F Me fit row is abiowys Men opfeact. a". “} 4.3 THE ROUTH-HURWITZ STABILITY CRITERION Let the characteristic polynomial of the system with T-F, G@= FQ. be As) = 0,5" $0915" + eS ae We define a table known as the Routh-array and having the form shown below. » An! -2 — Onda-d ant s Gn \ flu gent . ae st be (Anus Ont) = (da Ones? srt A i i a oes 3 y 4 Fat s s ® bbe ae. Geet) = (a, deer) The Routh-aray is produced by using the following rules: a, RULE 1: The first two roles are produced from the coefficient of A(s). When n is even the $" row contains coefficient of the even powers of s. including a. while the S”~' row contains the coefficients of odd powers of s. The situation is reversed when n is odd. RULE2: The entries in the third and subsequent rows are computed as follows: Sn-14n-2-% %n-3 by = Ae eo nnd m—1 Sn-19n-4 ~ 9 %n-5 = tod nn ote b a] — 4-3 ~ n-1%2 : q 2 ee a bian—5 - 4n-1b3 ete - a a, = W2= 42 or, 4 Note that: (i) The strow has 44+! entries when k is even and 4(¢+1) when k is odd so that the S' and S" rows each has only one entry. (ii) Multiplication of all the entries in any row by the same +ve real number is allowed. This point is especially useful in manual computation when it is more convenient to avoid fraction and decimals. (iii.) If the first entry in any row is zero, then the subsequent rows cannot be completed. There are two possibilities as follows.sSE 1 ‘When the first entry is zero but there is at least one none zero entry, the difficulty of prodtcing a Routh array can be avoided by using the polynomial * _ n n-1) on-2 A(Z) = Gz taz Wye eae Ay fF tO For which a Routh array can be produced without difficulty. Example ; If, A(s) = 5° +5‘ +45’ +457 +25+1, Then, the Routh array for this polynomial cannot be produced beyond the Sn row as shown below s 142 “ 141 2 o1 2 ° However for the polynomial ey = Saaz sardgaatazet Tearayls s ran a aad 2 it 2 4 a | z 2 CASE2: ‘When all the entries in a row are zero this is an indication that the system has complex conjugate roots of either of the types indicated in the diagram below. iw. ™. Fig. Type 1 purely imaginary roots at al,2 = Ijwo Fig. Type 2, 2 pairs of complex conjugate roots areSq = —oekJwe. and S34 = oot jwe Cases in which the system has roots of type | are of interest in application to F.B.C system in such cases the entry in the S? row is zeto and the system exhibits substained oscillation of frequency Wo. This frequency can be determined from the formula 2K, where yy and yp are the entries of the S? row ie the S? row is s? Y% 4.3.1 DETERMINATION OF STABILITY FROM ROUTH ARRAY Having produced the Routh array, the stability of the system can be determined by using the Routh stability criterion which states that Ze = NR, where NR denotes the number of changes of sign in the first column of the Routh array. In the example given in previously, NR = 2. This system is unstable, since Ze is not zero but if Zc = 0, the system is stable. Examples Qu 1. Consider a system with open-loop TF Ge) = = Hest as) 385% 4508 +20 Solution: Characteristic equation S* +1053 +3587 +505+24 =0 The Routh array become 135 24 5 0 30 30 24 2 42 a mM s ‘With no sign changes in the first column which indicates that the system is stable.Qn 2. Consider a fifth order system GG) = (45? a11549)(58 42544259 4 452 4135410) ‘The characteristic equation is 5425442534452 413541000 Therefore the Routh array will be BUG We replace the zero in row 3 by 4-16 8-100? 8. Now when e—> 0, the term “c:_ approaches large —ve number, while s=—iy —> +8. Therefore it indicates that the system has 2 changes in sign from + —>; and from - —> +. Hence it has: poles in r.hp. Therefore it is unstable. 43.2 Example ; $0 - ___s . Lc Given a FBC system with CLTF @(s) S(+5rX(1+0.5s)+50 . Determine the maximum vale of 7 for stability. 8 Ose 1 G0) | ___so GG) ~ S0rSeK+05550 s2 OS+r 50 The characteristic equation A(s) is sl S528 As) =0,57 57 +(0.5+1)57 +5450 so 30 Routh Array is For stabilit me my O57 >0,0.5+7>0,& 25=24F . 4 OS+r => r>0,and 05-247 >0 05>24r or 7< 08 >Ur oi <4 os O 1, the closed-loop root becomes complex and diserge to become infinite as K 0 52 Singular Point: Point at which two or mote poles coalese are called singular point. If singular point occurs in the real axis it is also known as break points. 4.4.1 THE ROOT LOCUS METHOD (INTRODUCED BY EVANS) The location of SISO system's close-loop poles have an important effect on the relative stability and the transient response of the system. The system’s open-loop gain is often subject to drift and variation. Therefore it is important for a designer to have an exact understanding of the location of the root of the characteristic equation as the open-loop gain varies. The root locus method is the scheme to do this. The root locus is a graphical description of the sensitivity of system,s poles with respect to a parameter, be it ‘open loop gain or any other one. Consider the simple case “f K}I{6@) l H(s) Assuming that the T.Fs Gp(s) and H(s) are known, K is the loop gain, unknown, and can be manipulated. ‘The roots (i.e clos-loop poles) of this system are given by the solutions to the equation 1+kGp(s)H(s)=O0_ (1). The root-locus is the path traced out by these roots as the loop gain (or the parameter under consideration) varies from 0 to infinite Example 1. Consider a system for which Gp(s)=sghq, H(s)=1. ‘The characteristic equation for the roots is k 1+—4_. S(s+2) ie $?