摘要 在一般的化工程序上,反向應答程序可算是較特殊的程序,常發生在如蒸餾 塔, 氣體吸收塔及鍋爐等化工廠中重要的設備上. 本研究乃針對一般常用 之反向應答程序模式,利用Powell搜索法來找出在不同模式參數下之PID與 PI控制器之最適調諧參數; 為了更廣泛且多樣的描述程序及控制器,而把 程序模式及控制器無因次化,然後再對無因次化後之程序模式作閉環路回 饋控制系統之模擬,基於最小IAE或ISTE之目標函數下, 利用Powell(1964) 搜索法則來獲得最適之控制器調諧值. 再利用類神經網路來做 最適調諧 關係之數據迴歸. 經由這些最適調諧關係之模式,便可由簡單之電腦計算 而得到PID與PI控制器之調諧值. 最後,由電腦模擬之結果與比較得到證 明,本研究之調諧方法是可行的. Abstract The process with inverse response is one kind of special chemical processes. A calculation method of PID controller tuning for thesecond-order process model with inverse response and time delay ispresented in this study. Optimum PID controller tuning data based onthe model and minimum IAE or ISTE criterion were obtained via Powellsearching technique. Then these data were correlated by ArtificialNeural Networks into the data regression model. Thus PID controllertuning based on the model can be easily obtained by calculation of thedata regression model. Simulation with reset-feedback PID algorithm isalso presented in the study. In addition, this study is practical andconvenient by the way of the examples simulation study.
