Detection of Drinking Water Quality Using CMAC Based Artificial Neural Networks

This paper describes the design, implementation and performance evaluations of the application developed for real-time drinking water quality detection using the Cerebellar Model Articulation Controller (CMAC) Artificial Neural Networks (ANNs). Most drinking water systems currently monitor a significant number of water quality parameters at the plant. These are required for compliance and maintenance of water quality as the water enters the distribution system. The threat of chemical or microbiological contamination of drinking water is obvious, and would be an effective way of causing devastating public health consequences. In this study we propose a practical approach for the detection and classification of chemical and microbiological contaminants in drinking water. CMAC ANN is intended for the use in the detection process of drinking water quality where real-time capabilities of the network are of prime importance. CMAC has been chosen as an alternative to Multilayer Perceptron (MLP) because of its exceptional learning capability. The learning speed of the CMAC ANN algorithm is much higher than that of MLP using a backpropagation (BP) algorithm. The detection (rate) of the new water samples has also been so fast (high) and accurate that the rate of success almost turned out to be 100% as compared to an ordinary MLP using the BP algorithm of which the success rate was found to be 98%.