axis =c_A_ m ole & =Ate Maat tet = 2At cos bt = 2At sin (bt + 90°) i m(f) Impulse response is linearly increasing sinusoidal Le., amplitude of oscillations linearly increases with time). Unstable sysiem. Mg)=4 3° $ o Single root at origin m4 — m(t)= offal. A pa ers t Impulse response is constant. Marginally stable system.Ci Ghapier 4 = Cancapis of Stability and Root Locus > 46 m(t) 1 A M)=5 io —%» | c < | Impulse reponse linearly increases with time. Unstable sysiem | Double root at origin 2. Ifany root of the characteristic equation has a positive real part (i.c., lying on right haif's-plane) then impulse response is unbounded, (Le., it increases to « as { tends to <). Hence E m1) dt is infinite and so system is unstable, we If the characteristic equation has repeated roots on the imaginary axis then impulse response is unbounded (i.e., it increases to «0 as t tends to «). Hence f |m(t)| dt is infinite and so the system is unstable. o 4, Ifone or more non - repeated roots of the characteristic equation are lying on the imaginary axis, then impulse response is bounded (i.e., it has constant amplitude oscillations) but is infinite and so the system-is unstable, 5. “Ifthe characteristic equation has single root at origin then the impulse response is bounded (i.e, it has constant amplitude) but j Im(t)| dt is infinite and so the system is unstable. q 6. Ifthe characteristic equation has repeated roots at origin then the impulse response is unbounded (Le., it linearly increases to infinity as t tends to 9) and so the system is unstable. 7. In system with one or more non-repeated roots on imaginary axis or with single root at origin, the output is bounded for bounded inputs except for the inputs having poles matching the system poles. These cases may be treated as acceptable or non-acceptable. Hence when the system has non repeated poles on imaginary axis or single pole at origin, it is referred as limitedly or marginally stable system. . In summary, the following three points depending on the location of roots of characteris ay be stated regarding the stability of the system ic equation. 1. [fall the roots of characteristic equation has negative real parts, then the system is stable. 2. If any root of the characteristic equation has a positive real part or if there is a repeated root on the imaginary axis then the system is unstable. 3. If the condition (j) is satisfied except for the presence of one or more non repeated roois on the imaginary axis, then the system is limitedly or marginally stable. In order to ascertain the stability of a system, it is necessary to determine if any of the roots of the characteristic equation lie in the right half s-plane. The characteristic equation is given by the denominator polynomial of closed loop transfer function, [equation (4.7)}.Consider the n* order characteristic equation shown below. ast tas says +. 1S +a Let the roots of n® order characteristic equation [equation (4.10)} bes =r, ty functions of the coefficients a, &, ayy wun @ yy Be Consider a second order polynomial, wet tingeters o (aed as] ap ag = ag (8) (5-1) = 98? ao (Fy +) S* ahh v4 11) Consider a third order polynomial aye tm stays ay(s'e Se Bar 3t) By = ay (8-4) =n) (6-5) = aap (q+ H +H)? tag (ith th) Sa thts (4.12) On extending this expansion to the n™ order polynomial, we get. ap 8" — ag(sum of all the roots) s"“! ‘sum of the products of the roots) 4/2 0 oes a mete if ap stay st tay 8 +8q (sum of the products ofthe roots) 5 0 | taken 3 at a time wre 413) + ag(-1)" (Product of all the n roots) If all the roots of ¢ polynomial are real and in the left half of s-plane, then all r, in equations (4.11) and (4.12) are real and negative. Therefore all polynomial coefficients are positive. This characteristic also applies to the general case of equation (4.13). If atleast one root is in the right half of s-plane then some of the coefficients will be negative. Also, it can be observed that if all the roots ate in the left half of s-plane, no coefficient can be zero. Since the characteristic polynomial coefficients are real, the complex roots should occur as conjugate pairs. From equation (4.13) it can be inferred that when polynomial coefficients are formed, the imaginary parts of roots/products of roots will cancel. Therefore, if all roots occur in the left half plane, (whether it is complex or real) then all coefficients of the general polynomial of equation (4.13) will be positive. Presence of a negative coefficient implies that there is atleast one root in the right half of s- plane. A zero coefficient indicates presence of complex-conjugate roots on the imaginary axis and/or one or more roots in the right half of s- plane. Insummary, following conclusions can be made about coefficients of characteristic polynomial, 1. If all the coefficients are positive and if no coefficient is zero, then all the roots are in the left half of s- plane. 2. Tf any coefficient a, Is equal to zero then, some of the roots may be on the imaginary axis or on the right half of s- plane,48 3. Many coefficient a, is negative then atleast one root is in the right half of s- plane. It can be concluded that the absence or negativeness of any of the coefficients of a characteristic Polynomial indicates that the system is either unstable or at most marginally stable. ‘Thus the necessary condition for stability of the system is that all the coefficients of its characteristic polynomial be positive. If any coefficient is zero/negative, we can immediately say that the system is unstable. In order for all the roots to have negative real parts, it is necessary that all of the coefficients of characteristic equation be positive, but it is not sufficient, because there may be roots in the right half plane and/or on the imaginary axis, even when coeflicients are positive.( ie., when roots have negative real part, then all the cocfficients of characteristic polynomial will be positive, but the reverse condition is not true always). always true Roots with negative real part. All coefficients positive. not always true Hence, when all the coefficients are positive, the system may or may not be stable, because there may be roots in the right half plane and/or on the imaginary axis. For example, consider the characteristic polynomial with all positive coefficients, Sts? +2s+8=0. ‘The characteristic polynomial can be written as, (S +5? 425+8)=(s+2) (- Now the roots are, tvs 1 =-2, +4428, ‘ “Fo 2 2 ‘The coefficients of the polynomial are all positive, but two roots have positive real part and so will lie on on right half of s-plane, therefore the system is unstable. 43 ROUTH HURWITZ CRITERION The Routh-Hurwitz stability criterion is an analytical procedure for determining whether all the roots of a polynomial have negative real part or not. The first step in analysing the stability of a system is to exantine its characteristic equation. The necessary condition for stability is that all the coefficients of the polynomial be positive. If some of the coefficients are zero or negative it can be concluded that the system is not stable. When all the coefficients are positive, the system is not necessarily stable. Eventhough the coefficient are positive, some of the roots may lie on the right half of s-plane or on the imaginary axis, In order for all the roots to have negative real parts, itis necessary but not sufficient that all coefficients of the characteristic equation be positive. If all the coefficients of the characteristic equation are positive, then the system may be stable and one should proceed further to examine the sufficient conditions of stability,49 A. Hurwitz and E.J. Routh independently published the method of investigating the sufficient conditions of stability of a system. The Hurwitz criterion is in terms of determinants and Routh criterion is in terms of array formulation. The Routh stability criterion is presented here. The Routh stability criterion is based on ordering the coefficients of the characteristic equation, into a schedule, called the Routh array as shown below. st tas"! +a,s"-2+,..¢a, S+a, =0, where ay >0 F 2 n1St By : s % & & & & x! a & Bw st Cn a a a ad cy ¢, cy c, c, s 8 S, h, The Routh stability criterion can be stated as follows. "The necessary and sufficient condition for stability is that all of the elements in the first column of the Routh array be positive. If this condition is not met, the system is unstable and the number of sign changes in the elements of the first column of the Routh array corresponds to the number of roots of the characteristic equation in the right half of the s-plane”. Note : Ifthe order of sign of first column element is +, +, = + and +, Then + to —is considered as one sign change and ~to + as another sign change. CONSTRUCTION OF ROUTH ARRAY Let the characteristic polynomial be, a2 aps" +a" | +a,s"* +a,s" 4.0, stags" The coefficients of the polynomial are arranged in two rows as shown below. St ay ap ay Agee Sy ay ag Gpece When nis even, the s* row is formed by coefficients of even order terms (i.e., coefficients of even powers of s) and s*? row is formed by coefficients of odd order terms (Le., coefficients of odd powers of s) . When n is odd, the s* row is formed by coefficients of odd order terms (i.e., coefficients of odd powers of 8) and s™! row is formed by coefficients of even order terms(i.e., coefficients of even powers offs). ‘The other rows of routh array upto s° row can be formed by the following procedure. Each row of Routh array is constructed by using the elements of previous two rows.Consider two consecutive rows of Routh array as shown below. PO ES SE By Hy SEE Yo Yn Ya Ya, V4 Yonee Let the next row be, BF sy My Ty Ey Bye The clements of s**? row are given by, (-Y iM XY —__o_Yil _ Yo%1= 9X0 Yo Yo ce [Xo Xa] * IYo _Ya|._ YoX:—Y2Xo Yo Yo x (Dp Xo Xs) Yo Ys} _ YoXs ~¥3Xo py = Ye Sad Yok = YN Yo Yo ) Xo Xs Yo YoXs —Y4Xo Yo Yo Xo A n= Ya_¥s1 _ YoRS Yeo and so on. Yo Yo The elements z,, 2,, 2,5 Z,... are computed for all possible computations as shown above. In the process of constructing Routh array the missing terms are considered as zeros. Also, all the elements of any row can be multiplied or divided by a positive constant to simplify the computational work. Case- In the construction of Routh array one may come across the following three cases. Case-] : Normal Routh array (Non-zero elements in the first column of routh array). Case-H : A row of all zeros. Case-H] : First element of a row is zero but some or other elements are not zero. Normal routh array In this case, there is no difficulty in forming routh array. The routh array can be constructed as explained above. The sign changes are noted to find the number of roots lying on the right half of s-plane and the stability of the system can be estimated. In this case, 1, If there is no sign change in the first column of Routh array then all the roots are lying on left half of s-plane and the system is stable.iL G Geateol Systems Triginwaring 2, If there is sign change in the first column of routh array, theri the system is unstable and the number of roots lying on the right half of s-plane is equal to number of sign changes. The remaining roots are lying on the left half of s-plane. Case-II row of all zeros Anall zero row indicates the existence of an even polynomial as a factor of the given characteristic equation. In an even polynomial the exponents of s are even integers or zero only. This even polynomial factor is also called auxiliary polynomial, The coefficients of the auxiliary polynomial will always be the elements of the row directly above the row of zeros in the array. The roots of an even polynomial occur in pairs that are equal in magnitude and opposite in sign. Hence, these roots can be purely imaginary, purely real or complex. The purely imaginary and purely real Toots occur in pairs. The complex roots occur in groups of four and the complex roots have quadrantal symmetry, that is the roots are symmetrical with respect to both the real and imaginary axes. The fig 4.1 shows the roots of an even polynomial * Jor jos — f c o Fig 4.1 : The roots of an even polynomial. The case-II polynomial can be analyzed by any one of the following two methods. METHOD-1 1. Determine the auxiliary polynomial, A(s) 2. Differentiate the auxiliary polynomial with respect to s, to get d A(s)/ds 3. The row of zeros is replaced with coefficients of dA(s)/ds. 4. Continue the construction of the array in the usual manner (as that of case-1 ) and the array is interpreted as follows. a, If there are sign changes in the first column of routh array then the system is unstable. The number of roots lying on right half of s-plane is equal to number of sign changes. The number of roots on imaginary axis can be estimated from the roots of auxiliary polynomial. ‘The remaining roots are lying on the left half of s-plane. b. If there are no sign changes in the first column of routh array then the all zeros row indicate the existence of purely imaginary roots and so the system is limitedly or marginally stable. The roots of auxiliary equation lies ‘on imaginary axis and the remaining roots lies on left half of s-plane. METHOD2 1. Determine the auxiliary polynomial, A(s). 