+25+k=0 The roots are then given by Sy =-14+I-k Sy =-1-vI-k Generally, § =8+ jw = for k=0, 8,43.1 lL 4 RULES FOR ROOT LOCUS SKETCHING Symmetry of Root-locus: Since the complex roots comes in conjugate pairs, the root-locus is Symmetrical with respect to the real axis. Number of Branches: If the order of the system is N, the root-locus has N branches, Each branch start at the open loop- gain (i.e when K = 0), The branches are normally symmetrical about the real axis as mention above i.e for any known real axis branch, there is another which is a mirrior image of the first one about the axis, Real axis sub-branch (or segment) Any portion of the real axis is part ofthe locus ifthe total number of poles and zeros to its right is an odd integer. Asymplotic Behaviour When some branches goes to infinity, the approach along a symplotic centered. E.g branches terminates at the M zeros (open loop) where N> m, the remaining n ~ m branches goes off to infinity either along or asymplotic to a family of n —m straight lines which radiates from the point. S=5 = {E poles of G(s) -& zeros of G(s)}/n-m _@Z)-F) n-m at angle given by, = (2L— 1180, L=1,2,3..n-m to the real axis, Determination of break points(Break-in or away) 1NG@ & A break point is a point where 2 ot mote branches of the root locus intersects. 4 ‘Break-In Point Le it is location of multiple roots, all break points are given by NGS) D'(s) = N'(s) Dis) = 0. ‘Where the prime denotes differential watt the complex vatiable §, Note however that every solution of the break point equation does correspond to a break point. The break points are the solutions of the above equation which lie on the real axis, Segment of the root locus joining two adjacent poles or 2 adjacent zero 5. Imaginary axis crossing: There are two methods that can be used to locate these crossings, The first method is the solution of ‘the polynomial equation obtained by substituting S = jw into the equation D(s) + KN(s) The second method uses the Routh-array to determine the value of K for marginal stability and the corresponding angular velocity w of sustained oscillation. 6. Angle of Departure or Entry: Defi; The angle of departure (Entry) from (to) is the angle which the tangent to the root-locus branch of the pole (zero) makes with the positive real axis. ‘The angle of departure # from a simple pole located at S = -P is given by § = 180-(0+¢) Where 6 is the sum of the angles of the vector drawn from all the other poles to the point, S =-P. 4 iis the sum of the angles of the vectors drawn from all the zeros to the same point. 6=6,4+0, +9, Similarly, the angle of entry © is a simple zero located at S = -z and is given by « =6-¢-180" Where 8 is the sum of angles of the vectors drawn from all the poles to the point $= -Z ¢ is the sum of the angles of the vectors drawn from all other zeros to the same point. Example 2: Consider the open-loop TF 8_ ENG) _ K+!) _ K(s4+1) oa Ds) 52495 5(5+2) for H=1,the CLTF ts G=—KGs) 82 4.254 k(S+!) By factorization of the denominator, we get the close-loop poles to be =-1 | 5,=-FQ++ lade Example 3:'The fig. Below shows the block diagram representation of a temp. control loop for an isothermal reaction process. Sketch the root locus for the system indicating clearly the important feature of the plot. Describe fully what the plot shows about the stability and transient behaviour of the system. And determine approx. the value of the controller gain Ke for which the damping ratio of the dominant pole is 0.6 Controller Heater Process ikp 28 08 _ “ 25+4] [56-1 4 SH Transducer Solution: The OLTF isKN(s) _ 2Ke Dis) 2541 (55-1541) _ 0.2ke © (5 +15 +0.5)(s — 0.2) => Ms)=1 D(s) = (S + 1S +0.5)(5 0.2) “M =0,n=3 and k =0.2ke pole located at 5 = p= b23 ‘Since poles are located at S = -Pi => Pi=-0.2, Pr2=0.5, P3=1 Plotting ) Real axis segments are from S$ = 0.2 to -0.5 and from $ =-1 to $= 00 2, P, (i) Asymtotes : The center is at Seas ‘a 180 , 180 , 180 180 |, 180 5 180 3 3 3 The angles are ° . 60, 180, 300. (ii). Break point location. One break point is expected atS =o, between $= 0.2 and S =-0,5 The break po int equation is N(s) D'(s)- N'(s) D(s) =0 D\(s) = 30} + 2.60, +0.2=0 “0, = —0.085 Apply magnitude criterion to get the value of K at break point D(oy) N(o,) = 0,915 x 0.415 x 0.285 1 K = 0,108 But K =0.2ke w). Imaginay axis crossing. The characteristic polynomial for the roots is D(s)+KN(s) =5° +1.35? +0.25+k-0.1=0 The Routh array s 1 02 s k-011 s For marginal stability ‘At the imaginary axis crossing {0.36-k _ 13 =036 ie ke=18. Put value of K in S? term or expression to obtain the point imaginary axis crossing. ie 1.38? +(K-0.1)S =0. 1.38? +0.265 =0. The crossings are at S,, =V-02 S,. =4j0.45Discussion of Stability and Transient Ressponse For 0 s Ke <0.5 one root lis inthe RHP and system is them unstable, For 0.5 < Ke < 0.54 all roots are real and in the LHP> Therefore the system is stable and the step response is over damped. For 0.54 < Ke < 1.8 two of the roots are complex conjugates, the third root is real, All roots are in the LHP and step response is underdamped. For Ke> 1.8, the two complex roots are in the RHP and the system is unstable. Now the dampin ratio € = 0.6 => $= Cos 12253 ‘Then we draw a line to meet the loci at 53 degree and substitute the value of S at the point intersection in the characteristic equation to obtain the value of K.