首页|Research from Chung-Ang University Yields New Findings on Machine Learning (Mult imodal machine learning for predicting heat transfer characteristics in micro-pin finheat sinks)
Research from Chung-Ang University Yields New Findings on Machine Learning (Mult imodal machine learning for predicting heat transfer characteristics in micro-pin finheat sinks)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News - Investigators publish new report on ar tificial intelligence. According to news reportingout of Seoul, South Korea, by NewsRx editors, research stated, “As three-dimensional integrated circuit(3D-I C) chip technology advances, thermal management has become increasingly importan t because ofincreasing heat flux from thermal stacking. Micro-pin fin-embedded cooling has emerged as a promisingsolution for 3D-ICs, offering better thermal and hydraulic performance than conventional microchannelheat sinks.”The news editors obtained a quote from the research from Chung-Ang University: “ It is also easy tointegrate into existing 3D-IC structures, such as through-sil icon vias between stacks. The utilization oftwo-phase flow in micro-pin fins fu rther enhances temperature uniformity and improves the heat transfercoefficient by leveraging latent heat. Nevertheless, predicting thermal performance in micr o-pin fin heatsinks under boiling conditions remains challenging owing to intri cate geometric shapes and diverse operatingconditions. The present lack of corr elation or theoretical models poses a significant obstacle. To addressthis prob lem, our study employed a Multimodal machine-learning (ML) approach, combining i magedata capturing boiling patterns of two-phase flow and information about geo metric shape and operatingconditions, to predict heat transfer characteristics in micro-pin fin heat sinks. We utilized experimentaldata comprising 155 types of boiling heat transfer data with the dielectric fluid FC-72 in two micro-fin shapes directly etched on Si. Four ML algorithms (XGBoost, LightGBM, Multilayer p erceptron (MLP), andMultimodal ML) were employed to predict thermal performance . The correlation coefficient analysis beforelearning revealed the influence of each type of measurement data on the heater surface temperature duringtwo-phas e flow. Prediction accuracy was measured using mean absolute percent error (MAPE ), and theresults were compared in terms of maximum and average temperature dep ending on the characteristicsof each ML algorithm. Overall, the Multimodal appr oach demonstrated superior capability in predictingtemperature distributions wi th spatial details, surpassing conventional decision-tree algorithms and MLPin performance.”
NETL
NSTL
