A Performance and Model Complexity Study of a Phase Change Material Thermal Energy Storage Heat Exchanger

Abstract

Phase change material (PCM) thermal energy storage (TES) heat exchangers (HXs) have the potential to shave and shift loads behind the meter if integrated into heat pumps as a substitute for the outdoor coil during discharge. Traditional design methods for PCM TES HX’s in this application require complex and computationally expensive models. This thesis compares a simple analytical model and a detailed finite difference model to determine if the analytical model can accurately capture the physics of these devices and be used for design. The analytical model predicts the time of full discharge given a phase change composite thickness for simulation problems. For design problems the model predicts the thickness of a phase change composite slab given a target discharge time. Therefore, the comparison is conducted by evaluating each model’s discharge time against experimental results of the PCM TES HX prototype. Good agreement would indicate an accurate thickness prediction by the analytical model. Experiments were run to capture on-design and off-design conditions for constant pressure inlet conditions. Off-design conditions include tests that simulate a requirement for greater load or less compressor power in a heat pump. However, the results are not directly extendable to constant power testing, which is what would typically occur in the field in a heat pump. The finite difference model can extend to constant power tests if properly validated, and the analytical model could be upgraded. The experimental results reveal that the analytical model predicts full discharge time for on-design cases with satisfactory accuracy (13.1%) for constant inlet conditions. However large errors in predicted heat transfer rate (148.66 W compared to a maximum heat transfer rate of 500 W) show the analytical model does not work well for simulation problems. In simulation problems instantaneous heat transfer rate is important because it reflects the models ability to predict the load the heat exchanger can handle. The finite difference model suffers from similar error but is able to predict temperature distribution, which could be helpful for device design. Therefore, the

finite difference model will still be an important part of product development. Additionally, the analytical model assumes the phase front moves vertically and does not include sensible heat transfer, which leads to errors in predictions in certain cases. This work recommends the following work flow for product development: Use the analytical model to select a prototype thickness, then build and simulate the model with a more complex continuum approach. Testing and simulating with a finite difference model will help solve problems in design commonly observed like PCM supercooling and hysteresis. Hysteresis was observed in the experimental results for this work. The capacity for melting tests was near the theoretical target, 0.846 kW-hr, but the capacity for evaporator tests varied from 0.77 kW-hr to 0.614 kWhr for final temperatures of 15.1°C and 18.97°C respectively. The results indicate higher discharged energy at lower temperatures, which shows that energy is still stored in the chemical bonds of the PCM and that differences in nucleation of crystals is leading to different behavior in freezing than melting (hysteresis). The temperature distribution prediction in the finite difference model were helpful for diagnosing this issue.