Performance analysis of PID controller, fuzzy and ANFIS in pasteurization process

Pasteurization engine uses a heat exchanger to heat the milk up to 71.6 °C. Therefore, the suitable type of controller for pasteurization process must be obtained. In this paper, there are five types of controller to be compared. These controllers are a PID controller tuning Ziegler Nichols, PID tuning Cohen Coon, Cascade PID, Fuzzy Inference System and Adapative Neuro Fuzzy Inference System (ANFIS). The system is also got disturbances that is caused by changes in flow of the product and changes of a step and impulse input for the system. This controller is created in MATLAB, which is connected to the PLC using OPC. The output of MATLAB will determine the control signals which serve as an input for the heater. It is shown from the data that the system that is controlled using ANFIS has an overshoot is only 0.27 percent, which is the smallest overshoot compared to other controllers. The required power for heater is only 0.48 kW which is the smallest one compare to other controllers. The MSE (mean squared error) value after the rise time had been reached is 0.114. The MSE when the system is given the flow disturbance is 1.86. When the system's step input is changed, the system that is controlled by ANFIS has a smaller overshoot and MSE compared to when the system is being controlled by PID cascade. However, when exposed to impulse disturbance ANFIS system controlled is highly reactive and has a lower performance than PID Cascade. Overall, it can be concluded that the performance of the ANFIS controller is better compare to other controllers.