基于极限学习机(ELM)的连铸坯质量预测

针对传统基于BP神经网络建立的连铸坯质量预测模型训练速度慢、适应能力弱、预测精度低等问题,本文提出一种基于极限学习机的连铸坯质量预测方法,对方大特钢60Si2Mn连铸坯中心疏松和中心偏析缺陷进行预测,并与BP和遗传算法优化BP神经网络预测模型的预测结果进行分析对比.结果表明:BP及GA-BP神经网络预测模型对连铸坯中心疏松和中心偏析缺陷的预测准确率分别为50%、57.5%、70%和72.5%;而基于极限学习机的连铸坯预测模型预测准确率更高,对连铸坯中心疏松和中心偏析缺陷的预测准确率分别为85%和82.5%,且该模型具有极快的运算时间,仅需0.1 s.该模型可对连铸坯质量进行迅速准确地分析,为连铸坯质量预测的在线应用提供了一种新的方法.

Quality prediction of the continuous casting bloom based on the extreme learning machine

To solve the problems of slow training, weak generalization ability, and low prediction accuracy in the traditional prediction model established in terms of the BP neural network, a method of the quality prediction of the continuous casting bloom based on the extreme learning machine (ELM) was proposed to predict the degree of the center porosity and the central segregation of 60Si2Mn continuous casting bloom produced by Fangda Special Steel. Comparing the prediction models of the BP neural network and the GA-BP neural network, the results show that the prediction accuracy of the model based on ELM is improved to 85% and 82.5% in the center loose and central segregation, respectively, and the operation time is reduced to 0.1 s. The model can rapidly and accurately analyze the quality of a continuous casting billet, thus providing a new method for the online application of continuous casting billet quality prediction.