An Efficient Parameter Identification Approach for 3-D Cauer Model of Multichip IGBT Module Based on Cooperative Co-evolution Algorithm
IEEE Transactions on Power Electronics · 2026-08 · 41 (8) · 13520-13535
DOI: 10.1109/tpel.2026.3674738 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Power Electronics, 2026-08. Verfasst von Hengzhi Liu und 2 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben. Gefördert von: Beijing Natural Science Foundation, Systematic Key Project of China State Railway Group Co., Ltd..
Abstract (Originalsprache)
Accurate identification of 3-D Cauer model parameters is crucial for obtaining the junction temperature distribution of multichip IGBT modules and subsequently assessing their reliability. However, existing methods suffer from low computational efficiency and limited applicability. Thus, an efficient parameter identification method for 3-D Cauer model of multichip IGBT modules is proposed. The proposed method transforms the parameter identification problem into a large-scale global optimization (LSGO) problem. By constructing the fitness function based on the temperature responses obtained from both the 3-D Cauer model and one-time finite element analysis, the method achieves high computational efficiency. For the first time, a cooperative co-evolution (CC)-based LSGO algorithm, with high universality for decision variables across diverse dimensions, is employed to identify 3-D Cauer model parameters. Within the algorithm, problem decomposition considers the inherent partitioned structure of the 3-D Cauer model for multichip IGBT modules, effectively preserving internal physical correlations and enhancing both convergence speed and solution quality. In addition, the probabilistic contribution-based computational resource allocation strategy prioritizes high-contribution subproblems for resource efficiency, while enabling low-contribution ones to run and thus avoiding local optima. Finally, the effectiveness and superiority of the proposed method were validated using experimental data from a 6500-V/750-A multichip IGBT module.
Förderung
- Beijing Natural Science Foundation
- Systematic Key Project of China State Railway Group Co., Ltd.
Zitieren
APA
Liu, H., Zhang, H., & Tang, Y. (2026). An Efficient Parameter Identification Approach for 3-D Cauer Model of Multichip IGBT Module Based on Cooperative Co-evolution Algorithm. IEEE Transactions on Power Electronics, 41(8), 13520-13535. https://doi.org/10.1109/tpel.2026.3674738
Vancouver
Liu H, Zhang H, Tang Y. An Efficient Parameter Identification Approach for 3-D Cauer Model of Multichip IGBT Module Based on Cooperative Co-evolution Algorithm. IEEE Transactions on Power Electronics. 2026;41(8):13520-13535. doi:10.1109/tpel.2026.3674738
BibTeX
@article{Liu2026,
author = {Liu, Hengzhi and Zhang, He-Sheng and Tang, Yicong},
title = {An Efficient Parameter Identification Approach for 3-D Cauer Model of Multichip IGBT Module Based on Cooperative Co-evolution Algorithm},
journal = {IEEE Transactions on Power Electronics},
year = {2026},
volume = {41},
number = {8},
pages = {13520--13535},
doi = {10.1109/tpel.2026.3674738}
}
EUVIMED