Two novel leakage identification techniques based on hydraulic transient flow in a pipeline

The inverse transient analysis (ITA) techniques are proposed to identify leakage parameters in a reservoir-pipeline-valve system based on the forward transient analysis method. First, the applicability of transient leakage models based on steady friction and unsteady friction (UF) models is validated by means of forward transient analyses based on the one-dimensional method of characteristics, which show that the leakage model with the UF model provides better performance compared to the experimental data. Based on the UF model, ITA techniques for leak parameter identification further incorporate successively and separately a genetic algorithm (GA) and a fusion algorithm of simulated annealing-genetic algorithm (SA-GA) to predict leakage parameters, respectively. Compared to experimental data, the numerical research on single-hole and multi-hole leak cases indicates that both algorithms can accurately predict the relative leakage location and flow rate, especially the ITA-SA-GA has better performance. In the single-hole leak case, the relative errors in the identified relative leakage location and flow rate by the ITA-SA-GA were 2.80% and 5.85%, respectively. In the multi-hole leak case, the relative errors in the identified relative leakage location and flow rate by the ITA-SA-GA were 10.00% and 26.95% for leak hole no.1, respectively, and the corresponding relative errors of 5.73% and 13.77% for leak hole no.2.