Advances in Neural Networks for Optimizing Drinking Water Pipeline Management - A Comprehensive Literature Review and Practical Application in Network Calibration with Roughness Analysis

Abstract

automation of a water distribution network calibration process in an aim to attain efficiency while eliminating sources of error. In this respect, a literature review forms part of modern applications of AI to water management and covers a practical case study about network calibration. The theoretical part describes the research status quo regarding AI and Neural Networks for water management, in general, and a bit more concretely towards network calibration. In that respect, the practical section covers the implementation of an artificial neural network to proceed with the automatic calibration for a real water supply network. Methods based purely on AI do not hold great hopes for network calibration. Therefore, further research is needed to test approaches such as Physically Informed Neural Networks or hybrid methods.