Multi-objective optimization of a cylindrical heat sink with straight and forked fins using artificial neural network (ANN)

This study investigated the optimized shape of a heat sink for circular chip-on-board (COB) type light-emitting diode (LED) equipment. The numerical model was validated through experiments, and an artificial neural network (ANN) model was constructed to predict thermal performance based on the data from numerical analysis. The distribution of the chimney-shaped airflow and the change in airflow based on the forked point were analyzed. The thermal performance trend was demonstrated using predictions from the neural network model. The finning and porosity factors were introduced to establish criteria for changes in thermal performance trends. After that, multi-objective optimization was performed, and several heat sink designs for a wide range of total fin mass and thermal resistance were proposed in the form of a Pareto Front. This study is expected to contribute to efficient and accurate thermal management of LED equipment by proposing a heat sink design that has not been extensively explored using machine learning techniques.