Leak diagnosis in pipelines using a combined artificial neural network approach

Leakages in pipelines affect the reliability of fluid transport systems causing environmental damages, economic losses, and pressure reduction at the delivery points. Therefore, this paper presents a methodology to detect and locate water leaks in pipelines by using artificial neural networks (ANN) techniques and online measurements of pressure and flow rate. Contrary to reported works in the literature, the proposed method estimates the friction factor of the pipe and uses this information as an input to compute the leak position. Data generated from a validated numerical simulator was used to enrich the data-training set for the ANN. Various leak scenarios were considered to characterize pressure losses and their differentials in different sections of the pipeline. Finally, the algorithm was tested experimentally in a pilot plant. The results demonstrate good performance and the applicability of the proposed method. [Display omitted] •A combined artificial neural network (ANN) for leak diagnosis in pipes is presented.•The ANN scheme estimates the location and friction factor based on measurement data.•The evaluation of the diagnostic method with experimental datasets are included.•An average error of 0.629% was obtained for leak location in the experiments.