Numerical investigation of water hammer effects in pressurised pipeli incorporating dynamic cavity evolution

Cavity formation, growth, and collapse during water hammer represent a highly nonlinear and potentially hazardous transient process. A major modelling challenge is to simultaneously retain the efficiency of one-dimensional characteristic-based water hammer computation and describe the distributed evolution of cavity boundaries, cavity shape, and cavity volume. Traditional discrete vapour cavity models (DVCM) exhibit significant limitations in this regard, often causing pressure prediction errors and non-physical phenomena. To address these challenges, this study develops a seven-classification cavity model based on refined flow-state subdivision. The proposed framework enables dynamic tracking of cavity boundaries, cavity shape, and cavity-volume distribution during transient cavity growth, collapse, and boundary movement. Experimental validation in simple and complex pipeline systems shows that the seven-classification cavity model captures pressure peaks, oscillation trends, and their occurrence timing more accurately than the classical DVCM. In the simple pipeline system, the mean absolute relative error of the peak values is reduced by about 90.0%, and the mean absolute time deviation of the main pressure rises is reduced by about 54.1%. In the complex pipeline system, the absolute relative error of the first major pressure peak is reduced by about 94.2%, while the timing error of the first major pressure rise is reduced by about 25.0%. The predicted cavity shapes, boundary evolution, and cavity-volume distributions also agree well with experimental observations and avoid the non-physical concentration of cavity volume produced by the classical DVCM. This study provides a more reliable framework for simulating vapour-cavity-induced water hammer in pressurised pipelines.