Abstract:
This research concerns the design and implementation of fuzzy model-based controller for non-linear system. The process of study is nitric acid and sodium hydroxide neutralization system. The modeling of the process was using Takagi-Sugeno method. The partition of the domain of the process is devised 2 cases namely, c-mean clustering and manual partitioning. It was found that the model derived using of c-mean clustering method make a better prediction for the process behavior than the manual partitioning method. Then c-mean clustering model was used for constructing internal model controller with multi-region control scheme. For IMC, the fuzzy controller was obtained by inverting fuzzy model, which the controller output was the combination of all region outputs. The controller was used to control pH system. Investigated in this research were conducted in 3 cases. Firstly changing set points control system from 3 different values 5, 7, and 9 to new values (+0.5), secondly disturbing the system by changing flow rate and concentration of acid +10%, thirdly testing the model mismatch by changing equilibrium the parameter of the model namely, the constant of ion in water. The performance is compared with a PI gain-scheduling controller. The fuzzy controller performed better than a PI gain-scheduling controller in all case.