基于模糊聚类的粒子群神经网络预测卷取温度

Coiling temperature prediction with PSO neural network based on fuzzy clustering

影响热轧带钢卷取温度的因素多而且复杂,采用传统的温度预测模型难以达到较高的精度。为了满足卷取温度高精度预测的要求,将模糊聚类分析、神经网络、粒子群算法结合起来,提出了一种基于模糊聚类的粒子群神经网络用于预测卷取温度。运用现场实际数据测试表明,该方法预测卷取温度效果良好。

Factors affecting coiling temperature are very complex.Traditional coiling temperature prediction model is difficult to get the high precision of prediction.In order to satisfy the demand we propose a PSO neural network based on fuzzy clustering to predict coiling temperature.Fuzzy clustering,PSO and neural network are combined in this method.The experiment and analysis with actual production data indicates that the method could predict coiling temperature well.