基于混沌优化支持向量机的板形预测与优化

针对带钢热连轧中板形控制问题, 提出了一种基于最小二乘支持向量机模型的预测和优化算法. 在分析最小二乘支持向量机数学预测模型的基础上, 提出了一种改进的变尺度混沌优化方法, 结合实数编码遗传算法, 进行最优模型参数的搜索. 利用在线实测数据对模型进行训练并进行带钢平直度指数的预测, 并对模型输入参数中的控制参数进行优化以实现板形控制的优化. 仿真结果表明, 与BP神经网络相比, 板形预测精度得到提高, 平直度指数优化约40%, 为进一步提高热连轧板形控制精度提供了一种新的有效方法.

Flatness Prediction and Optimization Based on Chaotic Optimized Support Vector Machine

Aiming at the exact shape control problem in hot strip rolling mills, a prediction and optimization algorithm based on the least square support vector machine is proposed. The prediction model is mathematically analyzed first, then an improved multi-scale chaotic optimization algorithm combined with the real-coded genetic algorithm is proposed to optimize the model parameters. Using the on-line data obtained from the factory, the model is trained and the flatness value is predicted. The control parameters among those input parameters of the model are also optimized to improve the shape control. Simulation results show that the prediction accuracy is improved compared to that achieved by the BP neural network, the flatness value is decreased by 40%. This proposed algorithm provides us a new method to improve the shape control of hot strip rolling.