Leak Detection in a Branched System by Inverse Transient Analysis with the Admittance Matrix Method

The diagnosis of water distribution systems by means of the inverse transient analysis requires efficient and reliable numerical models. In the network admittance matrix method (NAMM) the 1-D waterhammer governing equations are integrated in the frequency domain and organized in a laplacian matrix form. The NAMM is particularly suitable for complex systems because of this structure and can be used for the system diagnosis, including leak sizing and location. In this paper a damaged branched system is considered and the diagnosis is performed by means of the NAMM using experimental data from laboratory transient tests. Two different boundary conditions are used in the implementation of the NAMM and the leak is located and sized with a reasonable approximation. An extended numerical investigation is also presented and allows confirmation of the results for different leak locations. The use of the NAMM for the leak detection and the validation using experimental data on a branched system are the main original contributions of this work. The successful diagnosis indicates promising results for applications in more complex systems.