Question 14.3: Figure 14.14 shows a block diagram of a feedback control sys...

Figure 14.14 shows a block diagram of a feedback control system in which the transfer function is the same as the one used in Examples 13.6 and 13.8 and is subsequently repeated. For the purposes of this example, assume that the transfer function represents the dynamics’ large-capacity material-handling system in which the input signal u(t) is the desired position and the output signal y(t) is the actual, measured position of the payload, measured in meters:

T_{P} (s) =\frac{2}{6s^{2}  + 11s^{2}  + 6s + 1} .

Design a controller to meet the following specifications:

_ e_{ss} for step input = 0,

_ no overshoot for step input,

_ response to step input as rapid as possible.

The Blue Check Mark means that this solution has been answered and checked by an expert. This guarantees that the final answer is accurate.
Learn more on how we answer questions.

The simplest controller is the proportional control described in Subsection 14.5.3:

T_{C}(s) =k_{p}  .

The open-loop transfer function, combining the controller and the system under control, is

G(s) = T_{C}(s) T_{p}(s)=\frac{2k_{p} }{6s^{3}  + 11s^{2}  + 6s + 1}  .

For a unit step input, the steady state error is determined according to Eq. (14.10):

e_{ss} =\underset{s\rightarrow 0 }{\lim } \frac{1(1/s)}{1 + G(s)H(s)} .

e_{ss} =\underset{s\rightarrow \infty }{\lim } \frac{1}{1 + G(s)} =\frac{1}{1 + 2k_{p} }  .

This result indicates that, regardless of the value of kp, the specification for steadystate error will never be met. Therefore a different control structure is desired. In Subsection 14.5.3, the PI control law is presented. The main characteristic of this control law is that it increases the type of the system and hence improves steady state performance. The original system is a type 0 system. If the controller is a PI controller, then the combined system will be type 1, and, according to Table 14.1, the steady-state error to a step input will be zero. To achieve the desired transient response, further analysis is required.

The controller transfer function is

T_{C}(s)=k_{p} \left(1+\frac{1}{T_{i}s } \right) ,

and the combined open-loop transfer function is

G(s)=\frac{\frac{2k_{p} }{T_{i} }{\left(T_{i}s+1\right) } }{s\left(6s^{3}  + 11s^{2}  + 6s + 1\right) }  .

As a starting point for the values of kp and Ti, the Ziegler–Nichols tuning rules are used. Figure 14.15 shows a step response of the original system, computed with MATLAB’s step command. Indicated on the plot are the values of R and Lfor this response, as defined in Fig. 14.12. From the plot in Fig. 14.15, the following characteristics can be measured:

L = 1.5 s,

R = 0.27 m/s.

The Ziegler–Nichols tuning rules for PI controllers suggest the following values for gains:

k_{p} = 2.2,

 

T_{i} = 4.95 s.

To test whether or not the second design criterion is satisfied with these values, the following steps in MATLAB will generate a closed-loop step response:

14.12
14.15
14.16
14.17
14.18
14.19
The 'Blue Check Mark' means that either the MATLAB code/script/answer provided in the answer section has been tested by our team of experts; or the answer in general has be fact checked.

Learn more on how do we answer questions.

>> kp = 2.2
>> ti = 4.95 % set the gain values
>> num = 2*kp/ti*[ti 1];
>> den = [6 11 6 1 0];
>> tfol = tf(num, den) % establish the open–loop
%transfer function
>> tfcl = tfol/(1+tfol); % compute the closed–loop
%transfer function
>> tfcl = minreal(tfcl) % cancel poles and zeros
(see Section 11.7)
>> step(tfcl) % compute and plot step
%response

Figure 14.16 shows the step response for the tuning parameters computed with the Zeigler–Nichols tuning rules.

The oscillatory behavior of the closed-loop system clearly violates the second design constraint, that there be no overshoot. Note, however, that the first design constraint, zero steady-state error to step input, is clearly met. Oscillatory behavior can be mitigated by lowering the proportional gain factor. Through trial and error and use of MATLAB’s commands to reevaluate the closed-loop transfer function and step response for various gain values, it was discovered that when kp is reduced to a value of 0.48, the step response shows no overshoot, as demonstrated in Fig. 14.17.

Finally, the third design criterion, that the response be as rapid as possible, is more difficult to assess. Howcan it be determined if the response can be any quicker? Figure 14.15, for example, shows that the open-loop system can respond to a step input in approximately 20 s. Experience with feedback control dictates that closed-loop response can nearly always be more rapid than open-loop. Therefore the response in Fig. 14.17 seems a bit sluggish.

The next step in the design process is to consider a PID control architecture, as described in Subsection 14.5.5. The new combined open-loop transfer function is

G(s)=\frac{\frac{2k_{p} }{T_{i} }{\left(T_{i}T_{d}s^{2} +T_{i}s+1\right) } }{s\left(6s^{3}  + 11s^{2}  + 6s + 1\right) } .

The Ziegler–Nichols tuning rules suggest the following starting place for the PID parameters:

k_{p} = 2.96,

 

T_{i }= 3.0 s,

 

T_{d} = 0.75 s,

which lead to the closed-loop step response seen in Fig. 14.18. Note that the system is somewhat less oscillatory than the first attempt with a PI controller, but the overshoot is very high. As with the PI controller, the response can be improved by adjustment of the three control parameters. The overall gain kp must be reduced and the derivative term Td increased to eliminate the overshoot. In addition, decreasing the value of Ti strengthens the integral action and helps drive the system to its steady-state value. Although there are literally an infinite number of parameter combinations that will bring this system in compliance with the first two design goals, the following set of parameters result in the response seen in Fig. 14.19:

k_{p} = 1.25,

 

T_{i} = 1.6 s,

 

T_{d} = 1.8 s.

In comparing the step responses obtained with the PI controller (Fig. 14.17) with those obtained with the PID controller, it is important to note the change in time scale on the two plots. The PID controller offers a far superior performance in the speed of response and hence is a better choice in light of the third design criterion.

Related Answered Questions