Abstract—Continuous Stirred Tank Reactor (CSTR) presents a significant role in the most of chemical process industries. The inherent nonlinear dynamic of CSTR results in various difficulties of stabilizing and controlling of the process. Likewise, functional changes with elapsing of the time and different disturbances associated with real environment are the other hints which must put into perspective. In this paper, a novel strategy based on the combination of fuzzy gain scheduling and a socio-political optimization algorithm, imperialist competitive algorithm, has been employed to design an intelligent Fuzzy-PI controller so that it surmounts the problem of system’s parameter variation throughout the controlling of CSTR temperature. The controller has two inputs which are error and derivative of error and two outputs that are the proportional gain and integral gain. The controller is updated continuously to track the changes which are occurred in the system. To obtain the high performance control, imperialist competitive algorithm was employed to determine the optimum membership functions of the input and output variables. The performance of the proposed controller has been consider based on the Sum of the Square Error (SSE) and Integral Absolute Error(IAE) Criteria. In addition, conventional PI controller and classic online PI controller were applied to the system for the performance comparison. The results clearly indicate the robustness of Fuzzy-PI controller in dealing with uncertainties throughout the tracking of the reference signal. Fuzzy-PI controller also embodies an outstanding accuracy of tracking which leads to the least tracking error criterion.
Index Terms—Fuzzy-PI controller, Imperialist Competitive Algorithm (ICA), Gain Scheduling.
A. M. Yazdani is with the Faculty of Electrical Engineering, Universiti Teknologi Malaysia, 81310 Skudai, Johor Malaysia; (e-mail: firstname.lastname@example.org).
M. A. Movahed is with the Faculty of Electrical Engineering Universiti Teknologi Malaysia, 81310 Skudai, Johor Malaysia; (e-mail: email@example.com).
S. Mahmoudzadeh is with the Department of Computer Engineering, Islamic Azad University, Shabestar branch, Iran (e-mail: firstname.lastname@example.org).
Cite: A. M. Yazdani, M. A. Movahed, and S. Mahmoudzadeh, "Applying a Novel Fuzzy-PI Controller on the Model of Continuous Stirred Tank Reactor," International Journal of Machine Learning and Computing vol. 1, no. 2, pp. 218-223, 2011.