Transient electrothermal coupling optimization method of through-silicon via based on neural network

Based on neural network-assisted artificial bee colony, a silicon through-hole electrothermal transient optimization method is introduced. It efficiently and accurately analyzes transient electrothermal issues in three-dimensional microsystems. Electrothermal coupling simulations are conducted using finite element analysis software to examine the impact of design parameters (silicon through-hole radius, oxide thickness, silicon through-hole spacing) on silicon through-hole array performance, including copper pillar and micro-protrusion temperatures. A neural network is used to establish the mapping relationship between design and performance parameters. A collaborative optimization strategy with performance constraints is proposed to optimize design parameters using a bee colony optimization algorithm. The predicted performance closely matches finite element simulation results, with a maximum temperature deviation of 2.6%. This validates the feasibility of the optimization strategy. This method significantly reduces simulation time and simplifies mathematical analysis in multi-physics coupling compared to traditional finite element methods.