Solutions 1: Revision¶
Solutions¶
The closed-loop characteristic equation is given by the formula 1+GH(s)=0. Hence:
s+2=0; s2+3s+3=0; s+1=0.
s(s+1)(s2+s+1)+K(s+2)(s+1)=0.
The stability problem may be solved by Routh-Hurwitz. Char. eqn. is:
1+G(s)=(s+1)(s−5)+K(s+2)=0s2+(K−4)s+(2K−5)=0Routh array is:
s212K−5s1K−40s0−|12K−5K−40|K−4=2K−50First column of Routh array is [1,K−4,2K−5]. There is one sign change (one untsable pole) if 2K−5<0, ie. 0<K<5/2. There are two sign changes (two unstable poles) if K−4<0, i.e. K<4.
Hence system is untsable for 0<K<4, and stable for all K>4.
See Figure l. α=cos−1ζ. ωn is given by the radius of the circle with origin 0+j0 which passes through the point on the root-locus. For a given value of ζ use trigonometry to determine the corresponding value of ωn and hence the closed-loop pole locations. For a given point s0 on the root-locus K is given by the formula:
K=1|Go(s)|=n∏i=1|so−pi|m∏i=1|so−zi|.The terms |s0−pi| and |s0−zi| are the lengths of the lines drawn from the point s0 on the root-locus to the pole pi and zero zi respectively.
ζ |
ωn |
K |
closed-loop poles |
---|---|---|---|
0.0 |
∞ |
∞ |
s=−2±j∞ |
0.1 |
20 |
307 |
s=−2±j19.9 |
0.5 |
4 |
13 |
s=−2±j3.46 |
1.0 |
2 |
1 |
s=−2,−2 |
The constraint 2<ωn≤10 implies 1<ζ≤0.2 and 1<K≤97. For a peak overshoot Mp≤0.2 we require ζ≥0.48 which implies that ωn≤4.17 rad.s-1 for this system (because σ=ζωn=2). The constraint on rise time can only be satisfied if ωn≤0.56 rad.s-1. Hence the two constraints cannot simultaneously be satisfied by adjusting the gain K, since any valid roots would have to lie on the. root-locus.
The root-locus is shown in Figure 2. It verifies the stability result of Question 2. For ideal damping, the closed-loop poles must lie on line OA drawn at an angle of 135° to the positive real axis. The co-ordinates of (and hence the gain required to locate the dominant poles at) this point may be found by trigonometry. However it is easier to equate the coefficients of the model char. eqn. s2+√2ωns+ω2n=0. If this is done you should find K=9.16 poles are: s=−2.58±j2.58.
- E(s)=R(s)−C(s);C(s)=1G(s)1+G(s)H(s)R(s)
hence
E(s)=1+G(H−1)l+GHR(s)=s(1+sT)+(h−1)s(1+sT)+hR(s)Final value, when R(s)=1/s is: $e(t)|t→∞=sE(s)|s→0=s{s(1+sT)+(h−1)s(1+sT)+h⋅1s}s→0=h−1h$
hence system is Type 0. When h = 1, system is Type 1.
Use the relationships given in the handout for Tr, Ts, and Mp to relate these paremeters to the s-plane. Additional constraints are: b0=a0 to give zero step-error, b1=2σ−0.0lω2n to satisfy the ramp error requirement. A solution is:
b1=24.28
b0=400
a1=28.28
a0=400
Note: ignoring the zero, the characteristic equation required is s2+10s+400 which gives ωn=20, ζ=0.5 (σ=ζωn=10). The zero has the effect of increasing the overshoot and so additional damping is required - you will need to iterate a few times to get a value of a1 that yields a satisfactory solution, see Figure 3. (This solution has ζ=0.707).