System Identification and Optimization Using ARX Models for Thermal Management in WBG-based Power Electronics Applications

In this paper, we delve into the complexities of thermal management in power electronics, spotlighting the introduction of Wide Band Gap (WBG) components which, despite their superior performance, introduce challenges in managing increased heat flux densities. Temperature modeling emerges as a critical task due to its pivotal role in ensuring device reliability and performance. Traditional thermal models offer valuable insights during the design phase, yet they often necessitate a compromise between complexity and speed/accuracy, especially when accounting for thermal boundary conditions and thermal coupling. We propose the use of a black-box model, specifically AutoRegressive with eXogenous inputs (ARX) models, to adeptly identify and characterize the thermal behavior of systems post-implementation. The article addresses two main subjects: a detailed exposition of the mathematical development behind ARX models and a comprehensive methodology tailored for thermal modeling in electronic systems. Experimental results underscore the efficacy of our approach, demonstrating that ARX models can accurately estimate the temperatures of electronic boards under a variety of conditions.