Detection and compensation of dead time in chemical processes

Abstract

Dead time is commonly found in chemical processes. Signi cant amount of dead time can lead to unsatisfactory and unstable process operation. Therefore, dead time estimation and compensation is very important for processes with long dead time. In this study, a new dead time compensation technique for stable processes has been developed. The estimation of dead time is developed using a numerical technique. Numerical Integration is used to detect dead time from step response data of a process. The method was evaluated and validated in simulation. Later, this method has been used to detect dead time from step response data of a real process. The dead time estimation technique performs well up to a certain amount of noise present in the system. The dead time compensation method is developed by modifying the classical dead time compensator structure. Two new lters are introduced in the traditional compensator. New rules for tuning the controller and the parameters of the lters are developed. The performance of the newly proposed dead time compensator has been compared with the most recently appeared dead time compensator in the literature using both simulation and experimental studies. A pilot scale two tank heating system is used for experimental evaluation of the proposed dead time compensator. The dead time compensator outperforms the most recently appeared dead time compensator. The reliability analysis is performed to test the robustness of the proposed method. The reliability analysis shows that the newly developed dead time compensator can handle some uncertainties in the estimation of dead time and process model.