This paper aimed to investigate the high temperature resistance filtration material filter efficiency forecast model based on Least Square-Support Vector Machine(LS-SVM).According to the LS-SVM principle,it used Matlab as software platform and lssvmlab toolbox to learn experimental data.With air temperature,air velocity,generating dust concentration as input,and filter efficiency as output,it trained with the LS-SVM,predicting the filter efficiency under various conditions.The forecast results indicate that the method has achieved the satisfactory precision.Case study shows that LS-SVM based high temperature resistance filtration material filter efficiency forecast model has better robustness and forecasting accuracy and faster than RBF neural network based forecast model.%本文主要研究利用最小二乘支持向量机的方法预测耐高温高效滤料在不同测试条件下过滤效率的变化。笔者采用最小二乘支持向量机的方法,以Matlab作为软件平台,利用lssvmlab工具箱对于滤料测试的实验数据进行学习,以气体温度、气体流速,上游发尘浓度为输入,滤料过滤效率为输出训练算法,预测不同测试条件下滤料的过滤效率,并且与RBF神经网络的预测结果进行比较。实例分析表明,与基于RBF神经网络的耐高温滤料过滤效率模型相比,基于最小二乘支持向量机的过滤效率模型具有更高的精度,更强的鲁棒性,并且速度更快。
展开▼