基于CSKPCA-SLSSVM的艾萨炉故障监测研究

The research of ISASMELT furnace fault monitoring based on CSKPCA-SLSSVM

针对艾萨炉熔炼过程中炉子容易出现故障,但故障判断困难的问题,提出了一种融合模糊C均值聚类特征样本KPCA和稀疏LSSVM的故障检测模型。首先基于模糊C均值聚类算法获得样本的簇中心,在此基础上基于特征样本核主元分析法对数据进行处理,并结合T^2和SPE统计量对艾萨炉故障进行初步识别,然后基于稀疏最小二乘支持向量机对初步识别结果进行进一步划分。实验结果表明,该方法建立的艾萨炉监测模型,提高了故障识别的准确率,准确的反映整个生产过程的变化,适合在类似的工业过程中推广。

ISASMELT furnace melting process in the furnace is prone to failure, but failure to judge the difficulty of the problem, presents a fusion of the fuzzy C-means clustering characteristics of the sample KPCA and sparse least squares support vector machine fault detection mode/. First, based on fuzzy C-means clustering algorithm to obtain the cluster center of the sample on the basis of data processing based on the characteristics of the sample kernel principal component analysis, combined with the T2 and SPE statistics are preliminary identification of the ISASMELT furnace failure, and then based on sparse The least squares support vector machine further divided on the results of the preliminary identification. The experimental results show that the ISA furnace monitoring model, the method established to improve the accuracy of identification of failure to accurately reflect the changes of the entire production process, suitable for promotion in similar industrial processes.