Two phase flow distribution in parallel flow heat exchangers -- experimentally verified model

Maldistribution of refrigerant is a problem in parallel microchannel heat exchangers that lowers their potential effectiveness, especially when controlled by superheat. This paper describes modeling and numerical simulation of two phase mass flow distribution in a microchannel evaporator, on the basis of the pseudo 2-D finite volume method. Emphasis is placed on refrigerant-side heat transfer and pressure drop characteristics. The global flow distribution is based on the mechanistic fact that the pressure drop along each flow path containing an individual microchannel tube must be the same. Besides the primary pressure drop across the microchannel tube, other pressure drops are also considered and modeled. These include the frictional pressure loss along the inlet/outlet headers, as well as contraction and expansion loss associated with fluid entering and leaving the tube. Mass flow rate and quality in each microchannel tube, overall pressure drop and evaporator surface temperatures are calculated and then compared to data taken from the experimental facility. This maldistribution model simulates refrigerant distribution in a microchannel evaporator for two cases. In the first case, the flash gas bypass system, flash gas from the expansion valve is separated and only single phase liquid is supplied into the evaporator. The second case is a conventional direct expansion system where refrigerant enters the evaporator in a two-phase state. For the second case, a quality distribution profile along the inlet header is proposed to simulate the liquid-vapor distribution in the microchannel tubes. The second case is the same as the first case, with a quality distribution profile added. The maldistribution model is compared to a uniform distribution model, a common assumption made in open literature where all microchannel tubes receive identical flow rate. Comparing the maldistribution and uniform distribution models to experimental data, results indicate that the uniform distribution model predicts a higher superheat than the maldistribution model in every case. In order to compare cooling capacity prediction, the mass flow rate in the uniform distribution model was then increased until the exit superheat matched the superheat of the maldistribution model. For every case, the cooling capacity is over predicted by the uniform distribution model by an average of 34%. If an evaporator model is to be used as a design tool, it becomes necessary to use the mass flow distribution model to accurately predict both superheat and cooling capacity accurately