Block Averaging Based Leak Detection for Water Pipes Based On Vibration Sensors

This paper proposes a new time difference estimation method for leak detection based on transform-domain cross spectral densities (CSDs) between two vibration sensor signals. A new pipe channel model is derived from field measurement data, and a block averaging method is employed for robust estimation of CSDs. Parameter optimization is performed to further improve the estimation accuracy. The performance of the proposed method is evaluated through numerical simulations and field measurement data.