Leak detection in branched pipe systems coupling wavelet analysis and a Lagrangian model

Abstract

Transient analysis of pressurized pipe systems yields useful information regarding the actual state in an effective and rapid manner and then it can be used for the diagnosis of the system. Besides the framework of the inverse problem solution, transient based leak detection can also focus on the modalities with which pressure waves generated by a known manoeuvre travel through a system. This latter approach has been successfully applied by several authors to simple systems, usually comprising a single pipe with a constant head reservoir and a valve at the end section. The present work extends this approach to more complex systems, that is, those with at least one junction. To accomplish this, the pressure signal is analysed using wavelet analysis, a technique which has produced encouraging results. Moreover, the analysis of the pressure signals is generalized by means of a Lagrangian model, which directly simulates the pressure waves' generations and movements in the system. Coupling wavelet analysis with the Lagrangian model allows us to associate singularities pointed out in pressure signals to singularities in the system. In order to better ascertain the applicability of the method, tests were carried out on both a laboratory installation and a real system. © IWA Publishing 2009.