Computational-Efficient Thermal Estimation for IGBT Modules under Periodic Power Loss Profiles in Modular Multilevel Converters

One of the next challenges for modular multilevel converters (MMCs) is how to the size the critical components to their limits with reduced design margins, while at the same time fulfilling the reliability target. To do it confidently, better thermal modeling with quantitative uncertainty analysis is necessary to support the decision-making during the product design stage. Regarding the conversion of long-term mission profiles into thermal profiles for reliability evaluation, this paper proposes a computationally efficient thermal modeling method for insulated-gate bipolar transistor modules in MMCs. The proposed method considers the impact of the inherent thermal unbalance and has minimum computation with quantitative error analysis. Finally, both simulations and experiments have verified the theoretical results.