Model and Control of a Multi Input-Multi Output Nonlinear System Using Soft Computing

Soft computing researches focus on using artificial intelligence tools (Artificial Neural Networks, Fuzzy Technology and Genetic Algorithms…) simultaneously in a hybrid-supplementary frame to solve real problems. In this paper, two soft computing powerful tools (Artificial Neural Networks and Fuzzy Logic) have been implemented for modeling and controlling of a nonlinear system. The studied system is a multi component distillation column that is located in Tehran refinery. At first, the system is modeled with recurrent neural network using obtained data from column, then, optimum network is selected based on root mean square error (RMSE) criterion. Five fuzzy controllers are designed for handling system requirements. Results show that soft computing methods can be utilized for modeling and controlling nonlinear systems with high flexibility and accuracy.