Optimizing symmetric-sinusoidal microchannel heat sinks with butterfly prisms for high temperature superconducting applications: An ANN-based approach

As electrical devices (EDs) advance in power density and miniaturization, microchannel heat sinks (MCHSs) gain extensive attention for efficient heat dissipation. Among these, the symmetric-sinusoidal microchannel heat sink (SMHS), which balances thermal and hydraulic performance, is emerging as one of the most promising cooling solution for EDs. Yet its performance breakthroughs are often constrained by high pressure drop (ΔP) losses and suboptimal heat transfer efficiency. This study proposes a SMHS integrated with butterfly prisms (SMHS-BP). By optimizing the structural and arrangement parameters of the BP (wing distance D_u, tail distance D_v, and flow direction distance D_h), an artificial neural network (ANN) was employed to predict the Nusselt number and ΔP for different designs. The predictions, compared with simulation results, explore the heat transfer mechanisms of SMHS-BP and determine specific parameters for the overall ideal framework (OIF) and thermal ideal framework (TIF). Results demonstrate that the incorporation of BP significantly reduces fluid momentum decay at microchannel wave crests, promoting multidirectional fluid mixing and enhancing heat transfer. Furthermore, numerical simulations of the thermal behavior of the latest-generation high temperature superconducting synchronous condenser (HT-SSC) equipped with SMHS-BP demonstrated the promising application prospects of microchannels in superconducting power systems.