2. Divide the characteristic equation by auxiliary polynomial.3. Construct Routh array using the coefficients of quotient polynomial. 4, The array is interpreted as follows. a. Ifthere are sign changes in the first column of routh array of quotient polynomial then the system is unstable. The number of roots of quotient polynomial lying on right half of s- plane is given by number of sign changes in first column of routh array. The roots of auxiliary polynomial are directly calculated to find whether they are purely imaginary or purely real or complex. The total number of roots on right half of s-plane is given by the sum of number of sign changes and the number of roots of auxiliary polynomial with positive real part. The number of roots on imaginary axis can be estimated from the roots of auxiliary polynomial. The remaining roots are lying on the left half of s-plane. If there is no sign change in the first column of routh array of quotient polynomial then the system is limitedly or marginally stable. Since there is no sign change all the roots of quotient polynomial are lying on the left half of s-plane. The roots of auxiliary polynomial are directly calculated to find whether they are purely imaginary or purely real or complex. The number of roots lying on imaginary axis and on the right half of s-plane can be estimated from the roots of auxiliary polynomial. The remaining roots are lying on the left half of s-plane. Case-III : First element of a row is zero While constructing routh array, if a zero is encountered as first element of a row then all the elements of the next row will be infinite. To overcome this problem let 0—>< and complete the construction of array in the usual way (as that of case-I ) Finally Jet ¢>0 and determine the values of the elements of the array which are functions of €. ‘The resultant array is interpreted as follows. Note + If all the elements of a row are zeros then the solution is attempted by considering the |polynomial as case-if polynomial. Even if there is a single element zero on s! row, it is considered as a row of all zeros. a. If there is no sign change in first column of routh array and if there is no row with all zeros, then alll the roots are lying on left half of s-plane and tire system is stable. If there are sign changes in first column-of routh array and there is no row with all zeros, then some of the roots are lying on the right half of s-plane and the system is unstable. The number of foots lying on the right half of s-plane is equal to number of sign changes and the remaining roots are lying on the left half of s-plane. If there is a row of all zeros after letting €—0, then there is a possibility of roots on imaginary axis, Determine the auxiliary polynomial and divide the characteristic equation by auxiliary polynomial to eliminate the imaginary roots. The routh array is constructed using the coefficients of quotient polynomial and the characteristic equation is interpreted as explained in method-2 of case-II polynomial.4.13 Conard Systems Engineering) EXAMPLE 4.1 Using Routh criterion, determine the stability of the system represented by the characteristic equation, s*#8s?+ 1852 +165 +5=0. Comment on the location of the roots of characteristic equation SOLUTION ‘The characteristic equation of the systemis, s*+8s°+18s?+16s+5=0, ‘The given characteristic equation is 4” order equation and soit has 4 roots. Since the highest power ofs is even number, form the firstrow of routh array using the coefficients of even powers of s and fom the second row using the coefficients ofodd Powers of s. st 118 5 Rowt 3 ar s 8 6 ns Row2 | : | ‘The elements of s° row can be divided by 8 to simplify the oe 5 ae ros gi; 19x2-554 s i Rowt Te ty é 1.6875 = 1.7 @o5 Jas ~-Row2 Bia lhe 8 ROWS 1.7x5-0x16 s! riz o-- Rows © ae 25 st 15 ... Row-5 LAG —~Column-1 ‘On examining the elements of first column of routh array itis observed that all the elements are positive and there is no ‘sign change. Hence all the roots are lying on the left half of s-plane and the system is stabie. RESULT 1. Stablesystem 2. Allthe fourroots are lying on the left haif of s-plano. EXAMPLE 4.2 ~~ ConstructRoutheray an determine the stability of he system whose characterisite equation is s*+2s°+8s'+125%20s? +168+16=0. Also determine the number of roots lying on righthalf of plane, lefthalf of s-plane and on imaginary axis, SOLUTION ‘The characteristic equation of the systemis, s*+2s'+8s'+12s*+ 20s+16s +16 =0, The given characteristic polynomial is 6® order equation and so it has 6 roots. Since the highest power of sis even ‘number, form the first row of routh array using the coefficients of even powers ofs and form the second row using the coefficients of odd powers of, fs: 1 8 2 £6 Row-1 es: 2 2 46 so Rowe The elements of s* row can be divided by 2 to simplify the caiculations.4.4 8 : 5 gf , XB= 6x1 1x20-8x1 116-001 ! 7 7 i 16 she 2 12 16 ‘stole divide by 2 . a 2, etat3 in 4 6 8 tat ee, LO=OxT 1x8 8x7 Hag TF 7 st: 1033) sO 0 The auxiliary equation is, A= s4+6s%8. On bgt ue es me OT differentiating A with respect to s we get, —-Column-1 GA, Bina (Onexamining the elements of 1 column of outh array itis observed ——=4s'+128 that theres no sign change. The rowwith al zeros indicate the possibility ss ofrocts on imaginary axis. Hence the system is limitedly or marginally JA stable The coefficients of “—~ are used toform s*row. The awiliary polynomials, s:4 12 s!+652+8=0 divide by 4 Let, =x Sy T6-Sx1 128-01 coe +6x48=0 fo oe 62 V6? -4x8 3 8 The roots of quadratic are, X= —————— 3x3_8x4 3 : s': 033 ‘The roots of auxiliary polynomials, ye 0358-073 | ° 033 s=avx =4V-2 and 2/4 eg =+iv2.- V2+ 2 and —j2 The roots of auxilary polynomial are also roots of characteristic equation. Hence 4 roots are lying on imaginary axis and the remaining two roots are lying onthe left half of »-plane RESULT 1. The systemis imitedly or marginally stable. 2. Fourroots are ying on imaginary axis and remaining two roots ae tying onlefthalf ofs-plane. EXAMPLE 4.3 Construct Routh array and determine the stability of the system represented by the characteristic equation, s'+s'+2s°+2s?+3s+5=0. Commenton the location of the roots of characteristic equation. SOLUTION ‘The characteristic equation of the systems, s*+s"+2s*+2s*+3s+5=0. The given characteristic polynomial is 5* order equation and so it has 5 roots. Since the highest power of sis odd number, form the firstrow of routh array using the coefficients of odd powers of s and form the second raw using the coefficients of even powers ofs. gs: 42 3 Rowe so 1 2 5 ...Row2415 so: < 2 se, Row-3 s 2642 5 Row gy de2a2at 13-5x1 € 4 7 ; (32 4244) ss 0 -2 si; Se oa Row Replace 0 by € 2642 a % ° 5 Row Onletting <0, we get - e221) exS=0x4 sip 1) 2 3 ..Rowt eb € to ex sir, 4,2 5 ..Row2 aa 5. 8:1 012 ..Row-3 2e+2 4 1 x (-2)-(S% ¢) Row = Sethe 4| Hastat Row-5 gare € ' goits! an Rw-Bt +444) See A eh 2e42 On observing the elements of first colurnn of outh array, itis found that there are two sign changes. Hence two roots are lying on the right half of s- plane and the system is unstable. The remaining three roats are lying onthe left hatfofs-plane. RESULT (@). The systemis unstable. (b). Two roots are lying on right half of s-plane and three roots are lying on left halfaf s-piane. EXAMPLE 4.4 , By routh stability criterion determine the stability of the system represented by the characteristic equation, 95°- 20s'+10s*_s? 9s -10 = 0. Comment on the location of roots of characteristic equation. SOLUTION The characteristic polynomial of the systems, 95°-20s*+10s*-s*-Gs-10=0 ‘On examining the coefficients of the characteristic polynomial, itis found that some of the coefficients are negative and ‘80 some roots will lie on the right half of s-plane. Hence the system is unstable. The routh array can be constructedto find the number of roots lying on righthatf of s-plane. The given characteristic polynomial is 5 order equation and so it has 5 roots. Since the highest power of s is odd number, form the first row of routh array using the coefficients of odd powers of s and form the second row using the coefficients, of even powers ofs. s ils a | 20 2 | = i +s 2 = 10 Row-2 | : i 1 A in s955 -135 1 ‘Yilst 1955 1 135 ne ROWS Ql 1 Ns: 12931 1 1 1 1 10) x9 19-9... Rowe HMOIOS (EDAD) =20% CO) SC10) x9, sto: 1-168 s ~10 Rows | (2.955 x(-9-(13.9) (20) 955 x(-10) Row Rows | ae O55 O55 16 - “= Column-14.16 By examining the elements of I* column of routh array itis observed that there are three sign changes and so three roots are lying onthe right haifof s-plane and the remaining two rootsare lying on the lefthaifofs-piane. RESULT (@). The systemis unstable. (©). Three roots are fying on right half of s-plane and two roots are lying on left half of s-piane. EXAMPLE 4.5 ‘The characteristic polynomial of a systomis, s’+8s%+24s'+24s'+24s"+245%23s+15=0, Determine the location ofroots on s-plane and hence the stability of the system, SOLUTION METHOD-I ‘The characteristic equation is, s’+9s°+24s+24s"+24s4+24s%235+15=0. The given cheracteristic polynomial is:7" order equation and so it has 7 roots. Since the highest power of s is odd number, form the first row of array using the coefficients of odd powers of s and form the second row using the coefficients of even powers of s as shown below. shar A at Be. 28) towel 3x24-8x1 3x24-8x1 3x23-5x1 so: 9 2h 2h 15... Row? 3 Divide s* row by 3to simplify the computations. oe - as 2133 eae Divide by 21. Soi 712 23 Row nia Lot se: 4 1 1 sor, 3,8 8 8 ..Row2 et gt eon q : LA A ROMS gt, PRBa1K3 1xB-1x3 145-03 st yyaod sn Row 1 1 7 a4 . 8 21040 ea Rowe se 5 5 5 tot Divide by 5 Sonar sn Row-5 484 4 1 oly 821051 4 Row - 5. de 13t sn Row7 a a 1 sosteqe a ROW a “* column ‘On examining the first column elements of routh arrayitis | The auxiliary polynomials, found that there are two sign changes. Hence two roots are lying on Azsttstt the right halfof s-plane and so the system is unstable. Differentiate A with respect tos. ‘The ow ofallzeros indicates the possibilty ofroots on imaginary axis. This can be tested by evaluating the roots of auxiliary polynomial. a 4g 428 is The auxiliary equation is, s*+s?+1=0 e142 Put, s*=xin the auniliary equation, Divide by 2 sttstttaxtext1=0 s:2 4417 = (Gonna ystems Engineering) ap 2xt-txt 2x1-0%1 2 =12120° or 12-120" x astavxssVi2P or. tviZ=120° savizi2072 or eviz—120°2 #1260" or 212-60° = +05 +0866) or +(0.5~j0866) The roots of quadratic are, x But s? ‘Two roots of auxiliary polynomial are lying on the right half ofs-plane and the remaining two on the left hatf of s-plane. The roots of auxiliary equation are also the roots of characteristic polynomial. The two roots lying on the right half of s-plane are indicated by two sign changes in the first column of routh array. The remaining five roots are lying on the lefthaif ofs-plane. No roots are lying on imaginary axis. RESULT i 4. Thesystem is unstable. 2. Tworoots are lying on right hatf of s-plane and five roots are lying on left haifofs-plane.. METHOD-II ‘The characteristic equation is, 8’ + 9s° + 248° + 24s‘ + 248° + 248” + 235+15=0 ‘The given characteristic polynomial is 7* order equationand so . ithas 7 roots, Since the highest power ofsisodd number, fomnthefirst [_, 3x24-8x1 Gu24—-Bx1 Ox Zd—5n1 row of array using the coefficients of odd powers of sand formthe | S ° ———~3-—~ 3-3 send row using the coefficients of even powers of s as shown | os. 94 45 2193-2188 om Divide by 21.33 9 124 2h 23... Row ss 4 1 1 Sob 8 M15 Rowe gh BS THOS HBOS ; 7 7 7 Divide s*row by 3t simply the computations eke: 5 a So: 1 Mh 2 BB ..Row-t Divide by 5 fo: 3 8 8 5 ...Row2 sti 1 4 Txt ep ot 4 = Row-3 sa es tp 4A o=-Row-4 o et 0 0 w+ Row-5 Since we get.a row of zeros, there exists an even polynomial, the even polynomial is nothing but, the auxiliary polynomial. ‘The auxiliary polynomialis, st+st+1=0 Divide the characteristic equation by auxiliary polynomial to get the quotient polynomial. The characteristic polynomial be expressed as a product of quotient polynomial and auxiliary polynomial.Cichapher T= concep of Siability and Root Locus > 4.18 Qs° + 245% + 24s‘ + 24s? + 24s? +238 +15=0 L (s* +8? +1) (s* + 987+ 233 +15)=0 Even Quotient polynomial polynomial 2 i |v aj: & 28 18 /s gue The routh array is constructed for quotient polynomial as shown + Vgis below. < iy ee 2438 3% s 1 2B 7 #8 o|8 3x23-5x1 @ ty = s 1 SxB—5x1 2 “aly 9 15 of: ES a 8 a\z 21.33 Yo ala, Fe 123 arg ig 15 o. 21.33x5-0x3 le 1 . 2133 7 |8 1 ' 1 =+.Columnet ee The elements of column-1 of quotient polynomial are all positive and there is no sign change. Hence all the roots of quotient polynomial are lying on the left half of s-plane. To determine the stability, the roots of auxiliary polynomial should be evaluated. ‘The auxiliary equationis, s+s?+1=0. Put, s?=xintheauxiliaryequation, —_s*+s*#1=x'+x+1=0 ‘i ae viH4 1 v8 The roots of quadraticare, x= SSF) = 12120" or 12-120" Buts?