Lee J M, Yoo C K, Lee I B
Department of Chemical Engineering, Pohang University of Science and Technology, San 31 Hyoja Dong, Pohang, P.O. Box 790-784, Korea.
Water Sci Technol. 2003;47(12):49-56.
A new monitoring method using independent component analysis (ICA) is suggested for the wastewater treatment process (WWTP). ICA is an extension of PCA (Principal Component Analysis). While PCA can only impose independence up to the second order (mean and variance) with constraint on the direction vectors to be orthogonal, ICA imposes statistical independence up to more than second order on the individual component and has no orthogonal condition. When the variables have the Gaussian distribution, PCA itself provides a satisfactory result in monitoring performance. However, the measured variables are not often normally distributed. In this case, ICA can provide better monitoring results than PCA since ICA is based on the assumption that the latent variables are not normally distributed. In this paper, the ICA monitoring algorithm with kernel density estimation was applied to fault detection and diagnosis of the wastewater simulation benchmark. ICA with kernel density estimation gives better results than PCA in disturbance detection in spite of severe periodic features of the wastewater plant.
本文提出了一种用于污水处理过程（WWTP）的基于独立成分分析（ICA）的新型监测方法。ICA是主成分分析（PCA）的扩展。PCA只能在二阶（均值和方差）上施加独立性，并对方向向量施加正交约束，而ICA则对各个成分施加超过二阶的统计独立性，且没有正交条件。当变量具有高斯分布时，PCA本身在监测性能方面能提供令人满意的结果。然而，测量变量通常并非正态分布。在这种情况下，ICA能提供比PCA更好的监测结果，因为ICA基于潜在变量非正态分布的假设。本文将带核密度估计的ICA监测算法应用于废水模拟基准的故障检测与诊断。尽管污水处理厂具有严重的周期性特征，但带核密度估计的ICA在干扰检测方面比PCA能给出更好的结果。
