FEM-based Thermal Profile Prediction for Thermal Management of System-on-Chips

Abstract

In this paper, we propose a thermal profile based on the finite element method (FEM). The proposed model is used to predict the temperature profile of the Xilinx™ SPARTAN-3E Field-Programmable Gate Array (FPGA) board during one day. In addition, thermal measurements based on infrared thermography are performed to validate our thermal profiles. These predicted profiles are compared to the temperature maps obtained with a thermal camera over 24 hours. A good agreement, with a maximum error of 1.8 °C, between the predicted and measured temperatures is obtained, which helps a lot in the proper functioning and the thermal management of the system-on-chips (SoC).