The application of RBF neural network in the complex industry process

The complex industry process always has the characteristics of uncertainty, nonlinear, large time delay, strong coupling and so on. So it is difficult to establish an online control model. In order to overcome the impact of these factors on modeling the complex industrial system, this paper proposes a RBF neural network control method based on data and the optimization strategy through iterative learning control. According to the actual history data, the system output of the next iteration with RBF neural network is optimized, so that the error between the model and the measured value reduces in the iterative process gradually. The output track is close to the ideal track. Simulation results about the synthetic ammonia decarbonization process have shown that this method has better performance than fuzzy neural network and more effective to control the complex industry system. The control precision and the response speed of the system are improved obviously. It can provide an effective technical approach to solve a class of complex systems modeling and optimization control problems.