Time-domain analysis and control design of pure integrating plus dead-time processes

• Main Authors: Hong-Yi Zhu, 朱紘億
• Other Authors: Yeong-Iuan Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/33601798557280619260

碩士 === 國立聯合大學 === 化學工程學系碩士班 === 94 === The pure dead-time may be caused by transportation of materials or the actuation of final element. This transportation lag or dead-time is quite common in the chemical process industries where a process fluid is transported through a pipe. Even when pure dead-time is negligible relative to the process time constant, the response of many chemical processes may appear to exhibit dead-time due to approximating higher-order transfer functions by the combination of a lower-order component coupled with some dead-time. The dead-time is quite different from the other transfer functions mainly because it is not a rational function. Consequently, if the dead-time terms exist in the process dynamics, the time-domain solution of a conventional feedback control system is no longer obtainable by classical methods. The methodology of expanding the closed-loop transfer function with the dead-time term in the denominator into an open-loop representation is proposed. For simplicity, the pure integrating plus dead-time model (IPDT) is considered throughout this thesis. The closed-loop time domain series solutions for an IPDT model are developed based on the proposed methodology, and the results are comparisons with the numerical simulations. The P and PD controllers’ parameter tuning settings for set-point change along with the PI and PID controllers’ parameter tuning settings for disturbance input are obtained by minimizing the common performance index (i.e. ISE、ITSE、IT2SE、IAE、ITAE、IT2AE). Some new, simple and fast methods for estimating IPDT model parameters from closed-loop response data are proposed. Methods are presented for estimating the parameters from two points of the step change in set-point response data, similar to that of an open-loop process reaction curve method. The graphical and area-based methods that have been adopted by the classical open-loop test were developed for the closed-loop test as well. Finally, the pure capacity level apparatus was used to demonstrate the proposed methodologies by several experimental runs.