Model identification of non linear systems using soft computing technique

In Chemical and process industries modeling of non linear systems posses a major challenging task to design engineers due to multivariable process interactions. Innovative technology for process identification is on high demand. A model identification using Neural Networks and ANFIS for the nonlinear systems in series is proposed and designed using conductivity as a measured parameter and flow rate as manipulated variable. Real time experimental data of the non linear system is used to train the neural network by back propagation training algorithm and ANFIS using Matlab. The identified model using various estimators is compared with the actual process model. The error analysis was also performed. Neural Model Predictive Controller controller (NMPC) is designed to control the level. Performance of NMPC compared with traditional PID controller.