Sensitivity analysis of one-dimensional water hammer model in pump-pipeline systems

Water hammer models are crucial for simulating transient flows in pump-pipeline systems. However, uncertainties in input variable significantly affect model accuracy, potentially leading to equipment damage and water leakage. To improve model reliability, sensitivity analysis is an essential tool for identifying critical parameters affecting model outputs. This study introduces a one-dimensional water hammer model for analyzing pressure fluctuations in pump-pipeline systems, emphasizing the identification of critical parameters that influence maximum and minimum pressures. Two sensitivity analysis methods: Partial Rank Correlation Coefficient (PRCC) and Mutual Information (MI), are utilized. The results show that the rotational inertia, flow velocity, and valve closing time significantly influence pressure. Both PRCC and MI demonstrate similar ranking of parameter sensitivity. These findings contribute to developing more precise and reliable models for practical engineering applications and aid in designing strategies to mitigate pressure surges during power failures.