A data-driven model for heat pump water heater with wrap-around condenser coil in heating-up transients

To simulate the heating-up transients in heat pump water heaters, the traditional approach used CFD software to simulate the transient flow and temperature fields inside the water tank, which usually took more than ten hours to complete one simulation. To balance the computational speed and simulation accuracy, a data-driven model for heat pump water heaters with wrap-around condenser coils was proposed in this study that can accommodate water tanks and condensers with different geometries and operating parameters. The natural convection heat transfer coefficient was fitted as an explicit function, and the water temperature evolution and distribution can be solved using the normalized temperature distribution database summarized in this work. Compared to the traditional CFD-based model, the proposed data-driven model reduced simulation time by four magnitudes while achieving an energy deviation of less than 1 %. The water temperature predicted by the data-driven model deviated from the experimental data by less than 2 °C, while the discrepancy in heating energy was under 3 %. These results indicated that the proposed data-driven model can accurately simulate the dynamic temperature evolution within the tank and the performance of the refrigeration system with trivial computational resources.