Optimization of delta ribs in a microchannel heat sink using numerical analysis and Artificial Neural Networks

The miniaturization of electronic devices necessitates effective thermal management due to high power density. This study investigates the effects of different delta rib arrays on the flow and thermal performance of microchannel heat sinks under varying Reynolds numbers ( Re = 100, 250, 400, 550, and 700) and rib attack angles ( θ = 22.5°, 30°, 37.5°, 45°, 52.5°, and 60°), and an artificial neural network (ANN) model was used for optimization. Among the three arrays, the alternating array exhibited the highest performance due to intensified interactions between odd and even rows. At Re = 700, the Nusselt number increased with θ and slightly decreased at higher θ , while the friction factor rose over the same range. The maximum performance evaluation criterion ( PEC ) of the alternating array was 1.3600 at θ = 37.5°, showing a 6.12 % improvement over the optimized rectangular rib. ANN-based optimization further yielded a PEC of 1.3609 at θ = 40.5° and Re = 700, with a relative error of 0.03 % compared to CFD results, confirming high accuracy. These findings demonstrate that the delta rib can significantly enhance thermal performance, and ANN-based optimization provides an effective tool for designing high-performance in advanced thermal management systems.