=x .ssivx = +VIZi20° or + ViZ-120° 2 Jizz or s Vi 2-120°2 or 412-60" or =(05 - j0.866) “The roots of auxiliry equation are complex and has quadrantal symmetry. Two roots of auxiliary equation are lying on the righthalfofs-plane and the other two on the left half of s-plane. . ‘The roots of characteristic equation are given by the roots of auxiliary polynomial and the roots of quotient polynomial. Hence we can conclude that two roots of characteristic equation are lying on the right half of s-plane and so system is unstable. The remaining five roots are lying on left haif of s-plane. EXAMPLE 4.6 “The characteristic polynomial ofa systemis s’ + 5s° +9s° + Qs‘ +4s° + 20s* + 36s +36 = 0°. Determine the location of roots on thes-plane and hence the stability of the system. SOLUTION ‘The characteristic equation is, 5’ +5s° ~ ¥s° +9s‘+4s° + 20s? + 36s + 36=0. ‘The given characteristic polynomial is 7° order equation and so it has 7 roots. Since the highest power of sis odd. number, form the first row of array using the coefficients cf odd powers of s and form the second row of array using the coefficients of even powers of sas shown below419 5. 1x9-18x1 1x. ra og 4 36, we RoW-1 s 1 sf: 5 9 20 36 ...Row2 s°:7.2 0 Divide s® row by 5 to simplify the computations. Divide by 7.2 Fo: 1 9 4 36. Rowt se: 0 4 ss: t 184 7. wee Re 2 bow 1x 18-0x1 1x4-41 1x72 S21 0 4 ROWS a rn si: 10 4 ww Row jst: 18 0 72 Ss: 0 0 Row-5 Divide by 1.8 sd 0 4 The row of all zeros indicate the existence of even polynomial, Which is also the aunxiary polynomial. The auxiliary polynomialis, s!+4 0. Divide the characteristic equation by auxiliary equation to get the quotient 3: 4x O-Ox1 1x4-4x1 Polynomial 7 1 ‘The characteristic equation can be expressed as a product of |S: 0 o ‘quotient polynomial and auxiliary equation. ‘ \8 le | : lg 2 |8 eg 2.81 458° +955 +95! 445° +208" + 365+36=0 (g & gs 2 8] (+4) (s3+5s? +98+9)=-0 we 8) Even Quotient polynomial a al polynomial (= [8 S| I [wm 1+ le ‘The routh array is constructed for quotient polynomial as shown a EP | bolow. yh ts. a st; Se8=0e4 el? 8 [8 tee s 1419 tlm i lt. | to ss [e |e sl | s he to +f IS ° B® gy Wel als [oe “ale Pig oe 2 * E+ Column = ‘™ ‘There is no sign change in the elements of first column of routh array of quotient polynomial, Hence all the roots of quotient polynomial are lying on the left half of plane. ‘To determine the stability, the roots of auxiliary polynomial should be evaluated. The auxiliary polynomialis, s+4=0, Put, s?= xin the auxiliary equation, «. st#4=x¢ +4 = 0 : wets A = x st¥ 4 =4J2= 290° or 22-90" But, ssivx = +v2Z90° or + V2Z—80° V2290°2 or +2 2-902 sav2cae or 222-45 = £(1+j) oF #(1-j) ‘The roots of auxiary equation are complex and has quadrantal symmetry. Two roots of auxiliary equation are ly tthalf of s-plane and the other two on the left half ofs-plane. ‘The roots of characteristic equation are given by roots of quotient polynomial and auxiliary polynomial. Hence we can Conclude that two roots of characteristic equation are lying on the right half of s-plane and so the system is unstable. The remaining five roots are lying on the left half of s-plane,RESULT @ Thesystemis unstable. (0)_Two roots are tying on the right half of splane and five roots are lying on the loft half of s-plane. EXAMPLE 4.7 Uso the routh stability criterion to determine the location of roots on the s-plane and hence the stability for the system ‘Tepresented by the characteristic equation s* + 4s‘ + 8s° + 8s*+7s+4=0. SOLUTION The characteristic equation of the systemis, s°+ 4s‘ + 8s*+8s?+ 7s +4 ‘The given characteristic polynomial is 5* order equation and soit has 5 roots. Since the highest power of s is odd ‘number, form the first row of routh array using the coefficients of odd powers of s and form the second row using the coefficients of even powers of. es: 1 87 Row-t as dedtxh 1 si: 4 8 4 Row-2, s 6 Divide s¢ row by 4 to simplify the calculations. Divide by § sit 18 7 Row-1 3:4 1 tot sort 12 4 a= ROW ge te2rtxt txt-Ox1 Soutqt Row a j 1 eta Row eteted tot —_ - tet Row-5 1 it s':0 6 t eer RaW Let 0+ —Column-1 sie When © => 0, there is no sign change in the first column of t routh array. But we have @ row of all zoros(s" row or row-5) and so there isa possibilty of roots on imaginary axis. This can be found from the roots of auxiliary polynomial. Here the auxiliary polynomial isgiven bystrow. -lors aN-1=2i1 ‘The roots of auxiliary polynomial are +/1and--j1, lying on imaginary axis. The roots of auxiliary polynomial are also roots of characteristic equation. Hence two roots of characteristic equation are lying on imaginary axis and so the system sllimitedly or marginally stable. The remaining three roots of characteristic equation are lying on the left half of s-plane. RESULT (©) Thesystemis|imitedly or marginally stable. (0). Two roots are lying on imaginary axis and three roots are lying on left half ofs-plane, EXAMPLE 4.8 Use the routh stability criterion to determine the location of rocts on the s-plane and hence the stability for the system ropresented by the characteristic equation, s*+ s*+3s'+ 3s"+ 3s"+2s +1 = 0. The auxiliary polynomialis, s?+1 =0; SOLUTION ‘The characteristic polynomial of the systemis, s*+ s*+3s‘+ 3s*+ 3s* +2s +1 =0.4.21 ‘The given characteristic polynomial is 6" order equation and so ithas 6 roots. Since the highest power of sis even number, form the firstrow of routh array using the coefficients of even powers of s and form the second row using ¥e3—3x1 1x3-2x1 1x1-0%1 the coefficients of odd powers of s as shown below, 1 1 1 4, eee 303 1 ..Rowd ead) i ‘4 let Oe eoro4 3 2 ne Rowe? ath ve 1 1 foie toi on ROWS £3=1s1 €x € € s R oo, 3ecl 2et Bel 2e-1 3e-1 2st, 10x ? oe ROWS Sg Pe deat € SS 2F ae = Foire 1 eo: od eos ROWT, Ontetting <> 0, we get, s 1 3 301 ..Row-t r2e Hel, 2e-1_ Bet, Beni < € on Re s 1 32 1oW-2 anes s 0 4 = ROW-3 3e-t eo “g _Row4 s (22 4-1 (25-9-(3e-1) (3e-D y et) +4): 1 1 ... Row-5 Yossi: 0 ~- RowS “ete ae) 2244 evo: 4 » ROW-7 Since there is a row ofall zeros (s? row) theres a possibility of age roots on imaginary axis. The auxiliary polynomial is s? + 1 = 0. o, 42 tet (42-942 ter) i The roats of auxiliary polynomial are, $= 4¥—-1=2j1 s The roots of auxiliary polynomial are also roots of characteristic equation. Hence two roots are lying on imaginary axis. ‘Therefore divide the characteristic polynomial by auxiliary equation and construct the routh array for quotient polynomial tofind the roots lying on right haff of plane. ‘The characteristic polynomial can be expressed as a product of auxiliary polynomial and quotient polynomial. 8°+3s43s+3s?+2s+1=0-> (57+) (s¢+5°4+25?+25+1)=0 1x2-2x1 1x1-0x1 Even Cuotent pomomiat 1 Tie routh array for quotient polynomial is constructed as shown below. s z 1 2 1 s 1 2 2 e 1 s! cl RowOn letting <> 0,we get 5 eo FTh 2 4 cow Ses Det s 251 poke : i) § Rowe |s?+1[s® 45° 43st 43s? +35? +2541 1112 -Row-2 it Boye ge or ve Rowe S428) 359438" +2841 ys lw! Row-4 oS eth | 2s* +2s° +387 +2541 2 tad 7 5 a8 Lid Row5 : 2 east —Column-1 = 28% +374 2841 (On examining the first column of the routh array of quotient 2s? e428 polynomial, we found that there are two sign changes. Hence two | : oe Ge roots are lying on the right half of s-plane and other two roots of " ‘quotient polynomial are lying on the left half ofs-piane. a sod 0 The roots of characteristic equation are given by roots of auxiliary polynomial and quotient polynomial. Hence two roots are lying ‘on imaginary axis, two roots are lying on righthalfof s-plane and the Temaining two roots are lying on left half ofs-piene. Hence the system isunstable. RESULT (@) Thesystemis unstable, (©) Two roots are lying on imaginary axis, two roots are lying on right half of plane and two roots are {ying on lefthalf of s-plane. EXAMPLE 4.9 ; Determine the range of K for stability of unity feedback system whose open loop transfer function is K age K _-. = ea erD SOLUTION K Os) __GS)__ (#1) (+2) K The closed loop transfer function, —=~ RG) 7 G() te K 3E* 6D IK sis+1) (+2) The characteristic equations, 8 (s+ 1) (s+2)+K=0 .8(8?+3842)+K= a $°+3s*+25+K=0 ‘The routh array is constructed as shown below, ‘The highest power of in the characteristic polynomial is odd number. Hence form the first row using the coefficients of odd powers of S and form the second row using the coefficients of even powers of. Seg. PT | toa Kxt s 13 9K a = iy 16-Kt ES a 3 \ { xK-0x3 - LK @-0rs | —Column-1 | ck423 TT ae Peghncoring) Forthe system to be stable there should not be any sign change in the elements of frst column, Henice choose the value of Kso that the first column elements are postive. From s* row, forthe system to be stable, K> 0 6-K From s* ow, forthe system to be stable, ——>0 For ox >0, the value of K should be less than 6 ~The range of K forthe system to be stable is 00. ‘Since (686.25-K) > 0, should be less than 666.25. From s° row, for the system to be stable, (200+K) >0 Since (200+K) > 0, K should be greater than -200, but practical values of K starts from 0. Hence K should be greater then ©The range of K for the system to be stable is 0 of Kand aso that he system oscillates ata frequency of 2radisec. SOLUTION K(s+1) ie ees td, £8) __G)_._ Sy as" e2554 Kis+1) transfer function] R(s) ~ 1 G(s) $1) Spas? 525514 KG FT) +2541 The characteristic equation is, s* +as? +2s+1+K(s+1)=0. Sttas’+2s+1+Ks+K=0 =>. shtas?+(2+K)s +1+K=0 ‘The routh array of characteristic polynomial is constructed as shown below, The maximum power of sis odd, hence the first row of routh array is formed using coefficients of odd powers of s and the second row of routh array is formed using coefficients ofeven powers of. lfthe elements of s' row are all zeros then there exists an even polynomial (arauxiliary polynomial). f the roots of the auxiliary polynomial are purelyimaginary then the roots are lying on imaginary axis and the system oscillates. The frequency of oscillation is the root of auxiliary polynomial,4.25 ‘Bouth array’ s 1 2K se a 1+K gi. 222K) - (4K) s: HK From s* row, the auxiliary polynomial is, as?+(1+K)=0 > = + Given that, s= 222. E-2 = us = From s" row, SW) 2 => aQ+K)-(1+K)=0 => 2a+Ka 2a-1+ K(a-1)=0 Put, K = 4a 1 “2a—-1+(4e-t)fa-1)=0 => 2a-1+4a?—da-at1=0 => 4a?-3a=0 (or) a(4a-3)=0 Since 840, 4a-3=0, ass Whena=(3/4), K=da-1=4x(3/4)-1=2 RESULT When the system oscillates ata frequency of2radisec, K=2 and a= 3/4. EXAMPLE 4.12 ‘A feedback system has open loop transfer function of Gis) = wes . Determine the maximum value of K for stability of closed loop system. SOLUTION Generally control systems have very low bandwidth which implies thetithas very low frequency range of operation. Hence for iow frequency ranges the term e*can be replaced by, 1-s, (L¢., e*%.1-sT). ___ Ket. K(1-s) *GS)= Fe 5520) ~ He? + 8548) K(t-s) The closed loop| C(s)__ G(s) _ s(s? +5s+9) K(i-s) sanster uncon} Rs) 1+G(s) 4, _KC ‘s{s* +58+9)+K(1~s) 3s? + 58+9) “The characteristic equation is given by the denominator polynomial of closed loop transfer function. 5 The characteristic equationis, s (s+5s+9)+K(t-s)=0 =. s(g¢+3873} +K(1- 8) =s*+5st+9s+K-KS=0 => 81+ 5+ (0-K)s+K=0 The routh array of characteristic polynomial is corstructed as shown below.iia A ZAPATA] HABER ad Rook Tos > 4.26 ‘The maximum power of sin the charactetistic polynomial is odd, hence form the firstrow of routh array using coefficients, -af odd powers of s and second row of routh array using coefficients of even powers of s. so: 1 9-K eo: 5 K s' oo: 9-4.2K eo: K From s' row, for stability of thesystem, (9-1.2K)>0 He @-1.2K)>0 then 1.2K<9 ; Ke g=75 From s? row, for stability of the system, K>0 Finally we can conclude that for stability ofthe system K should be in the range of 0 > 7 > U4 Fig! QS? re? SS AS gs I 1 Encircled : A point is said 10 be encircled by a closed path if it is found inside the path, With reference to fig 1, the point A is encircied in the clockwise direction and the point B is not encircled. | Enclosed : Any point or region is said to be enclosed by a closed path, ifit is found to lie to the | right of the path when the path is traversed in the prescribed direction. The shaded regions infig 2 and 3 are the regions enclosed by the closed path. With reference to fig 2, the point A is enclosed by closed path and the point B is not enclosed. With reference to fig 3 the point A is not enclosed by closed path but point B is enclosed. TABLE-4.2 The function, F(s) and s-plane contour F()-plane contour jog Fls) = sh2ex6 = (s-1-2)(6-1+/2) Zers:2, = 142, 2, = 142 F(s)-plane+12, Hf | “}— (=F re: pa242, p22 S-plane iv. F(s)-plane iva F(s)-plane D, Fis)= s?-2s+6_ (s-1-]2)(s-1+)2) s2 s-plane ive Fis)-plane ate Poles : pj=2+i2, pj=2-/2 Zeros : z= 14/2, z,=1-/2 ive F(s)-plane OG. s-2 FS) = Braye) Poles : p, = 1,0,=3 Zeros :z,=2 We F(s)-plane 1. The summary of relationship between the enclosure of poles and zeros by the s-plane closed contour and number of encirclements of the origin of F(s)-plane by the corresponding F(s)-plane contour, are given below.429 1. If s-plane closed contour encloses Z number of zeros in the right half of s-plane then the corresponding contour in F(s)-plane will encircle, the origin of F(s)-plane Z times in the clockwise direction. 