Research on Hydraulic Looper System Modeling and RBF Neural Network Decoupling Control

With the development of industry, in the production process, there are increasingly higher requirements for product accuracy and performance. However, there are serious coupling and strong uncertainty in complex engineering, especially in multivariable systems, the design is more complicated. Multivariable systems can choose a variety of algorithms to optimize parameters of complex models, including particle swarm optimization algorithm, genetic algorithm, and ant colony algorithm. This article introduces the RBF neural network based on the improved weed optimization algorithm into the coupled control system. It introduces the RBF neural network optimized by the improved weed algorithm into the coupled control system. On the basis of the state space dynamic model, using the two advantages of the weed algorithm's strong population competitiveness and wide spatial distribution range, the accuracy of the perceptron of the RBF neural network is accurately optimized, and finally the actual engineering is better controlled. It overcomes the problems of the basic weed algorithm (IWO) that are easy to fall into the local optimum, low convergence accuracy, and slow convergence speed. Finally, comparison are made with other optimization algorithms. The simulation results show the effectiveness of this method. The control scheme has high robustness to meet certain external disturbance coupling, and at the same time minimizes the relationship between the coupling variables, and the control effect has been significantly improved.