【摘要】 针对传统轧制力模型的固有缺陷,提出了一种基于灰色关联分析的ELM(极限学习机)轧制力预报模型.首先通过灰色关联分析对输入变量进行相关性分析,用于提高模型的性能;然后结合10次10折交叉验证确定ELM模型的隐含层节点数,建立热轧薄板的轧制力预测模型.运用现场数据对该网络进行训练和测试,并与传统的模型相比较.实验结果表明,该模型能快速、准确地预报轧制力,能满足在线预测的要求.更多还原

【Abstract】 According to the inherent shortcomings of traditional rolling force models,an ELM( extreme learning machine) model based on grey correlation analysis is proposed. First,the correlation analysis of input factors is carried out by grey correlation analysis in order to improve model performance. Second,the 10-fold cross validation method is applied to determining the node number in the hidden layer of the ELM model,and the rolling force prediction model is then established. The network is trained and tested with field data and compared with traditional models. Simulation results show that the model proposed can predict rolling force quickly and accurately,and is able to meet the requirements for online prediction in actual rolling processes.更多还原