Time Delay Dynamic System Model Identification of Hot Rolling Mill Based on an Improved PSO Neutral Network

摘要

Combining improved particle swarm optimization (IPSO) with neural network, we propose a model identification method for time lag and nonlinear of hot rolling mill thickness control system. The method is used for automatic gauge control identification of hot rolling mill and the result is applied to Smith predictor. The result of simulation show that the model based on IPSO neutral network is used for Smith predictor control can avoid hypothesis of thickness control system and solution nonlinear equation. The change of thickness of hot rolling mill can be traced accurately, thus the system showing its favorable dynamic properties and precision.