Page 102 - Dynamics and Control of Nuclear Reactors
P. 102

96     CHAPTER 8 Reactor control




                         The final, steady-state value of the error, e(t), when x(t) is a unit step, is given by.
                                                      lim         a
                                                etðÞ ¼   sE sðÞ ¼                       (8.11)
                                                     s ! 0      a + K p
                         Note that the steady-state error is not equal to zero, and may be decreased by
                         increasing K p , the proportional gain. For a¼0.02 and K p ¼1, the steady-state
                         error is e(∞)¼1/51¼0.0196. If K p is increased to 10, then the steady-state error
                         is e(∞)¼0.002.


                           Remarks
                           1. As the proportional gain constant K p increases, the steady-state error decreases. However, the
                              steady-state error is never zero.
                           2. As the gain K p increases, the time constant of the system decreases, and the rate of system
                              response increases. For the open-loop system in the above example (a¼0.02), the time constant,
                              τ¼50s. For the closed-loop system with a proportional gain K p , the time constant is given by
                                                            1
                                                                                      (8.12)
                                                       τ ¼
                                                          a + K p
                              Thus, for K p ¼0.1, the time constant decreases to τ¼8.33s.
                           3. An increased K p results in a faster response of the control system. Since there are limitations on
                              the achievable responses of control devices, K p should not be increased arbitrarily. Furthermore,
                              an increase in K p above a certain value can make the system unstable. The stability margin will
                              decrease with increasing values of K p .

                            All these points must be considered in choosing the proportional gain. The pro-
                         portional controller increases the speed of response of a system. The steady-state
                         error is nonzero.


                         8.3.7.2 Integral controller
                         An integral control is expressed as:
                                                           Z t
                                                    ftðÞ ¼ K i  evðÞdv
                                                           o
                         In the Laplace domain, the control action has the form.

                                                           K i EsðÞ
                                                      FsðÞ ¼                            (8.13)
                                                             s
                         A very important caution to be exercised in using integral controllers is the consid-
                         eration of actuator limits. When an actuator reaches its lower or upper limit (because
                         of actuator saturation), the error may remain large and the integral action will con-
                         tinuously increase the error term. This characteristic is called integral windup and
                         must be avoided by changing the sign of the error using appropriate logic. A large
                         transient may be observed if the actuator reaches its limit.
   97   98   99   100   101   102   103   104   105   106   107