核主成分分析的高炉故障检测研究

Blast furnace fault detection based on KPCA

由于高炉冶炼系统的复杂性,传统的故障检测方法在高炉故障检测中的应用效果不佳.同时,高炉冶炼过程中的数据具有明显的非线性特征,利用主成分分析(PCA)等线性多元统计方法也难以取得良好的故障检测效果.针对这种情况,提出了利用核主成分分析(KPCA)方法对高炉冶炼过程中的故障进行检测,以适应高炉的非线性特征,实现对高炉故障的快速检测.

Traditional fault detection methods are ineffective in complex blast furnace processes.On the other hand,the data in the blast furnace process had significant nonlinear characteristics;and in PCA and other multiple linear statistical methods it is also difficult to obtain good fault detection effect.We proposed a method to detect fault in the blast furnace process based on the KPCA method.It fully adapts to blast furnaces' nonlinear characteristics to achieve fast and accurate detection of blast furnace failures.