Modeling and analysis of flow and heat transfer maldistribution in a supercritical CO₂ hybrid mini-channel heat exchanger based on spatial thermal resistance networks

Flow maldistribution in mini-channel heat exchangers (MCHEs) may severely degrade the thermal performance. Conventional analyses often neglect the coupling between flow and heat transfer, which leads to an overestimation of the maldistribution. To address this issue, a new model is developed to simultaneously predict the flow and heat transfer maldistribution. A back-propagation neural network (BPNN) is first constructed to predict the local hydraulic loss coefficient, and then it is coupled with a spatial thermal resistance network to jointly resolve the flow and temperature distributions in the MCHE. Two major findings are presented: (1) flow and heat transfer maldistribution is negatively correlated with the fraction of frictional pressure drop; and (2) neglecting heat transfer effects results in a substantial overestimation of flow maldistribution, with about 20% at a fixed Reynolds number and more than 50% under turbulent conditions. Furthermore, the proposed predictive correlations reveal that the maldistribution decays exponentially with increasing dimensionless length and varies non-monotonically with the Reynolds number. The maximum errors for these correlations are less than 5% and 3%, respectively. These results demonstrate that accounting for flow and heat transfer coupling is essential for accurate prediction, reliable design, and performance optimization of MCHEs.