Uncertainty analysis of transient flow modeling and transient-based leak detection in elastic water pipeline systems

Abstract

In the transient flow modeling and analysis of practical water pipeline systems, discrepancies commonly exist between numerical simulation results and experimental measurement data. Such discrepancies are usually accounted for the inaccuracy and inadequacy of the used mathematical models and/or the lack of understanding of the hydrodynamic physics for transient pipe flows in the literature. However, the variability or uncertainty of the parameters for such numerical model inputs may also attribute to these discrepancies, especially in complex pipeline systems. This paper investigates the effects of different system uncertainties on the transient flow modeling and analysis such as pipe system design and leak detection. Different factors of pipe and fluid properties as well as system operations and complexities are considered for the uncertainty analysis. The one-dimensional (1D) water hammer model and the Monte-Carlo simulation (MCS) method are used in this investigation. The analysis results in this study demonstrate that the variability of the transient flow modeling can be easily affected by the uncertainty factors of wave speed and system complexities, while the transient-based leak detection is more sensitive to the uncertainty factors of wave speed and data measurement than the factors of valve operation and initial hydraulic condition.