In this research, a CFD model is used to examine the effects of microchannel flexible elements with a circular vortex generator on flow dynamics and heat transfer. A series of CFD simulations, guided by a design of experiments approach and integrated with a radial basis function neural network, is employed to investigate the effects of flexible wall amplitude, oscillation frequency, and the Reynolds number (Re). The main focus is on three key objectives: the time and space averaged Nusselt number, 𝑁𝑢 ̿̿̿̿, pressure drop, ΔP, and the thermal performance ratio (TPR). Results reveal that the ΔP is most sensitive to the Re, while the amplitude strongly governs the TPR, and both amplitude and Reynolds number have a greater impact on the 𝑁𝑢 ̿̿̿̿ rather than frequency. Optimization results indicate that the maximum Nusselt number (𝑁𝑢 ̿̿̿̿ ≈ 51.3) occurs at Re = 1500, Ae ≈ 0.112 mm, and fe ≈ 5.45 Hz, the minimum pressure drop (ΔP ≈ 1272 Pa) at Re = 750, Ae = 0.125 mm, and fe = 2 Hz, and the best TPR (≈ 1.24) at Re ≈ 1384, Ae ≈ 0.113 mm, and fe ≈ 5.42 Hz. The surrogate-based optimization framework shows an efficient and accurate strategy for balancing heat transfer enhancement and hydrodynamics efficiency. This highlights the potential of elastic-wall microchannels as high-performance thermal management devices.
