Models for Better Environmental Intelligent Management within Water Supply Systems

Abstract

The paper presents models for better environmental intelligent management within water supply systems. The following computer models were developed: supervising parameters (pressure and flow) of water supply network (classification models in the form of neural networks, hybrid neural networks, decision trees and multiple decision trees), forecasting of water supply network load in different intervals of time (prediction models in the form of neural networks and hybrid neural networks), preferences for network operator and consumer in the form of decision rules and decision trees, classification of exceptions, typical examples and preferences for controlling water flow, controlling of pumps in the water supply network in the form of decision and activity rules and controlling of pumps for filling up retention tanks in the form of decision and action rules. These models were compared with a view to obtaining optimal models to control the parameters of water supply networks. The models are embedded in intelligent decision support system with a knowledge acquisition module. The research was done for Municipal Water Supply and Sewage Company in Rzeszow, Poland.