Fault detection and isolation for wastewater treatment plants using interval prediction and Bayesian reasoning

The wastewater treatment plant that operates abnormally may lead to poor effluent quality, resulting in the destruction of the environment, and even more serious situation. Therefore, it is necessary to detect and isolate faults in the wastewater treatment plants. This paper proposed a method of fault detection and isolation using interval model. Radial basis function (RBF) neural network is utilized to model the wastewater treatment plant, and the linear output weights of the neural network are estimated by the set membership identification algorithm. After that, the confidence interval of the predicted effluent variables can be obtained, and then the interval boundary is used as the threshold for fault detection. After detecting the fault, based on this interval model and the Bayesian reasoning, the posterior probability of the considered faults can be calculated. When the probability in exceed of a certain threshold, the fault can be successfully isolated. The final experimental results verified the method.