Comparative analysis of modelling approaches for pit thermal energy storage: Towards reliable long-term predictions

Pit thermal energy storage (PTES) is important to address the intermittency of renewable heat supply and achieve cross-seasonal utilization. This study examines several simplified modelling approaches that are suitable for long-term simulations of PTES. The models are analyzed in terms of governing equations, geometry, grid generation, temporal discretization, mixing treatment, and fluid-solid coupling. A controlled benchmarking strategy is adopted to test the applicability and failure boundaries of key modelling choices under unified conditions, using analytical solutions, CFD results, and field measurements as case-dependent comparison baselines. Results show that in the one-dimensional conductive case, all simplified models closely match the analytical solution, with the semi-analytical approach offering better stability than the explicit Euler. The adiabatic mixing treatment method also performs well in two-dimensional natural convection problems, with temperature discrepancies accumulating at about 0.35 °C/day. In three-dimensional forced convection scenarios, however, the absence of a forced-convection transport correction under high inflow jets can lead to a maximum temperature deviation up to 17.6 °C. This case defines a clear failure boundary for the simplified governing assumptions under high-flow conditions. Comparisons with one year of measurements show that all simplified models predict charging and discharging energies with errors below 3%. Nonetheless, some geometric simplifications (such as cuboid or cylindrical shapes) can cause thermal loss deviations exceeding 30%. For the models with fluid-solid coupling, flow-control deviations can reach 10%, and heat-pump deviations can be amplified by a factor of 3.6. Overall, a 2D semi-analytical model with adiabatic mixing, a truncated-pyramid geometry, internal coupling, and adaptive/refined grids is recommended for annual PTES simulations under large-volume, low-flow operation where stratification is maintained. In contrast, high-flow jet conditions can trigger strong forced convection and rapid disturbance propagation, under which the current simplified governing assumptions are not applicable and forced-convection corrections or CFD-based modelling are required. These findings provide a solid foundation for designing large seasonal heat storage systems and integrating multiple renewable energy sources.