Fault Diagnosis of WWTP Based on Improved Support Vector Machine

Abstract

Wastewater treatment is a complicated process where sensors and equipment are operated at harsh conditions, and there are often long time delays in variables' response to disturbances. In this work, support vector machine is applied to diagnose fault. Because of the unbalanced distribution of the fault classes data quantity or importance, the risk functional R _WLOO(α) with weight coefficient based on leave-one-out errors is presented; then Genetic Algorithms (GA) is used to globally optimize the risk functional R _WLOO(α). Because of the size of the data is large, we present a simple algorithm of R _WLOO(α) to reduce the amount of calculation. The improved SVM is used to classify dataset of WWTP, and the results indicate that compared with the standard SVM and BP neural network (NN), the improved one can gain higher classification accuracy.