A New Flatness Pattern Recognition Model Based on Variable Metric Chaos Optimization Neural Network

摘要

Aim at the problems occurring in a least square method model and a neural network model for flatness pattern recognition, a new approach of flatness pattern recognition based on the variable metric chaos optimization neural network is proposed to meet the demand of high-precision flatness control for cold strip mill. The model is shown to fit the actual data pricisely and to overcome several disadvantages of the conventional BP neural network. Namely: slow convergence, low accuracy and difficulty in finding the global optimum. A series of tests have been conducted based on the data of the actual flatness pattern. The simulation results show that the speed and accuracy of the flatness pattern recognition model are obviously improved.