摘要
污水生化处理过程的严重非线性给过程监视带来困难。核主元分析(KPCA)可以通过集成算子与非线性核函数有效计算高维特征空间的主元成分,从而有效捕捉过程中的非线性关系。基于KPCA方法构造污水生化处理过程监视策略,可以有效监测污水处理过程中出现的异常状态,与线性PCA监视方法相比,显示出更好的监视性能。
The inherent high nonlinearity of wastewater treatment processes brings exceptional difficulties in process monitoring. Kernel principal component analysis (KPCA) can efficiently compute principal components in high-dimensional feature space by mean of integral operators and nonlinear kernel functions and seize the nonlinear relations of processes. Monitoring strategy developed using KPCA approach can detect the abnormal behavior presented in processes. Compared to PCA, KPCA shows better performance and has very good application prospect.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2005年第z1期157-158,共2页
Chinese Journal of Scientific Instrument
关键词
KPCA
主元分析
非线性过程
故障检测
Kernel PCA Principal component analysis Nonlinear process Fault detection
