Iterative Learning Control Method Based On Neural Network Abs Resin Polymerization Process

Iterative learning control is a new intelligent control algorithm in these two decades. In industrial process control, most complex industrial process has some repeatable properties, which enables controller designed and improved online by iterative learning, accumulating knowledge from controlled plant using a learning mechanism. In other words, on-line learning, on-line control and improvement of control system are integrated in an algorithm and realized by the repetition of industrial processes. Iterative learning control has strong advantage in solving the uncertainty problem which is caused by nonlinearity and external disturbance of the plant. It improves the system performance gradually by completing the system experience in control process.Neural network has the satisfactory capability of approximating any nonlinear mapping, furthermore, it can learn and adapt to the dynamical properties of uncertain system. So neural network based control system has fairly strong adaptability and robustness. A new iterative learning control algorithm optimized by using neural network is proposed in this paper. This algorithm optimizes the parameters of the controller by using the computational ability and the satisfactory capability of approximating any nonlinear mapping of the neural network. Its basic principle is that iterative learning control learns the property of the plant during every iterative learning trail to track the desired output trajectory; meanwhile, the RBF neural network optimizes the plus of P-type iterative learning criterion. After the trail, RBF neural network finds an optimal learning plus according to the output information to replace the quondam one in order to accelerate the learning speed to reduce the learning trail. A PD-type controller as a feedback compensating controller is added to improve the robustness of the controlled system. The optimal iterative learning control algorithm is used to the temperature control of ABS resin polymerization reactor. The simulation result indicates that the algorithm can overcome the nonlinearity and lag of the ABS resin polymerization process, thus possess better robustness and control performance.