Performance of manifold microchannel liquid cooling
As advancements in information technology progress,chip development is moving towards large-area,high-power configurations,posing significant challenges in heat management.Microchannel liquid cooling has emerged as a viable solution to address the thermal management difficulties associated with high-power chips.Microchannel liquid cooling has emerged as a solution to address the thermal management difficulties associated with high-power chips.However,traditional straight microchannels suffer from high flow resistance and poor temperature uniformity.This paper introduces a novel manifold microchannel heat sink by incorporating a manifold inlet/outlet liquid structure,distributed jet impingement,and micro-pin fins.Under conditions where the average heat flux density exceeds 330 W/cm² and the total power reaches 2500 W,the chip's average temperature remains below 70℃,indicating that high efficiency cooling is achieved.Through numerical simulation,it is found that increasing the height of the jet impingement chamber leads to a significantly increase in heat transfer coefficient,albeit at the expense of an obvious increase in overall pressure drop.An optimum height of the jet impingement chamber is thereby identified.The dimensions of the micro-pin fins on the base plate,along with their relative proportions to the size of jet impingement chamber,emerge as critical structural parameters for the proposed microchannel heat sink.The presence of micro-pin fins does not uniformly contribute to heat transfer enhancement.A dimensionless parameter,micro-pin fin ratio,is introduced to signify the relative height between the micro-pin fins and the jet impingement chamber.A critical micro-pin fin ratio is realized by balancing the obstructing and enhancing effects.When the proportion of fins is higher than this critical value,the effect of enhanced heat transfer can be achieved.This study provides valuable guidance for the systematic design of the novel microchannel heat sink.
国家科技期刊平台
NSTL
万方数据