2. If s-plane closed: contour encloses P number of poles in the right half of s-plane then the corresponding contour in F(s)-plane will encircle the origin of F(s)- plane P times in anticlockwise direction. : 3, If the s-plane closed contour encloses Z zeros and P poles in the right half of s-plane and if P > Z, then the corresponding contour in F(s)-plane will encircle the origin of F(s)-plane (P-Z) times in the anti-clockwise direction. 4, If the s-plane closed contour encloses Z zeros and P poles in the right half of s-plane and if P > oy é; os o oma a. Nyquist Contour when b. Nyquist Comour when _¢, Nyquist Contour when there is no pole on there are poles at origin there are poles on imaginary imaginary acis aocis and at origin Fig 4.5 : Nyquist Contour 2. The Nyquist contour should be mapped in the G(s)H(s)-plane using the function G(s)H(6) to determine the encirclement -1+ j0 point in the G(s)H(s)-plane. The Nyquist contour of fig 4.5b can be divided into four sections C,, C,, C, and C,. The mapping of the four sections in the G(S)EX(S)-plane can be carried sectionwise and then combined together to get entire G(s)H(S)- contour. ~ 3. In section C, the value of @ varies from 0 to + «0. The mapping of section C, is obtained by letting s = jo in G(s)H(G) and varying o from 0 to +, i.e, G(S)H(S) =G(jo) HG) [ 0m. The locus of GGw)HGe) as o is varied from 0 to +0 will be the G(s)EH(s)-contour in G(s)H(s)-plane corresponding to section C, in s-plane. This locus is the polar plot of G(jo)HGo). There are three ways of mapping this section of G(s)H(s)-contour, they are, (i) Calculate the values of G(ja)HGo) for various values of @ and sketch the actual locus of GGa)HGe). (or) (i) Separate the real part and imaginary part of G(jo)HGo). Equate the imaginary part to zero, to find the frequency at which the G(j@)HGa) locus crosses real axis (to find phase crossover frequency). Substitute this frequency on real part and find the crossing point of the locus on real axis. Sketch the approximate locus of G(j@)H(jo) from the knowledge of type number and order of the system (or from the value of Gda)H(o) at @ = 0 and @ = 0). (iii) Separate the magnitude and phase of G(jo)H(ja). Equate the phase of G(j@)HGa) to - 180° and solve for «. This value of « is the phase crossover frequency and the magnitude at this frequency is the crossing point on real axis. Sketch the approximate root locus as mentioned in method (ii). 4. The section C, of Nyquist contour has a semicircle of infinite radius. Therefore, every point on section C, has infinite magnitude but the argument varies from +7/2 to ~w/2, Hence the mapping ofsection C, from s-plane to G(s)H(s) plane can be obtained. “ letting s= Lt Re! in G(s)H(s) and varying @ from +2/2 to —n/2. Consider the loop transfer function in time constant form and with y number of poles at origin; as shown below.4.34 ctgety - S048 45h) 04h & (1+sT,) (+sh,) tsh).. Let G(s)H(s) has m zeros & n poles including poles at origin. For practical systems, n >m, Since, sRe® and R->«, the term (1+8T) can be approximated to sT, [i.e., (1+8T) = sT]. T, xsTy xs s"_ kK, © G()H(s) # KS See FS HC) 9 xsT, XSF, xsT, tyes On letting, salt Re® we get, kK, )H(8) -— A __ = 0 Pm “ Lt (Re*)™ 5 rote nn BM When 0 = >: G{s)H(s) = 002 JF om) When 9=-5, G(@)H(s)= oe" From the above two equations we can conclude that the section C, of Nyquist contour in seplane is mapped as circles/circular are around origin with radius tending to zero in the G@H(6)-plane. In section C,, the value of @ varies from ~c to 0. The mapping of section C, is obtained by letting s = +jo in G(s)H(6)-and varying o from -20 to 0. Le, G(S)H(s) = G(jo)HGo) io lo==t00. 2100 The locus of G(ja)H(a) as @ is varied from — to 0 will be the G(s)H(s)-contour in G(S)H(6)-plane corresponding to section C, in s-plane. This locus is the inverse polar plot of G(jo)HGa). The inverse polar plot is given by the mirror image of polar plot with respect to real axis. . The section C, of Nyquist contour has a semicircle of zero radius..Therefore every point on semicircle has zero magnitude but the argument varies from ~1/2 to +7/2. Hence the mapping of section C, from s-plane to G(s)H(S)-plane can be obtained by letting s= Lt, Rein G(S)H(S) and varying 0 from -1/2 to #x/2. Consider the loop transfer function in time constant form and with y number of poles at origin as shown below. K(+sh) (+sh) d+sh) 2 tsk) (+sh) (+sh) G{s)H(s) = Let G(S)H(S) has m zeros & n poles including poles at origin. For practical systems, n >m. Since, s->Re® and R00, the term 1+ sT can be approximated to 1, [i-e(1+sT) =1].435 2-G(QH@) *K 3 On letting, so Lt, Re® we get, ke nee GHG) [s— Lt. ie ata Be When 6=-5, G(S)H(6) = xe?” When d=, - G(@)Hs)= 02” From the above two equations we can conclude that the section C, of Nyquist contour in s-plane is mapped as circles/eircular arc in G(s)H(s)-plane with origin as centre and infinite radius: Note: 1. If there are no poles on the origin then the section C, of Nyquist contour will be absent. 2. Ifthere are poles on imaginary axis as shown below then the Nyquist contour is divided into thd following 8 sections and the mapping is performed sectionwise. Section C, sO=0 tt+o- / SectionC, = Lt Re®;0=—St0+ 5 mo 2° 2 Section, > s=ja;a=+0, 10+ 2 Section C, SectionC, > s=j0;0=-2t0-o, SectionC, : s= Lt Re®;@=—2i0+% RO Z 2 Section C, : s=ja;@=-0, to0 RoR SectionC, + 82}, Re8:0=-Tio~> EXAMPLE 4.13 i . = K Draw te Nyeuistpletfor the systom whose openlopiransterfunctonis, sHis) = ay Determine the range of K for which closed loop system is stable. SOLUTION K 005K Given that, SHO START Ta(8 S4}x10( 1) 5 (105s) (1+ 0.45) The open loop transfer function has a poie at origin. Hence choose the Nyquist contour on s-plane enclosing the entire right half piane except the origin as shown in fig 4.13.1.4.36 ‘The Nyquist contour has four sections C,, C,, C, and C,. The mapping of each section is performed separately and the overall Nyquist plotis obtained by combining the individual sections. MAPPING OF SECTION C, Insection C, ,@ varies from 0 to 420. The mapping of section C, is given by the locus of G(ja)H{ja) a8 « is varied from 0 to =. This locus is the polar plot of Go)HGo). Fig 4.13.1 : Nyquist Contour in s-plane 2 008K GUSH) = arp COTS Lets=jo. - 0.08K +. G{jo)H(o) =~ at . Be i jo (1430.50) (1+ 0.40) jo (1+ j0.60 00507) —0.60? + jo (10.0507) When the locus of G(fo} (jo) crosses real axis the imaginary term will be zero and the corresponding frequency isthe phase crossover frequency, 2 i JMO = Ope, Opel O0505.)=0 > 1-00523,-0 = On = aie = 4472 red/s00 05K 05K 080? 0.6 x (4.472)? ‘The open loop systems type-1 and third order system. Also tis a minimum phase system with allpoles. Hence the polar plot of G(ja)!4(n) starts et-90° axds at infinity, crosses real axis at-0,00417K and ends at origin in second quadrant. The section C, and its mapping are shown in fig 4.13.2. and 4.13.3. j v io ofl land [04 1(s)-plane lomo Ato =0,, = 4.472 rad/ sec, ‘G(jo)HYjo) = =~0.00417K -0.00417K od Fig 4.13.2 : Section C, in s-plane Fig 4.13.3 : Mapping of section C, in G(s)H(s)-plane MAPPING OF SECTION G, The mapping of section C, from splane to Gi(s)H(¢}-planeis obtained by letting s= Lt Re InG(s)H(s)and varying 8 from +22 to-2/2. Since s-»R ? and Rs, the G(s)H(s) can be approximated as shown below, [i.e., (1+sT)=sT]. 0.05K 005K K GO) H) =F Gx05e) (F01s) ~ Sx05s o=] o i -0.00417K | ono . Fig 4.13.7 : Mapping of section C, in Fig 4.13.6 : Section C, in s-plane GHGS) plane MAPPING OF SECTION C, ‘The mapping of section G, froms-plane to G(s)H(s}-plane is obtained by letting $= Lt, Re” in G(s)His) and varying ® from —w/2to +n/2. Since sof eF and R-> 0, the G(s) H(s) can be approximated as shown below, (ie, (1#8T)= 1} 0.05K = 005K _ 005K s(1#0.5s)(1*0.1s)sxtxt s G(s)H(s) = Lets= Lt Re®. Roo = GfS)H(s) When 0 = 3 GIs)H(s) =~00 72 eat (4) From the equations (3) and (4) we can say that section C, in-plane (fig 4.13.8.) is mapped as a circular arc of infinite radius with argument (phase) varying from +n/2 to ~n/2 as shown in fig 4.13.9.4.38 Gis)H(s}-piane v Fig 4.13.8 : Section C, in s-plane Fig 4.13.9 : Mapping of section C, in COMPLETE NYQUIST PLOT Riyals ‘The entire Nyquist plotin G{s)H(s)-plane can be obtained by combining the mappings of individual sections, as shown Infig 4.13.10. i G{s)H(s}-plane G(s)H(s)-contour Fig4.13.10: Nyquist plot of K St Ete (s+ 2\(s-+ 10) STABILITY ANALYSIS ‘When, -0.00417K =—1, the contour passes through (—1+0) point and corresponding value of Kis the limiting value of Kforstability. 5‘ <. Limiting value of K = imiting value TnOaT = 240 When K < 240 ‘When Kis less than 240, the contour crosses real axis ata point between 0 and —1+)0. On travelling through Nyquistplot along the indicated direction itis found that the point —1+)0 is not encircled. Also the open loop transfer function has no poles ‘on the right half of s-plane. Therefore the clased loop system is stabie. When K > 240 ‘When Kis greater than 240, the contour crosses real axis ata point between —1-+0 and—xo. On travelling through Nyquist plotalong the indicated direction itis found thatthe point -1+0 is encircled in clockwise direction two times. [Since there are two clockwise encirclement and no right half open loop poles, the closed loop system has two poles on right half of s-plane}. ‘Therefore the closed loop systenvis unstable. RESULT ‘The value ofK for stably is 0 1-03, =0 => Ope = Tred / sec Ato =O;¢ = Trad/ sec, 2K _ 2K Gllo)Hija) =~ Gr 72K (ty K(t+ jo)? _K¥1+@? Ztan“o yi+o* Aan K (+07) jo? .K vite? Atan"o Vito? diana f ley — 270° Gijo) HGe) = Wo? ora Sr e(atan co — 270°) Aso+0, © Gljo)HGa) + 22-270" vi) Aso, G(jo)HGa) > 02-90" (3) Fig 4.14.3 : Mapping of section C,in G()HG)-plane From equations (1), (2) and (3) we can say that the polar plot starts at-270° axis at infinity, crosses real axis at-2K and ends at ongin in third quédrant. The section C, and its mapping are shown in fig 2 and 3. MAPPING OF SECTION C, Fig 4.14.2 : Section C, in s-plane “The mapping of section C, froms-plane to G(s) H(s)-plane is obtained by letting $ = Lt Re” in G(s)H(s) and varying 1.) from +n/2 to -n/2. Since s > Re” and R-+ «, the Gis) ..“s) can be approximated as shown below, [i-e., (1+sT)=sT ]apa AE CREA] Sd Ra Ts > 4.40 K(t+s)* Ks? _K s s s Gis)H(s) = Lets= Lt Re? + GIS) His) KI K = 4 _ =0eP ps Lt ReF issu Rel issu Re Roe Se cs When @ G(s) H(s) =002 (4) When o=-%, G(s) H(s)=0e2 enl5) From the equations (4) and (5) we can say that section G, in-plane (fig 4.14.4.) is mapped as circular arc of zero radius around origin in G(s)H(s)-plane with argument (phase) varying from —z/2 to +x/2.as shown in ig 4.14.5. jo, iv Li s-plane G()H@)-plane Roo 5 > t | Fig 4.14.4 : Section C, in s-plane Fig 4.14.5 : Mapping of section C,in G(S)HI(s)-plane Mapping of section C, Insection C,, @ varies from ~o to 0. The mapping of section C, is given by locus of G(jo) H(j@) as @ is varied from — to 0. This locus is the inverse polar plot of G(ja) H(jca). The inverse polar plotis given by the mirror image of polar plotwith respect to real axis. The section C, in-plane and its, corresponding contour in G(s)H{(s) plane are shown in fig 4.14.6 and fig 4.14.7. jo iv s-plane GG@HGS-plane Fig 4.14.6 : Section C, in plane Fig 4.14.7 : Mapping of section C, in G(H(s)-plane Mapping of section C, ‘Themapping of section C, from s-plane to Gis) H(s)-plane is obtained by leting s~ Lt Re” inG{s)H(s) and varying 6 from-ni2 to +72. Since s» Re* and R00, the G(s)H(s) can be approximated as shown below, [ie., (1+8T)= 1]. K (+s? Kxt_K er441 When 6= 4 Gis) H(s)= G(s) H(s)=~e ” From the equations (6) and (7) we can say that section C, in s-plane (fig 4.14.8.) ls mapped as a circular are ofinfnite radius with argument (phase) varying from +3n/2 to ~3r/2.as shown infig 4.14.9. iva jo. G(s)H(s}-plane Fig 4.14.8 : Section C, in s-plane Fig 4.14.9 : Mapping of section C, in G(s)HG)-plane COMPLETE NYQUIST PLOT ‘The entire Nyquist plot in G(s)H(s)-plane can be obtained by combining the mappings of individual sections, as shown infig4.14.10 STABILITY. ANALYSIS When, -2K =~1, the contour passes through ~1+40 pointand corresponding value of Kis the limiting value of K for stably -. Limiting value of K=—=0.5 Nila When K < 0.5 When K is less than 0.5, the contour crosses real axis ata point between 0 and —170. On travelling through Nyquist plot along the indicated direction itis observed that the ~1 +0 pointis encircled in clockwise direction two times. Therefore the system: is unstable. [Since there are two clockwise encirclement and no right half open loop poles. the closed loop system will have two: poles on righthalf of s-plane] When K > 0.5 When Kis greater than 0.5, the contour crosses real axis ata point between —1 +0 and—o. On travelling through Nyquist plotatong the indicated direction itis observed that (-1+0) pointis encircled in both clockwise and anticlockwise direction one time. Hence net encirclement is zero. Also the open loop system has no poles at the right haffof s-plane. Therefore the closed loop systemis stable.pd 9 and Roi Locus 4.42 iva -14)0 for K>0.5, G(s)His}-plane RESULT The system Is stable when K> 0.5. EXAMPLE 4.15 o__(1+4s) The open loop transfer function ofa systemis XSIN(6) == 55 (fe 95) « Determine the stability of closed loop system. ifthe closed loop system is not stable then find the number of closed-loop poles tying on the righthalfof s-plane. SOLUTION ont] ___ (4s) seplane Given tas, HSIN) = at oo The open loop transfer function has tiwo poles at origin. Hence choose ie the Nyquist contour on s-plane enclosing the entire righthalf plane exceptthe é origin as showm in fig 4.15.1. The Nyquist contour has four sections C,, C,, C,and C,, The mapping of each section is performed separately and the overall Nyquist plotis obtained by combining the individual sections. Fig 4.15.1 : Nyquist Contour in s-plane MAPPING OF SECTION C. ee ee ey In section C,, « varies from 0t0-+0. The mapping of section C, is given by the locus of G{jo)H(ja) as w is varied from O too. This locus is the polar plot of Gda)Hue). " (1+4s) SUSIE) 25) (125) Let s= jo. roars ae £6 (}o)H(ja) = (+o) Vit 160? Ztan“‘4o (oy? (1¥ Jo) (14/20) @? 