Improving the accuracy of leak localization through the use of an equivalent pipeline model

This study focuses on the localization of single leaks in liquid transmission pipelines. Among the analytical methods used for this purpose, the standard gradient-based leak localization procedure is noticeable. This method relies on calculating pressure gradients using pressure measurements captured along the pipeline. Specifically, it focuses on the period following a leak occurrence, corresponding to a stable flow process. Usually, the procedure applies a basic model that considers the transmission pipeline in terms of global friction resistance effects (i.e., linear pressure losses) while often disregarding local resistance effects (i.e., local pressure losses). In this work, we evaluate this leak localization procedure with regard to the applying of two different models of the examined pipeline. The former simplified model corresponds to the above-mentioned practice. The latter enhanced model accounts for both global and local pressure losses, which are recalculated into a new model consisting of equivalent straight lengths of all parts of a pipe duct. Our investigations employ pressure data obtained from a computational flow model that is based on the geometry and flow conditions of an existing laboratory water pipeline. Notably, the used pipeline flow model is characterized by local pressure losses that constitute only approximately 0.8% of the overall losses caused by friction resistances. During our research, we consider leakages with intensities equivalent to 1%, 2%, 5%, and 10% of the nominal flow rate, situated at different points along the pipeline.