Modeling, Identification, and Experimental Implementation of PID, Smith Predictor, and Fuzzy-PID Controllers for a Water Level Process

This paper presents the modeling, identification, and control of a water level process using three different controllers: a conventional PID controller, a Smith Predictor based PID controller, and a Fuzzy PID controller. The study was carried out on a real laboratory liquid level setup based on the TE۳۳۰۰ process trainer, in which the process variable was measured by a differential pressure transmitter and the control signal was applied through an electro pneumatic valve system. To establish a suitable model for controller design, the process was identified from experimental data and represented using first order plus time delay and second order plus time delay models. A conventional PID controller was first designed using the internal model control tuning method and implemented on the process. Since the overall system exhibited a noticeable time delay due to communication and implementation conditions, a Smith Predictor structure was then employed to reduce the adverse effect of delay on closed loop performance. In the next step, a Fuzzy PID controller was developed using a Sugeno fuzzy inference system with different membership functions. Both theoretical analysis and experimental implementation were used to evaluate the controllers under setpoint tracking and disturbance conditions. The results show that using the Smith Predictor improves the performance of the conventional PID controller in the presence of delay, while the Fuzzy PID controller provides greater flexibility and can achieve improved control behavior depending on the selected fuzzy surface and tuning parameters.

کلیدواژه ها:

: Water level process ، System identification ، PID control ، Smith Predictor ، Fuzzy PID control ، Time delay compensation ، TE۳۳۰۰