180° 140? Ztan'e ¥t+ 40? Ztan™ Ret prise fran 4 180°—tan*@ —tan”20) Wie? fiedo™ +07 ¥144074.43 14160? +.|G(jo fe nai eae 07¥140? V14+407 ZG(jo)H(Go) = tan!40 — 180° tan@ ~tan“"20 When the G(jo)H(j0) locus crosses real axis, the phase will be -180° and the corresponding frequency is the phase crossover frequency, ,.. ALO = Ope, ZG(jo)H(lo) = 180° ctan”4eop¢ — 180% tan ‘o,¢ ~ tan” 20. = -180° tan dog. = tanta g +B 250 ‘(On taking tan on both sides we get, tan flan “4o501 pe t tan” 20,0] tanA + tanB tan tan“ bg. tan tan "2056 | Note: tan( A + B= ee 30, Se amg =$-1 Ato =0,, = 0.354 rad / sec, i+ 16050" iy 16% 0354? |G(jo)HGo)|- = = =) eye ot fi+o.” Ytt40,.” (0354)? Vir0.36e Vis 4x 0.364 Hence G(jo)H(jo) locus crosses the real axis at 10.64. Ata 0, Gijn)Hjo) > 22-180" 2) Ato >=, G(o)Hfio) 202-2700 ra 43) From equations (1), (2) and (3) we can say that the polar plotstarts at-180° axis at infinity, travels in third quadrantand crosses real axis at -10.64 to enter second quadrant and then ends at origin in second quadrant. The section C, andits mapping are shown infig 4.15.2. and4.15.3. jo ome! ¥ s-plane G@H(@)-plane > Fig 4.18.2 : Section C, im s-plane Wig 4.15.3 : Mapping ey Chin MAPPING OF SECTION C, COHg pam “The mapping of section C, from s-plane to G(s)H(s)-plane is obtained by letting s= Lt R e” inGis)H(s) and varying @ from +2/2 to -n/2. Since s+ Re and R-+, the G(s)H(s) can be approximated as shown below, fie., (14ST) =sT].4.4% _ (14s) 4s 2 SSNS) aaa (ed) Ss DSS Let, s= Lt Re”. roe 2 Z +. G(S)H(s) - 3 =a = 00 el at (re*y When 8= = GteyHts)=0e *2 (4) When @=—3, Gis)H(s)= 082 8) From the equations (4) and (5) we can say that section C, in s-plane (fig 4.15.4.) is mapped as circulararc of zero radius around origin in G(s)H(s}-plane with argument (phase) varying from ~3n/2 to +3n/2.as shown in fig 4.15.5. iva G(s)H(s}-plane Fig 4.15.5 : Mapping of section Cyn Fig 4.154 : Section C, in s-plane Guts) plane MAPPING OF SECTION C, In section C, «varies from —20to 0. The mapping of section C, is given by the locus of G(jw) H(ja) asa Is varied from — eto 0. This locus is the inverse polar plot of G(jo) H(ja). ‘The inverse polar plotis given by the mirror image of polar plotwith respect to real axis. The section C, in s-plane andits corresponding contourin G(s)H(s) plane are shown in fig 4.15.6 and fig 4.15.7. jo. — Hig ALAS siSecto Ct pag Fig 4.15.7: Mapping of section C, in G()H(3)-plane MAPPING OF SECTION C, “The mapping of section G, froms-plane to G(s)H(s)-plane is obtained by letting s= Lt Re inG(s)H(s) and varying 6 from -n/2 to +2/2. Since s > Re” and R > 0, the G(s)H(s) can be approximated as shown below [i.e., (1+ST) = 1]445 (144s) SISIH(s) = s%(148) (1+ 2s) Let, s= Lt Re® Roo =. G(s}H(s) “4 ogy eee are? Toure he) When 0= te G(s)H(s) = wel" a) When @ =) GIs}H(s) =e From the equations (6) and (7) we can say thet section C, in s-plane (fig 4.15.8.) is mapped asa circle of infinite radius ‘with argument (phase) varying from +2 to—ras showin fig 4.15.9. jv G(s)H(s)-plane = ig 4.18.8: Section Cin i Sg SSA nee Fig 4.18.9 : Mapping of section C, in COMPLETE NYQUIST PLOT G(s)H(s)-plane ‘The entire Nyquist plotin G(s)H(s}-plane can be obtained by combining the mappings of individual sections, as shown infig 4.15.10. STABILITY ANALYSIS. Ontravelling through Nyquist contour in G(s)H(s}-piane it is observed that (—1+)0) pointis encircled in clockwise direction two times. Therefore the closed loop system is unstable, Since the— sion esse Nepean opines ‘wo poles of closed loop system arelying on the righthalfs-plane. (s)H(s)-plane Gis) [ : fl a ‘0 144s t Fig 4.15.10 : Nyquist plot of G(s) His) = Piesyiezy4.46 RESULT (@ _ Closed oop systemis unstable. ©)_Twopoles of closed loop system are iying on the right half s-plane. EXAMPLE 4.16 Stetch he Nyquist pet fora system ith the open loop vensterkncton Gis)H(e)= SAPS) = . Determine the TOs) (6 range of values of K for which the system is stable. SOLUTION K(1+0.5s) (14s) (1+ 10s) (8-1) “The open loop transfer function does not have.a pole at origin. Hence choose the Nyquist contour on s-plane enclosing the entire right half plane as shown in fig Fig 4.16.1: Nyquist 4.16.1. Contour in s-plane Tho Nyquist contour has three sections C,,C, and C,, Tho mapping of each section is performed soparatoly and tho overall Nyquist plot is obtained by combining the individual sections. Given that, G(s)H(s) = MAPPING OF SECTION C. InsectionC,,« varies from 0 to +x, The mapping of section C, is given by the locus of Gju)H(jw) 2s « is varied from 010-0, This locus is the polar plot of GGa)H{jao). (1405s (145) G(S)H(s) (105) ( Let s= jo. 2. G(jo)H Go) = £110.52) (1+ Jo) a K(1+}1.50-0507) _ K(1- 0.50?) + jiSoK (1+ j10) (-1+ jo) =1- 9-100? 151007) 80 ‘On multiplying the numerator and denominator by the complex conjugate of denominator we get, K(- 0.507) + jt5ak i ~(14100?) +90 =(141007)— [80 (14+ 107) + j90 -K(1- 0502) (141002) -1350*K + -_iIBok(1-0507)- 150K(1+ 1007)] 100" F (Gay G{jo)H(jo)= When the Glja}H(jw) locus crosses real axis the imaginary term is zero and the corresponding frequency is the phase crossover frequency, ALO = O50, 900 K(1- 0.5003,) - 1505, K(1+ 1002.) =0 15 £90). K(1-0.502.)=150,.K(1+1002) => 10508 = F(1+ 1003) 21-0503,=0467+1670%. => = 2.1703, = 0833. => =0.62rad/sec Ato=o,,= 0.62 radisec4.47 -K(1- 0.5005.) (14 100.2) = 13.50 5.7K tio)HGo) = Cir ide Fs Go, (10.5 x 0.627) (1410 x 0,627) + 13.5% 0.62?]__f_ 3.913 +5.189 (1410 x 0.62%)? +(9 x 0.62) See | =-0.18671 23.464 +31. 5 eK Therefore, G(ja)H(jo) locus crosses real axis at a point-0.1667K. ‘The exact shape of G{jo)H(0) locus is determined by calculating the magnitude and phase of G(ja)H(je) for various values of ©. eva (14+ j050) (14 jo) ¥1+(0.50)? ztan105 V1+07 Ztan'o Se Ri8) = os) Tee) A+(100)? Ztan“10o View? 2180--tan) 2x Yi0250" (tan“1050 + 2tan“e ~ tan""100 ~ 180°) V1#1000? 1402507 vi+1000? 2G(jo)H(e) = tan'0.50 + 2tan™’o —tan-'100— 180° = |G(jo)HGo)] = K. Aso +0, |G(o)HGo)|=K Asoa>0, 2G(jo)H(ije)=-180" 2 Vie 02507 _ io [or(S +025) 1.4028) ola) KW [e029] Dak Kye a = 005K | (Se t00} 100] * R100" — “eeey ote +100 M AS @ > 0, ZG(jo)= tan” bo + 2tan'~ tan!z0— 180" = 90° +180°— 90° 180" =0 * na | 1 @ 1 —q 1 rad/sec 0 OL 05 LS 2.0 50 | am | Gleo)Hgo) K o7ork | 202K | 0.083K | 007K | 0.054K | 05K | t <6 )HGe) -180 -210 -191 116 95 43 | 0 | fa | | Fromthe above analysis, the following conclusions are made, 1... The locus of G(jo)H(o) starts at K2—180° when «=0 and travels in second quadrant. 2. The locus crosses real axis at-0.1667K and enters third quadrant. 3. Then the locus crosses negative imaginary axis and enters fourth quadrant. 4. Finally the locus ends at 0.05K 20° whena==, Note: The exact plot can also be sketched on polar graph sheet. ‘The section C, in s-plane and its corresponding mapping in G(s)H(s) plane are shown in fig 4.16.2. and 4.16.3.4.48 im o-D> 667K Fig 4.16.2 : Section C, in s-plane Fig 4.16.3 : Mapping of section C, in MAPPING OF SECTION C, Senile pane G()HCs)-plane ‘Themapping of section C, froms-plane to Gis)H(s}-plane is obtained by letting s=,Lt_R e® in G(s)H(s) and varying 6 from—a/2to +02. Since s > Re® and R->~, G(s)H(s) can be approximated as shown below |ie., (1+sT)=sT: Here (s-1)=s}. _K(I+0.58) (148) GISH(8) = ear. iva ve ~ ‘s-1) |GG)H(s)-plane we ROSS *S _ O05K 10s xs t The approximate G(s)H(s)is independent pc0BK: fs and so the contour of section C, in s-planeis mapped as apointat0.05K in G(s)H(s}plane. Big 4,164 : Section Cjin s-plane Fig 4.16.5 : Mapping of section C, in MAPPING OF SECTION C, G()H(s)-plane Insection C,, varies from —0to 0. The mapping of section C, is given by the locus of Gjn) H(o) as o Is varied from ~ ©1000. This locus is the inverse polar plot of G(jo) Hj). ‘The inverse polar plotis given by the mirror image of polar plotwith respect to real axis. The section C, in s-plane and its corresponding contourin G(s)H(s) plane are shown in fig 4.16 6 and fig 16.7. iv G(S)H()-plane -0.1667K Fig 4.16.7 : Mapping of section C, tn Fig 4.16.6 : Section C, ia s-plane G(jH()-plane COMPLETE NYQUIST PLOT ‘The entire Nyquist plotin G(s)H(s)-plane can be obtained by combining the mappings of individual sections, as shown infig 4.16.8. iv G(S)H(s)-plane 14/0 for -140 for -1+j0 for O6 \ -0.1687K K(I+05s)(1+5) Flg 4.16.8: nyaust pot of Gio HO) = Try449 STABILITY ANALYSIS When —0.1667K =—1, the contour passes through -1+/0 point and this value of K is the limiting value for stability, ‘The limiting value of K =——— limiting value of K = When 06 When K>6, the locus crosses reat axis between —1+)0 and —<=, On travelling through the locus itis observed that the —14)0 points encircled anticlockwise one time. Also the open loop system has one pole at the right half s-plane. Hence the systemis stable, i RESULT (2) Theopentoop systemis unstable. (0) Forstabilty ofthe closed loop system, K>6. EXAMPLE 4.17 Construct Nyquist pltfora feedback contro system whose open oop transfer functions given by, G{S)Hi(s)= ‘Commenton the stability of open-loop and closed loop system. jo, ota SC SOLUTION cy s-plane —_—_ o=0 Given that, GS)H(8) = 5 ont, Gi ‘The open loop transfer function has a pole atorigin. Hence choose the Nyquist Geo contour on s-plane enclosing the entire right half s-plane except the origin as shown in Fig 4.17.1 : Nyquist fig 4.17.1 Contour in s-plane ‘The Nyquist contour has foursections C,,.C., C, and C,. The mapping of each section is performed separately and the overall Nyquist plotis obtained by combining the individual sections. MAPPING OF SECTION C, Insection C,,« varies from 0 to +x, The mapping of section C, is given by the locus of G(ja)H(je) as o is varied from Oto. This locus is the polar plotof G(ja)H(w), 5 GOH) FS Lets= jo. -Gllo)H(o)= Jol jo) 5 0290" Vita? Z-tan'o (-90+ tanto)4.50 5 HG Qo)Hjo)| = ee ovis Note: (7-je) represents 2 point £G(jo)H{jo) = - 90° + taro in fourth quadrant ‘The exact shape of G(ja)H(jo} locus is determined by calculating the magnitude and phase of G(jo)H(je) for various values ofa. re ‘ees 0 os 10 20 10.0 © (Go)HGoy * 745 | 338 142 0.05 ° <6Hie) -90 59 45 26 | 3s 0 ms | Fromthe above analysis, we can conclude that Gija)H(jo) locus starts at -90° axis at infinity for o=0 and meets the origin along 0° axis when @=c. ‘The section C, in s-plane and its corresponding mapping in G(s)H(s)-plane are shown infig 4.17.2 and 4.17.3. = beCd od Mg 417-2 Section Cin splone) Fig 4.17.3 : Mapping of section C, in MAPPING OF SECTION C, G()H(s)-plane ‘Themapping of section C, from -jlane to G{s)H(e)-planeis obtained by ering s = Lt Re" inGis)H(s)and varying? from +w/2 to-w/2. Since s+ Re” and R-» =, the G(s)H(s) can be approximated as shown below, [Le., (1-8) =~s)] oe Maw wot SSNS) Sea) eF Let s= Lt Re®, eke G(s)H(s) = oes 5 aun ROTO G(s)H(s) = Oe"? geet =0e” Whenge=>. (SekMeaDe . a x a eh OF ER (2) From the equations (1) and (2)we can say that section C, in s-plane (fig 4.17.4) is mapped as circulararc of zero radius around origin in G(s)H(s) plane with argument varying from —2 to +0.as showninfig 4.17.5. io, jv s-plane |G()H)-piane a0 Cy | Fig 4.17.5: Mapping of section C, in ig 4.17.4 : Section C, in s+ Fig 4.17.4 : Section C, in s-plane GigHa olane4.51 MAPPING OF SECTION C, Insection C,, « varies from -:00 0. The mapping of section C, is given by the locus of G(ja)H(jo) as a is varied from~ coto0. This locus is the inverse polar plot of Gio) H(jo).. ‘The inverse polar plotis given by the mirror image of polar plotwith respect to real axis. The section C, in s-plane and its corresponding contour in G(s)H(s) plane are shown in fig 4.17.6 and fig 4.17.7. joa | s-plane Fig 4.17.6 : Section C, in s-plane Fig 4.17.7 : Mapping of section C, in . G(s)H(s)-plane MAPPING OF SECTION C, The mapping of section C, rom splane to G(s)H(6)-planeis obtained by leting s = Lt R ef in G{s)!