Multiphysics simulation of static water freezing process in pumped storage power stations using MATLAB

Pumped Hydro Energy Storage (PHES) facilities in high-altitude regions face severe operational challenges during winter due to ice formation. This study develops a two-dimensional computational model integrating four-way heat transfer interactions among concrete, water, ice, and air domains to analyze coupled thermal processes in PHES upper reservoirs. The MATLAB-based model employs finite volume method, central difference scheme, and implicit time stepping to maintain computational stability. Experimental validation at Dunhua Pumped Storage Power Station (Jilin Province, China) demonstrates the MATLAB model’s superior accuracy with 8.33% prediction errors versus 41.03% for Stefan’s degree-day model. Results showed decreasing air temperature from −10 °C to −25 °C increases ice thickness from 4 to 24 mm after 48 h, while increasing wind speed from 0 to 12 m/s raises the thickness from 4 to 44 mm. Solar radiation reduces ice thickness by 23.5%, and higher initial water temperatures (8 °C vs 0 °C) delay ice formation by 6.86 h. Initial ice formation occurs 3.17–10.03 h after cooling initiation, depending on thermal conditions. Therefore, the model enables reliable ice formation prediction and supports effective ice management strategies, with operational recommendations focusing on maintaining elevated water temperatures and targeted heating of critical infrastructure.