Study of frost growth on heat exchangers used as outdoor coils in air source heat pump systems
Abstract
Scope and Method of Study: During winter heating operation, the outdoor coil of a heat pump acts as an evaporator and when the ambient temperature is near freezing, the moisture in the atmospheric air freezes on the coil surface. The frost growth affects the air flow and also adds resistance to heat transfer thus reducing the capacity and the efficiency of the heat pump. An experimental facility is designed and built to test a small scale heat exchanger working in frosting conditions. Tests are carried out using small scale fin-tube and microchannel heat exchanger over a range of glycol inlet temperatures, air velocities and ambient air RH. A semi-empirical frost model based on a scaling approach is developed and applied to both fin-tube heat and microchannel heat exchangers. The model is developed to handle non-uniformities in both refrigerant and air side. The frost model is integrated with a segment-by-segment heat exchanger calculation algorithm and is validated against experimental data. Frost growth model is also incorporated in a quasi-steady state system simulation algorithm. Findings and Conclusions: From the experimental study it was observed that the temperature of the surface and air RH affected the rate of frost growth significantly while the air velocity did not have a great influence. It is demonstrated that a dry heat transfer correlation can be used during frost growth period by correctly accounting for frost thickness in the hydraulic diameter calculations. Ignoring the phenomenon of air redistribution was found to result in errors in the range of 20% to 50% in predicted frost thickness. Frost thickness predicted by the frost model is within 5% of the measured values for most of the cases. Frost mass accumulation predicted by the simple model is higher than the measured values due to a uniform thickness assumption in the model. An improved frost model based on 1-D finite volume discretization is also presented which improved the frost mass prediction to within 13% of measured values. The system simulation is validated against experimental results and found to match reasonable well. The discrepancy between simulation and experiments were due to the effects of system transience in the initial stages.
Collections
- OSU Dissertations [11222]