(s) and varying from—w2to +7/2. Since s+ Re” and R->0, G(s)H(s) can be approximated as shown below, [i.e., (1-8) SSHG)" Fig axis Lets= Lt Re®, 0 =. G(S)H(s) | =a pre? Lt Re® ‘seus eB ROO Ro! Gin? (3) When 6= GIsJH(p=e Fg (4) From the equations (3) and (4) we can say that section G, in-plane (fig 4.17.8.) is mapped as a circular arc of infinite radius with argument varying from 2/2 to -r/2.as shown in fig 4.17.9. jo, s-plane & aod Fig 4.17.9 : Mapping of section C, in G(s) H(s)-plane Fig 4.17.8 : Section C, in s-plane COMPLETE NYQUIST PLOT ‘The entire Nyquist plotin G(s)H(s)-plane can be obtained by combining the mappings of individual sections, as shown infig4.17.10.Siabiliy and Roo? Locus 4.52 iv G(S)HG6)-plane Fig 4.17.10 : Nyquist plot of G(s) HO) STABILITY ANALYSIS s(i-s) The Nyquist contour in G(s)H(s)-plane does not encircle the point (—1+j0) but the open loop transfer function has one pole on the right half s-plane, Therefore the system is unstable. RESULT Both open loop and closed loop systems are unstable. EXAMPLE 4.18 By Nyquiststabiliy criterion determine the stabily of ciosed loop system, whose open loop transfer function is given by, (s+2) #Ne-1 SOLUTION CIS)H(s)= - Commenton the stability of open-loop and closed loop system. (s+2) {s#1) (s-1) The open loop transfer function does not have a pole at origin. Hence choose the Nyquist contour on s-plane enclosing the entire right half plane as shown in fig ig 4.18.1 : Nyquist 4184 Contour in s-plane Given that, G(s)H(s) = ‘The Nyquist contour has three sections C,,, C, and C,, The mapping of each section is performed separately and the overall Nyquist plotis obtained by combining the individual sections. MAPPING OF SECTION C, Insection C,, varies from 0 too. The mapping of section C, is given by the locus of G{jo)H\ja) as a is varied from 0 toc. This locus is the polar plot of G{ja)H(o) [Re Frapesons opin +2 2(1+05s) second quadrant SOHO SHED Tsctes) 2(1+ j0.5e) 2140250? /tan'0.50 Lets=jo. + G(jo)H§o)=—“" Gra? oats hee Aipoe sens), i {Io Fj) (-T+i0) ~ iva? ano fire? 2(180°-tan-)453 _ 21502507 at -£(-180 +tan'0.50) fico? *1GGayAjoy = 2YOZ50° ZGXjo)Hjo) = -180°+ tan050 The exact shape of Gja)H(j) locus is determined by calculating the magnitude and phase of G(je)H(ja) for various values of. @ I | T nadie ° 04 to | 20 | 100 = | KhoyHGoy |. 2 1.76 12 | os7 [on 0 eH) 180 | 168 -1939 | -195 | 01 | -20 lL deg | From the above analysis, we can conclude that Gjo)HJe) locus starts at-180° axis ata magnitude of -2fora=0and meets the origin along-90" axis when @=+-, The section C, in s-plane and its corresponding mapping in G(s)H(s}-plane are shown in fig 4.18.2. and 4.18.3. iv G(s)H(s)-plane ‘Fig 418.2: Section C, in s-plane Fig 4.18.3 : Mapping of section C, in G(SH()-plane MAPPING OF SECTION c ‘The mapping of section C,, from s-plane to G(s)H(s)-plane is obtained by letting s = at Ref inGis)H(s) and varying from +2/2to~a/2. Since s+ Re” and R->, Gis) H(s) can be approximated as shown below, {ie (1+3T)*ST]. ane SORBED 22884 Let, s= bt Re®. -G(s)H(s) ce nah “oR oe G(S)H(s) =| oo? (1) Z| G{siHis)=0e!2 wnf2) 2 From the equations (1) and (2) we can say that section C, in s-plane (fig 4.18.4) is mapped as circular arc of zeroradius ‘around origin in G(s)H(s)-plane with argument varying from ~r/2 to +4/2as shown in fig 4.18.5.ge a) iy and Rook Locus 4.4 iv fatein(sy-plane Ro0 ! Fig 4.18.5 : Mapping of section C, in G(s)H(5)-plane Fig 4.184 3 Section C, in s-plane MAPPING OF SECTION cy In section C,, « varies from -soto 0. The mapping of section C, is given by the locus of G(ja)H(ja) aso is varied from-to 0. This locus is the inverse polar | plotof G(j@)H(je). The inverse polar plotis given by the mirror image of polar plot with respectto real axis. The section C, ins-plane and its corresponding contour in G(s)H(s) plane are shown in fig 4.18.6 and fig 4.18.7. jo s-plane iv 1G(s)H(s)-plane ae re) > Cc, oma] " ; , Fig 4.18.7 : Mapping of section C, in Fig 4.18.6 : Section C, in s-plane GOH) plane COMPLETE NYQUIST PLOT ‘The entire Nyquist piot in G(s)H(s)-plane can be obtained by combining the mappings of individual sections, as shown in fig 4.18.8, STABILITY ANALYSIS ‘On travelling through Nyquist contour itis observed that 4 ~1+0 point is encircled in anticlockwise direction one time. Also the open loop transfer function has one pole at righthalf's-piane. ‘Since the number of anticlockwise encirclement is equal ionumber cof open loop poies on righthalf's-plane, the closed loop system is stable. . RESULT @ —_ Opentoopsystem is unstable @)__Closed loop systemis stable. 46 RELATIVE STABILITY The Relative stability indicates the closeness of the system to stable region. It is an indication of the strength or degree of stability. (s+2) Fig 4.18.8 + Nyquist plot of Gls} H()= eae Tn time domain, the relative stability may be measured by relative settling times of each root or pair of roots. The settling time is inversely proportional to the location of roots of characteristic equation. If the root is located far away from the imaginary axis, then the transients dies out faster and so the relative stability of system will improve. The transient response and so the relative stability for various location of roots in s-plane are shown in fig 4.6.4.55 s-plan * a Fig 4.6 : Transient response and relative stability for various locations of roots on s-plane In frequency domain the relative stability of a system can be studied from Nyquist plot. The relative stability of the system is given by closeness of polar plot to -14j0 point. As the polar plot gets closer to -1+j0 point the system moves towards instability. Unity crete The relative stability in frequency domain are quantitatively measured in terms of phase margin and gain margin. Consider a G(jo) HG) locus as shown in fig 4.7. Let this locus cross the real axis at point-A and a unit circle drawn with origin as centre cuts this Jocus at point-B. Let G, be the magnitude of Gj@)H(Ga) at point-A by and y be the angle between negative real axis and line OB: Fig: 4.7 If the gain of the system is increased, then the locus will shift upwards and it may cross real axis at —14j0 point. When the locus passes through -1+J0 point, G,->1 and y->0. Hence the closeness of GG) HG) locus to the critical point -1+40 can be measured in terms of intercept G, and angle y. The vaiue of G, and y are quantitative indications of relative stability. These values are used to define gain margin and phase margin as practical measures of relative stability. The concepts of gain margin and phase margin are defined for open loop systems but from the values of gain margin and phase margin the stability of closed loop system can be judged. 47__GAINMARGIN AND PHASE MARGIN Gain margin is a factor by which the system gain can be increased to drive the system to the verge of instability, With reference to fig 4.7 the magnitude of GGo)H@a) is G, when it crosses real axis and the phase corresponding to that point is ~180°. The frequency corresponding to that point be o,,. If the gain 1 of the system is increased by a factor ~ then the magnitude at the frequency @,, will be, 0, * Now the G(jo)HGa) iucus will pass through -10 point driving the system to the verge of instability. Hence the gain margin, K, of the system may be defined as the reciprocal of the gain at which the phase angle is 180°. The frequency at which the phass angle is 180° is called phase crossover frequency.Gain margin, K,=——__ = e IGG@)HU}\o -0,, Gs Gain margin in db = 20 log: = 20 log. I 5 [GGo)HGo)| GOS Note : Gain Margin in decibels is given by negative of db magnitude of G(ja) at phase crossover ncy. Hence at @,, if db magnitude is negative, then gain in is positive and vice versa. e The phase margin is defined as the amount of additional phase lag at gain crossover frequency required to bring the system to verge of instability. The frequency at which the magnitudes of GGa)HGo) equals unity is called the gain crossover frequency, «,.. With reference to fig 4.7, the phase angle corresponding to the meeting point of unity circle and GG@)H(ja) locus is —180° + 7. Now with magnitude remaining unity, if an additional phase lag equal toy is introduced then the net phase angle becomes —180° and G(j@)HGo) locus will pass through —1+J0 point driving the system to the verge of instability. This additional phase lag y is known as phase margin. Let, Og. = £GJ@) H(i )igno, ; Now -180°+7=6,, ©: Phase margin, y= 180° +$,. For stability of closed loop system the gain margin of open loop system should be greater than t or if itis expressed in db it should be positive and phase margin of open loop system should be positive. EXAMPLE 4.19 K ‘The open loop transfer function ofa unity feedback system is given by, G(s) = xivatiraty forgain Kin terms of T,, T, and specified gain margin, K,, SOLUTION Derive an expression K Given that, a) rn Menthet, G8) = eras eT) io) x = x = SAT Se o(l+joT)0+ jo.) Jo(1+ jot, + joT;-2*T,) K k = SNS Poca a jo[t+ jo(F, +T,)-07F7,] — jo-07(T,+7,)- jo FT, Ps K “Ta +) 0-0")4.57 “The gain margin, K, is defined as the reciprocal of the magnitude of G(jo) at phase crossover frequency. At phase ‘crossover frequency the magnitude is purely real. Hence at phase crossover frequency, oy, the imaginary partof Gijo) iszero, ALO FO, Ope(1-ORTi)=0 > ORT =0 = 1 -—— vile Ato~a,,the imaginary partis zero, . 0, k_ | K SHOE amr FTO OTH ATA 1 1 = e+ Ta) ladey., K/on( +7) 2.Gain margin, K, = Put, 02, =giin the above equation. Te 1 \ +T) la) * a1 T+h 4 (= it | R = tn ite Tit RESULT “The expression for gain K in terms of K,, T,and 1, iS, K= x ( 3 tt) EXAMPLE 4.20 Determine the Gain crossover frequency, Phase crossover frequency, Gain margin and Phase margin of a system with ‘ 1 openloop transfer function, G{s)= + SOLUTION To find phase crossover frequency and gain margin 1 Given that, 68) G95 Gray 1 1 : i 2 4 “jo Re) (+) jo(1Je+a—20%) jo (Ba+1-207) Bo? + jo (1- 20%) Atphase crossover frequency the imaginary partof G(ja)iszero, Hence pute ~a,, inimaginary partand equate to zero to solve fora, + Ge) 4 Wye(1- 20%) =0 =0.707 rad/ sec since @, #0, 1-202 =0 > -20%= rn “The gain margin, K, is defined as reciprocal of magnitude of G(jo) at phase cross over frequency. ‘i a [r-30?|, Gain margin in db = 20 log K, = 20 log 1.53.0 db Gain margin, K, = 30%, = 3x 0.707? = 15lo(t* Pais fo)” 290° Vis 40% Ztan 2oVi+02 tanta 1 ovis 4a" Jive 2G(jo) = -90" - tan"20 —tan“"o Atgain crossover frequency, «, the magnitude of Ga) is unity. 1 fried finer Solving the above equation for w , will be tedious, Hence by trial and error find the root of the above equation. -Ato=0%, — 1Gho)|= Wheno=%4 — |G(io)l= 1 1 ee ee = 03 oltedo? Jive? Nira vivt at = 4 =126 viedo? Vivo? O5VIr 4x05" Vie05" From the two calculations shown above, we can conclude that the unity magnitude will occur for a frequency between 0.5and 1.0, : When =05, 1GGo)) Whene =0.6, |GGa)= = 0915 1 08 J1+4 0.8 +067 1 = 1005 057 Vi¥4x057 Vie057" Leta = 0.57 be the gain crossover frequency, since for this valle of a the magnitude of Gfja) is approximately equal to When @ = 0.57, [GGw one. +. Gain crossover frequency, «,, = 0.57 rad/sec. Letthe phase of G(jo) ata, be}. Ato og =057, $y, =-90° -tan'2e- tanto =-90° ~tan-'(2 x 0.57) —tan'0.57 = -168° -.Phase margin, y = 180° +89, = 180° - 168° = 12° RESULT (©) The phase crossover frequency, o,.= 0.707 radisec ©) The gain crossover frequency, o,, = 0.57 radisec (© Thegainmargin, K,=1.5 The gain margin in db = 3.5 db ‘ (@\ The phase margin, y= 12°459 EXAMPLE 4.21 The open loop transfer function ofa systemis Gis) = re 0 ds Gas Winds eric @® Determine the vatue of K so that phase margin is 40°. SOLUTION (i) __ To find K for specified gain margin Given that, G8) =. 5.59.16) ee) Lets =jo. BS) a ee 9s ___, Ue) =O) (jo) Jol |te-Om) te? Fj 0-0.m) ‘Atphase crossover frequency a,., the GG) is real and so equate the imaginary partto zeroto solve fora, . ALO=O5, — Me(T-0405,)=0 => 1-02. =0 = -0103, Ta 162 rad/ sec 1QUO)aaye= 6, “The Given that gain margin = 6db, 20 fog K, =6 => log qe ©. Gain mat K =102 = 19953 By definition of gain margin, Gain margin, K, = ate, a SO 1 =o BI: = 9900. 18088 se (ii)_To find K for specified phase margin Given that, G(s) = Lets 8 (1+0.18) (1+5)" ee te K jo (14 J.) (1+ Jo) 290° fi+(0.t0)? Ztan"Oto Vite? ctano K SON Foon? Viva™ £G(jo) = -90° - tan'0-40 - tana < Gio)bapter =o: iSof Stability and Root Locus >) 4.60 Let, @, = Gain crossover frequency 4,.= ZGGo) ata =o,. At ge Oye = ZG(j0)] og. = -BO— tan"104 «55 — tana, By definition of phase margin, Phase margin, y= 180°+,. ‘The required phase marginis 40°, = 40° 2 tan OA ag, + tan"'oy, = 50° On taking tan on eithar side we get, ton (arp) = 20 Attn B Lee i tanAtanB tan [tan10.405. + tan", |= tan 50° 180° - 90° —tan*Ot0.