Control of Continuous Stirred Tank Reactor Using Artificial Neural Networks Based Predictive Control

In this work, a Neural network based predictive controller is analyzed to a non linear continuous stirred tank reactor (CSTR) carrying out series and parallel reactions: A -> B -> C and 2A -> D. In the first step, the neural network model of continuous stirred tank reactor is obtained by Levenburg- Marquard training. The data for the training the network is generated using state space model of continuous stirred tank reactor. The neural network model of continuous stirred tank reactor is used in model predictive controller design. The performance of present neural network based model predictive controller (NNMPC) is evaluated through simulations for servo & regulatory problems of CSTR. The performance of neural network based predictive controller is found to be superior than conventional PI controller for setpoint tracking problems.