Water Supply System operation regarding consumer safety using Kohonen neural network

A Water Supply System (WSS) ought to be high reliable continuous operating system. Drinking water supply utilities are responsible for providing a safe and reliable supply of potable water to their customers. Assessment procedures were implemented using the Kohonen neural network. In studies a set of attributes objects was used, including input variables such as, e. g.: the total time of lack of water supply in one year, the failure rate, the mean repair time of water pipes, economic efficiency. On the basis of maps and classification data a detailed synthesis analysis was performed. Trained network is able to assess the new systems which are not presented during the network learning. This paper presents a framework for the analysis of WSS operation that can be applied to the other systems. It is expected that the methodology using Kohonen neural network would provide the city authorities with support for decision making.