~tanto, => tan 0.10, +tan tog, = 180° - 90° — 40" tan tan“'Oog. + tan tan™o,. O-t0gs +g 1.4095 : - = 1192 ou Toten tr'Oo,, x tn taro, "> To, Ko, = Come Ne ‘On cross multipiing the above equation we get, Wg, = 1192 (1- Ose) => 0119203, + 1hrg, - 1192-0 " 1192 5%. -10=: 8 + ayag3 0 GaaeD™ 0%, +9.228 wg, -10=0 9, = 29228 = 02287 44x10 _ -9.228 2111873 iil a ~ 2 ‘Onteking positive value we get, 9.228 + 114873 = 0.98 rad/sec K IGGo)l=1 5 [GGo)L og our O0m,? frog? + 0.0409.” 1+ Wg, + K2 Og 4140.00,” Y1t 05." =0. 98 ¥1+0.01, 0.08" [1+ 0.98" =1.3787 1 RESULT Foragain margin of6 db, K=5.5135 Foraphase margin of40°,_K= 1.3787 48 ROOTLOCUS The root locus technique was introduced by W.R.Evans in 1948 for the analysis of control systems. ‘The root locus technique is a powerful tool for adjusting the location of closed loop poles to achieve the desired system performance by varying one or more system parameters. Rn 8 (S+p,) (S+P2) ‘The closed loop transfer function of the system with unity feedback is given by, K CS) __Gs)___ 88+) +P.) K RG) 1+G(s) yj, Ks (8+ p,) +P.) +K s(s+pi) +P2) Consider the open loop transfer function of system G(s)=4.61 The denominator polynomial of C(sV/R(s) is the characteristic equation of the system. The characteristic equation is given by, : s(s+p)(S+p,)+K=0. The roots of characteristic equation is @ function of open loop gain K. [In other words the roots of characteristic equation depend on open loop gain K J. When the gain K is varied from 0 to'=n, the roots of characteristic equation will take different values. When K = 0, the roots are given by open loop poles. When K-09, the roots will take the value of open loop zeros. ‘The path taken by the roots of characteristic equation when open loop gain K is varied from 0 to are called root loci (or the path taken by a root of characteristic equation when open loop gain K is varied from 0 to 0» is called root locus). Note : In general the roots of characteristic equation can be varied by varying any other system Parameter other than gain. In general the closed loop transfer function of system with multiple loops is obtained from the signal flow graph of the system using Mason's gain formula. LO) _ RO T(s)= < ERA, (Refer chapter 1 séction 1.12) The determinant, A is the denominator polynominal of C(s)/R(s). The characteristic equation of the system is given by , A= 0 . For the single loop system shown in fig 4.8 Ris) lava f(s) cH) _ Gs) "Ga RG) 1+6@HO H(s) }« The Characteristic equation is, a 14+G(s) H(s) = 0 ©. G(s) H(s) = -1 (42) From equation (4.22) it can be concluded that the roots of the characteristic equation occur only for those values of s for which, G(s)H(s) = -1. The equation (4.22) can be converted to two Evans conditions given below, {G@H@|=1 ZG(S)H(s)= £180" (2q +1), where q = 0,1,2,3, The equation (4.23) is called magnitude criterion and equation (4.24) is called angle criterion. The magnitude criterion states that 5 = 3, Will be a point on root locus if for that value of s, IG) H(| = 1.[chapter 4 ~ Conecepisof 4.62 The angle criterion states that s = s, will be a point on root locus if for that value of s, ZG(8) HG) is equal to an odd multiple of 180°. The function G(s)H(s) can be expressed as-a ratio of two polynomials in s as shown below. (+2) (s+2) (+5) GH =k B42) G22) 6%) (4.25) eae (s+ Pr) (+ Pa) (+ Ps), |G@HO] = K +ai)x[s+z2|x[p*zs).~ {s+p,]x|s+ palx|s+p. where, m= Number of zeros of loop transfer function. . n= Number of poles of loop transfer function. The magnitude criterion states that |G(s)H(S)| = Tils+2 Dijs+p, xattal Ls de sca i (4.26) Tijs+ pil Hs al The open-loop gain K corresponding to a point s= s, on root locus can be calculated using equation (4.26), It can be shown that |s + p| is equal to the length of vector drawn from s =p, to s = s, and js +z] is equal to the length of vector drawn from s =z, to $= s,. Hence the equation K can be written as, Product of length of vector from open loop poles to the point s= sq. Product of length of vectors from open loop zeros to the point s= sq From equation (4.25), LG(s)H(6) = 2(8+ 2) + Z(s+ 23)+ L(S + 25)+.-L(S+ Py) — L(S+ Pr) — L(S+ Ps )ene = 26+2)-Y) 26+p) a a where, m= Number of zeros of loop transfer function. n= Number of poles of loop transfer fimetion. * The angle criterion states that £G(s)H(s) = #180° (2q + 1) L8+4)-}) As+p))=4180°2q+1) son(.27) The equations (4.27) can be used to check whether a point s = s, is a point on root locus or not. It can be shown that (s+p) is equal to the angle of vector drawn from s~ p, to s ~ s, and Z(s+z) is equal to the angle of vector drawn from s = z, to s = s,. Hence equation (5.27) can be written as ( Sum of angles of vector) (Sum of angles of vector | from open loop zeros |~} from open loop poles |= £180° (2q+1) ‘ to the point s=sq (to the point s=sq46 Beara Systems Taginivering ) CONSTRUCTION OF ROOT LOCUS The exact root locus is sketched by trial and error procedure. In this method, the poles and zeros of G(S)H() are located on the s-plane on a graph sheet and a trial point s = s, is selected. Determine the angles of vectors drawn from poles and zeros to the trial point. From the angle criterion, determine the angle to be contributed by these vectors to make the trial point as 2 point on root locus. Shift the trial point suitably so that the angle criterion is satisfied. ‘A number of points are determined using the above procedure. Join the points by a smooth curve which is the root locus. The value of K for a particular root can be obtained from the magnitude criterion. ‘The trial and error procedure for sketching root locus is tedious. A set of rules have been developed to reduce the task involved in sketching root locus and to develop a quick approximate sketch. From the approximate sketch, a more accurate root locus can be obtained by a few trials. RULES FOR CONSTRUCTION OF ROOT LOCUS Rule 1 ; The root locus is symmetrical about the real axis. Rule 2: Each branch of the root locus originates from an open-loop pole corresponding to K = 0 and terminates at either on a finite open loop zero (or open loop zero at infinity) corresponding to K =0o, The number of branches of the root locus terminating on infinity is equal to n—m, (.e., the number of open loop poles minus the number of finite zeros) Rule 3 : ‘egments of the real axis having an odd number of real axis open-loop poles plus zeros to their right are parts of the root locus. Rule 4; The n—m root locus branches that tend to infinity, do so along straight line asymptotes making angles with the real axis given by, _ 180° q+) n-m * bs q=O1 Rule 5: The point of intersection of the asymptotes with the real axis is at s = 6, where, Sum of poles—Sum of zeros nm Rule 6; The breakaway and breakin points of the root locus are determined from the roots of the equation dK/ds = 0, If r numbers of branches of root locus meet at a point, then they break away at an angle of +180°/r. Rule 7: The angle of departure from a complex open-loop pole is given by, 6, = £180° (2q+ 1) +5 9= 01,2) on where @ is the net angle contribution at the pole by all other open loop poles and zeros. Similarly the angle of arrival at a complex open loop zero is given by, , = £180° 2q+ 1) +9; 9= 0, 1, pve where 6 is the net angle contribution at the zero by all other open-loop poles and zeros.4.64 Rule 8: The points of intersection of root locus branches with the imaginary axis can be determined by use of the Routh criterion. Alternatively they can be evaluated by letting s= ja in the characteristic equation and equating the real part and imaginary part to zero, to solve for @ and K. The values of @ are the intersection points on imaginary axis and K is the value of gain at the intersection points. Rule 9 ; The open-loop gain K at any point s = s, on the root locus is given by, I. +Pil _ Product of vector lengths from open loop poles to the point sq fiz] Product of vecior lengths from open loop zeros to the point sq a Note : The length of vector should be measured to scale. If there is no finite zero then the product of vector lengths from zeros is equal to 1. TYPICAL SKETCHES OF ROOT LOCUS PLOTS Angle of departure ep S-plane Root locus on real axis ofimaginary axis Centroid (meeting point of 5. © asymptote with real axis) Se 85 25 Angle of departure ze jor Angle of — aeval s-plane o |__Angle of ‘Arrival PROCEDURE FOR CONSTRUCTING ROOT LOCUS. Step 1: Locate the poles and zeros of G(s)E(6) on the s-plane. The root locus branch starts from open loop poles and terminates at zeros. ‘Step 2: Determine the root locus on real axis. Step 3; Determine the asymptotes of root locus branches and meeting point of asymptotes with real axis. Step 4: Find the breakaway and breakin points.465 contd § Step 5: If there is a complex pole then determine the angle of departure from the complex pole. If there is a complex zero then determine the angle of arrival at the complex zero. ‘Step 6: Find the points where the root loci may cross the imaginary axis. Step 7: Takea series of test points in the broad neighbourhood of the origin of the s-plane and adjust the test point to satisfy angle criterion. Sketch the root locus by joining the test points by smooth curve. . Step 8 : ‘The value of gain K at any point on the locus can be determined from magnitude condition, ‘The value of K at a point s = s, , is given by, product of length of vectors from poles to the point, product of length of vectors from finite zeros to the point, s | ‘Note : When there is no finite zero, the denominator is taken as unity. The length of vectors | | should be measured to scale. | EXPLANATION FOR THE VARIOUS STEPS IN THE PROCEDURE FOR CONSTRUCTING ROOT LOCUS Step 1 : Location of poles and zeros Draw the real and imaginary axis on an ordinary graph sheet and choose same scales both on real and imaginary axis. The poles are marked by cross "X" and zeros are marked by small circle "o". The number of root locus branches is equal to number of poles of open loop transfer function. The origin of a root locus is ata pole and the end is at a zero. Let, = number of poles m = number of finite zeros Now, m root locus branches ends at finite zeros. The remaining n-m root locus branches will end at zeros at infinity. Step 2 : Root locus on real axis In order to determine the part of root locus on real axis, take a test point on real axis. If the total number of poles and zeros on the real axis to the right of this test point is odd number, then the test point lies on the root locus. If it is even then the test point does not lie on the root locus. Step 3 : Angles of asymptotes and centroid If nis number of poles and m is number of finite zeros, then n—m root locus branches will terminate at zeros at infinity. These n—m root locus branches will go along an asymptotic path and meets the asymptotes at infinity, Hence number of asymptotes is equal to number of root locus branches going to infinity. The angles of asymptotes and the centroid are given by the following formulae. Angles of asymptotes = 2180@a*1) n—m where, q = 0, 1, 2, 3, ......(t-m) Sum of poles Sum of zeros Centroid (meeting point of asymptote with real axis) = comStep 4: Breakaway and Breakin points ‘The breakaway or breakin points either lie on real axis or exist as complex conjugate pairs. If there is a root locus on real axis between 2 poles then there exist a breakaway point. If there is a root locus on real axis between 2 zeros then there exist a breakin point. If there is a root locus on real axis between pole and zero then there may be or may not be breakaway or breakin point. Let the characteristic equation be in the form, B(s) + K A(s) = 0 ~Bis) A(s) The breakaway and breakin point is given by roots of the equation dK/ds = 0. The roots of dK/ds = 0 are actual breakaway or breakin point provided for this value of root, the gain K should be positive and real. Step 5: Angle of Departure and angle of arrival Angle of Departure i (om of angles of vector to the ] (Sum of angles of vectors to the) (from a complex pole A), complex pole A from other poles) (complex pole A from zeros ‘Note : The angles can be calculated as shown in fig 4.9 or they can be measured using protractor. 9, = 180%tan' = joa b 3 / a | 8, = 180% tan® = \s-plane e | 8, = 90° a | 9, =tant 2 d @, =tan” & e A Fig 4.9 : Calculation of angle of departure Example: | Consider the two complex conjugate poles A and A* shown in fig 4.9.(if poles are complex then they exist only as | conjugate pairs) | Angle of departure] _ s goe_¢g, 405465) + (0) +04) | atpoleA | Laie | Angle of departure] | eel a “h —[Angle of departure at pole A] |467 GDRonirol Systems Engineering) Angle of arrival at a] complex zero A Sum of angles of vectors to the) _ ‘Sum of angles of vectors to the complex zero A from all other zeros) \ complex zero A from poles} Note : The angles can be calculated as shown in fig 4.10 or they can be measured using rotractor| 9, =180°— tan? 2 b Poa 8, = 180% tan = é @, = 90° a 0, = tan? — . d 6, = tant e x Fig 4.10 : Calculation of angle of arrival Example: Consider the two complex conjugate zeros B and BY as shown in fig 4.10.(Ifzeras are complex ‘then they exist only as conjugate pairs) Angle of arrival “Lamnts |. 