Page 499 - Mechanical Engineers' Handbook (Volume 2)
P. 499

490   Closed-Loop Control System Analysis


                                                               2
                                                        J    e (t) dt                         (98)
                                                            0
                             A system designed by this criterion tends to show a rapid decrease in a large initial
                          error. Hence the response is fast and oscillatory. Thus the system has poor relative stability.
                                                              2
                          The ISE is of practical significance because   e (t) dt resembles power consumption for some
                          systems.

           7.6 Integral of Time-Multiplied Absolute-Error (ITAE) Criterion

                          According to the ITAE criterion, the optimum system is the one that minimizes the perform-
                          ance index:

                                                       J    t e(t)  dt                        (99)
                                                            0
                             This criterion weighs large initial errors lightly, and errors occurring late in the transient
                          response are penalized heavily. A system designed by use of the ITAE has a characteristic
                          that the overshoot in the transient response is small and oscillations are well damped.

           7.7  Comparison of Various Error Criteria 2
                          Figure 34 shows several error performance curves. The system considered is

                                                     C(s)       1
                                                                                             (100)
                                                     R(s)  s   2 s   1
                                                            2
                             The curves of Fig. 34 indicate that     0.7 corresponds to a near-optimal value with
                          respect to each of the performance indices used. At     0.7 the system given by Eq. (100)
                          results in rapid response to a step input with approximately 5% overshoot.
                             Table 5 summarizes the coefficients that will minimize the ITAE performance criterion
                          for a step input to the closed-loop transfer function 5
                                               C(s)             a 0
                                                                                             (101)
                                                     n
                                               R(s)  s   a n 1 s n 1          as   a 0
                                                                       1
                                   2
                             Table 6 summarizes the coefficients that will minimize the ITAE performance criterion
                          for a ramp input applied to the closed-loop transfer function:
                                               C(s)          bs   b 0
                                                              1
                                                                                             (102)
                                                     n
                                               R(s)  s   b  s n 1          bs   b
                                                          n 1          1    0
                             Figure 35 shows the response resulting from optimum coefficients for a step input ap-
                          plied to the normalized closed-loop transfer function given in Eq. (101), for ISE, IAE
                          (integral absolute error), and ITAE.
           8 SIMULATION FOR CONTROL SYSTEM ANALYSIS
                          Often in control system design the governing differential equations are reduced to a form
                          that facilitates the shaping of a controller. The model order reductions involve things such
                          as linearization, neglecting fast dynamics, parasitics, time delays, and so on. Once a con-
                          troller is synthesized for a system, after making simplifying assumptions, it becomes nec-
   494   495   496   497   498   499   500   501   502   503   504