论文部分内容阅读
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。
Aiming at the problem that the furnace is prone to failure in the process of Isaac furnace smelting, but the fault is difficult to judge, a fault detection model based on fuzzy C-means clustering feature KPCA and sparse LSSVM is proposed. Firstly, the cluster centers of the samples were obtained based on the fuzzy C-means clustering algorithm. Based on this, the data were processed based on the principal component analysis (PCA) of characteristic samples, and the fault of Isaac furnace was preliminarily identified by T ~ 2 and SPE statistics. Based on the sparse least squares support vector machine, the initial recognition results are further divided. The experimental results show that the ISAS furnace monitoring model established by this method improves the accuracy of fault identification and accurately reflects the changes of the entire production process and is suitable for promotion in similar industrial processes.