180°-(8, +63) + (@, +6, +8s) Angle of arrival at zero B* \. ~[Angle of arrival at zero B] Step 6 : Point of intersection of root locus with imaginary axis The point where the root loci intersects the imaginary axis can be found by following three methods. 1. By Routh Hurwitz array. 2. By trial and error approach. 3. Letting s =j in the characteristic equation and separate the real part and imaginary part. Two equations are obtained : ane by equating real part to zero and the other by equating imaginary part to zero, Solve the two equations for @ and K: The values of @ gives the points where the Toot locus crosses imaginary axis. The value of K gives the value of gain K at there crossing points. Also this value of K is the limiting value of K for stability of the system. Step 7 : Test points and root locus Choose a test point. Using a protractor roughly estimate the angles of vectors drawn to this point and adjust the point to satisfy angle criterion. Repeat the procedure for few more test points. Sketch the root locus from the knowledge of typical sketches and the informations obtained in steps | through 6.4.68 | Note : In practice the approximate root locus can be sketched from the informattons obtained in | steps 1 through 6 and from the knowledge of typical sketches of root locus. DETERMINATION OF OPEN LOOP GAIN FOR A SPECIFIED DAMPING OF THE DOMINANT ROOTS The dominant poles a pair of complex conjugate pole which decides the transient response of the system. In higher order systems the dominant poles are given by the poles which are very close to origin, provided all other poles are lying far away from the dominant poles. The poles which are far away from the origin will have less effect on the transient response of the system. The transfer function of higher order systems can be approximated to a second order transfer function: The standard form of closed loop transfer function of second order system is, ee: R(s) on joe Weoys+ or | | 8,.%-———H+J0, f1-2 s-plane ‘The dominant poles, s,and s,*, are given by the roots of quadratic factor, s4#2Ca,sto2 = 0. Wat naa? = iors feet 808 oe, aE 238 The dominant pole can be plotted on the s-plane as shown in fig 4.11. In fig 4.11, the right angle triangle OAP, Fig 4.11 : Dominant pole, 8, oa =cos tt To fix a dominant pole on root locus, draw a line at an angle of cos"! with respect to negative real axis. The meeting point of this line with root locus will give the location of dominant pole. The value of K corresponding to dominant pole can be obtained from magnitude condition. Let, K,, be the value of gain at dominant pole s,» Now, x., = Product of length of vectors from open loop poles to dominant pole » “sd Product of length of vectors from open loop zeros to dominant pole Importance of root locus The root locus technique is an important tool in designing control systems with desired performance characteristics. The desired performance of the system can be achieved by adjusting the location of its closed loop poles in the s-plane by varying one or more system parameters. ‘The root locus can be plotted in the s-plane by varying a system parameter (usually gain, K) over the complete range of values. The roots corresponding to a particular value of the system parameter can then be located on the locus or the value of the parameter for a desired root location can be determined from the locus. The root locus technique is also used for stability analysis. Using root locus the range of values of K, fora stable system can be determined. It is also easier to study the relative stability of the system from the knowledge of location of closed loop poles. The dominant roots are used to estimate the damping ratio and natural frequency of oscillation of the system. From ¢ and @, the time domain specifications can be calculated.469 EXAMPLE 4.22 K Aunity feedback control system has an open loop transfer function, G(s) =; Sketch the root locus: siete ds+ 13) SOLUTION Step 7 ; To locate poles and zeros The poles of open loop transfer function are the roots of the equation, s(s*+4s+13)=0. ‘The roots of the quadratic are, s= ~The poles are lying at s= 0, —2 +j3 and—2-)3. Letus denote the poles as P,, P,, and P,. Here, P,= 0, P=-2+j3andP,=-2-33. The poles are marked by X (cross) asstiown in fig 4.22.1. Step 2 : To find the root locus on real axis Thereis only one pole on real axis atthe origin. Hence if we choose any test point on the negative real axis thento the right of that point the total number of real poles and zeros is one, which is an odd number. Hence the entire negative real axis willbe part of rootiocus. The rootlocus on real axis is shown as a bold line in fig 4.22.1 ‘Note : For the given transfer function one root locus branch will start at the pole at the origin and meet the zero at infinity through the negative real axis. Step 3: To find angles of asymptotes and centroid Since there are 3 poles, the number of root locus branches are three. There is no finite zero. Hence allthe three root locus branches ends atzerosat infinity. The number of asymptotes required are three. 180° Angles of asymptotos = ="9° 2a") : g#0,4, -m Heren= =0. 0,1,2,3. When 4=0, When g=1, When q=2, Angles =+“°°*° - 1300°= 760° When q=3, Angles =2420°= 460° ‘Note: [tis enough if you calculate the required number ofangles. Here itis given by tirstthree values ofangies. The remaining values vill be repetitions ofthe previous values. Sum of poles ~ Sum of zeros _ 0-2+/8-2-j8-0_-4_ 139 n-m 3 3 Centroibiliiy aad Root Locus 4.70 The centroid is marked on real exis and from the centroid the angles of asymplotes are marked using a protractor. The asymptotes are dravm as dotted lines as shown in fig 4.22.1. 4.33 centroid] Pe Fig 4.22.1 : Figure showing the asymptote, root locus on real axis and location of poles and centroid Step 4-: To find the breakaway and breakin points K The closed loop] C(s)__ Gis) s(st44s+13) _ K transfer function| R(s) 1+G(s) 4, Ks (s7+48433)+K 8 (8°74 48413) ‘The characteristic equationis, s(s*+4s-+13)+K=0 . SHH4s241354K=0 => K=-shds?135 _, On differentiating the equation of K with respectto s we get, So (35? + 85.413) put Ko ds + ~@s?+8s+13)=0 = (3s?+85+13)=0 yo \ 82 NB 40133 «passing 2x3 Check for K: When, s= 1,33 +]1.6, the value of Kis given by, Ka{o'+4s?+ 13s) = (1.33 +j1.6 + 4 (1.33-+)1.6)* + 13 (1.33 +)1.8)) Se positive and real. Also itcan be shown that when s =—1.33-j1.6 the value of Kis notequalto real and positive. Since the values of K for, s=1.33:)1.8, are not real and positive, these points are not an actual breakaway or breakin points. The roctiocus has neither breakaway nor breakin point. Step 5: To find the angle of departure Letusconsider the compiexpoie p, shown in fig 4.22.2. Draw vectors from all other poles to the pole p, as shown in fig 4.22.2. Let the angles of these vectors be®, and ,. Here, 0, = 180°-tanr'(3/2)= 123.7" ; @,=90" ‘Angle of departure from the complex pole p, = 180°—(0, +8,) = 180° -(123.7° +90") =-33.7°Fig 4.22.3. : Root locus shet-h of, 1+ G(s) = 1+—- K s(6 445413).hie A= Concepis'of Stability Gnd Root Locus ~) 4.78 The angle of departure at complex pole p, is negative of the angle of departure at complexpole A - Angle of departure at pole p, = + 33.7° Mark the angles of departure at complex poles using protractor. Step 6 : To find the crossing point on imaginary axis ‘The characteristic equation is given by, 4452+ 130+K=0 Puts =jo (@)+4(o)+13§0)+K=0 >-Jo*~40?+ 13j0+K=0 On equating imaginary partto zero, weget, | On équating real part to zero, we get, 0+ 1300 -4o?+K = 0 -07 =-130 K = 40? oO =13 > o=tyi3=236 The crossing point of root locusis = j3.6. The value of K atthis crossing point is K= 52. (This is the limiting value of K forthe stability of the system). The complete root locus sketch is shown in fig 4.22.3, The root locus has three branches one branch starts at the pole at origin and travel through negative real axis to meet the zero at infinity. The other two root locus branches starts at complex oles (along the angie of departure), crosses the imaginary axis at-+ 3.6 and travel parallel to asymptotes to meet the zeros atinfnity. EXAMPLE 4.23 x13 = 52 ‘Sketch the root locus of the system whose open loop transfer function is, Gis) = eT Find the value of k sothat the damping ratio of the clased loop system is 0.5. SOLUTION Step 1: To locate poles and zeros ‘The poles of open loop transfer function are the roots of the equation, s(s#2) (s*4) = 0. +The poles are lying at, s= 0,-2,-4. Let us denote the poles as p,, p,, and p,, = Here, p,= 0, p,=-2, p,=-4, ‘The poles are marked by X(cross)as shawn in fig 4.23.1 Step 2 : To find the root locus on real axis There are three poles onthe real axis. Choose a test point on real axis between s = 0 and s =-2. To the right of this point the total number of real poles and zeros is one, which is an odd number. Hence the real axis between s= 0 and s = —2will be a partof rootlocus. Choose a test pointon real axis between s = -2 and s=—4, To the right of this point, the total number of real poles and Zer08 is two which is an even number. Hence the real axis between s = -2 and s = — 4 will not be a part of root focus. Choose test point on real axis to the left of s =~ 4. To the right ofthis point, the total number of real poles and zeros is three, which is an odd number. Hence the entire negative real axis from s =—4 to~cowill be a part of roat locus. ‘The root locus on real axis are shown as bold lines infig 4.23.14.73 Step 3 : To find asymptotes and centroid Since there are three poles the number of root locus branches are three. There is no finite zero. Hence all the three rom: locus branches ends atzeros at infinity. The number of asymptotes required are threo. 20" at) nm Angles of asymptotes Here,n=3andm=0. -.q=0,1,2,3 When g=0, Angles e 280° When q=1, . Angles=+ 1803 - s120° Note: itis enough if you calculate the required number ofangies. Here it is given by first three vatues of angles. The remaining values will be repetitions of the previous values. Centroid = Sum of Roles — Sum of zeyos 0-2 4-0 nm 5 ‘The centroid is marked on real axis and from the centroid the Figazza.s Figaro showing ‘td asinpieny angles ofasymptotes are marked usinga provacior. Theasympioles page Jacus on real ais and location of poles are drawn as dotted lines as shown in fig 4.23.1 cared cad bevoizuay prints Step 4 : To find the breakaway and breakin points K The closed loop| Cis) __Gis)__s (#2) (64) _ transfer function{ R{s) 1+ G(S) 44. s(s+2)(s+4) + K $(8+2)(6+4) The characteristic equations given by, s(6+2)(s+4)*K=0 => — S(s*46848)+K=0 = 59+ 6s?+B5+K=0 K=-s?- 652-85 On differentiating the equation of K with respectto s we get, Kia? ak +1288) SK = (Gs? + 128-8) pur SK. ds «43s? +125 +8)=0 => (387+125+8)-0 (Check for K: When s = -0.845, the value of K is given by, [(-0.845)? + 6(-0.845) + B(-0.845)] = 3.08 Since K, is positive and real for, s =-0.845, this pointis actual breakaway point. When s =-3.154, the value of Kis given by, K == [(-3.154) + 63.154)? + 8-3. 154)] = -3.08 54, this Is nota actual breakaway point. Since K, is negative for, s ‘The breakaway pointis marked on the negative real axis @s shown in fig 4.23.14.74 Fig 4.23.2. : Root locus sketch of , 1+ G(8) =1+ = 1 s(6+2)(6+4)"4.75 Step § : To find angle of departure Since there are no complex pole or zero, we need not find angle of departure or arrival. Step 6 : To find the crossing point of imaginary axis ‘The characteristic equation is given by, s°+687+8s+K=0 Puts=jo Gop + 6fo)’ + 890) +K=0 jo? 80? + 80+K=0 Equating imaginary part to zero Equating real parttd zero -fo? + Bo =0 602K =0 i K=60? =6x8=48 V8 = 228 | ‘The crossing point of root iocus is * 2.8, The value of K corresponding to this pointis K = 48. (Thisis the limiting value of Kfor the stability ofthe system). The complete root locus sketch is shown in fig 4.23.2. The root locus has three branches. One branch star at the pole ats =~ 4 and travol through negative real axis to meetthe zero at mfnity. The other two root locus branches staris ats = and 5=~2and travel through negative real axis, breakaway from real axisats = -0.845, then crosses imaginary axis ats=+)2.8.and travel parallel to asymptotes to meet the zeros at infinity To find the value of K corresponding to ¢ = Given that¢=0.5 Leta= cos" = cos? 0.5=60° Drawa line OP, such that the angle between line OP and negative real axis is 60° (a = 60°) as shown in fig 4.23.2. The meeting point of the line OP and root locus gives the dominant pole, s, Letk,, be value of K corresponding to the points =s, _ Product of length of vector from atl poles to the point, Product of length of vector from ail zeros to the point, s 4xh 1 ‘Note : The length of vectors are measured to scale. od 13x 1.7535 =7.96=8 EXAMPLE 4.24 The open lop transfer functon ofa uy feedbacksystemisgivenby, G(s) = — e oe <7 Sketch he rotlocus ares ofthe system. SOLUTION Step 1 : To locate poles and zeros The poles of open loop transfer function are the roots of the equations, (s°+ 4s + 11) =0. At v4? 4x11 2 ‘The roots of the quadratic are, s = 2+j264