Real-Time Implementation and Simulation of Comparative Fuzzy-PID Controllers for a Fluid Level System

This paper develops and compares three control strategies for a water level process: a conventional proportional-integral-derivative (PID) controller, a PID controller embedded in a Smith Predictor structure, and a Fuzzy PID controller. Experiments are carried out on a TE۳۳۰۰ process trainer, where a differential pressure transmitter reports the level, and an electro-pneumatic valve sets the inlet flow. First-order plus time delay (FOPTD) and second-order plus time delay (SOPTD) models are fitted to step-response data recorded on the setup, and the second-order model serves as the common plant against which all three controllers are evaluated. The conventional controller is tuned by the internal model control (IMC) method, and the inner controller of the predictor is independently retuned for the delay-free plant, a step that keeps the comparison focused on control structure rather than on inherited tuning. The Fuzzy PID controller is built on a zero-order Sugeno inference system, and three distinct control surfaces are constructed and tested. All three controllers are subsequently deployed on the physical setup and run in real time through an Open Platform Communications (OPC) interface, which introduces a delay of approximately five seconds between the controller and the plant. The paper reports what this delay compensation is worth in practice and examines how the geometry of the fuzzy control surface balances the error accumulated during a transient against the speed of correction and the resulting valve activity.

کلیدواژه ها:

Fuzzy PID control ، Smith Predictor ، System identification ، Control Surface geometry ، Time delay ، Water level system