Detection of Leakages in a Pipeline Network based on Hydraulic Laboratory Modelling with Artificial Intelligence

Abstract

Pipeline transportation of resources is considered a vital method due to low operational cost, and simple design and implementation. However, the presence of leakages within pipeline networks gives rise to noteworthy jurisdiction regarding environmental impact, economic implications, and safety considerations. The prompt identification and precise localization of such leakages are of utmost importance in order to get rid of their potential consequences on human existence. This project aims to detect leakages in a pipeline network based on hydraulic laboratory modelling with artificial intelligence systems. The dataset from both the hydraulic laboratory network and EPANET simulation respectively were used to train and test a model, then validate using for leakage prediction and localization using artificial neural network. The results shows that pressure is a more valid parameter to detect leakages to flowrate in a pipeline network. Also, artificial neural network developed model performed very well in predicting leak sizes with an accuracy of 96.89% respectively. The model developed based achieved validation accuracies which vary broadly between about 85% and 90%. Also, the F-score ranged between 80% and 91% which makes the model is valid to be used to predict and localize the leaks in real time.