数字孪生驱动的高可靠性电力电子系统特刊主编寄语

Motivated by the demand of developing zero-emission sustainable energy across the world, many industry sectors such as transportation and utility power generation are experiencing rapid revolution, in which power electronics is playing a backbone role in numerous power conversion systems such as electric vehicles, electric ships, airplanes, solar/wind power generations, and energy storage. However, the reliability of power electronics has not received sufficient attention, especially for safety-critical applications where reliability is the first priority in power electronic designs. On the other hand, to embrace the industrial revolutions (i.e., Industry 4.0 and 5.0) that focus heavily on interconnectivity, automation, intelligence, and real-time condition monitoring for all the physical systems such as power electronic converters, digital online preventive maintenance, and optimization will be of paramount importance. The reason is that, it will not only fundamentally transform the accuracy and detection speed of health monitoring of power converter systems, but also streamline an industrial product cycle from the stages of conceptual design, and smart manufacturing, to operation and maintenance (O&M). Conventional offline and sensor-based health monitoring technologies perhaps are not well compatible and synergized to the emerging technologies of interconnectivity, automation, as well as the intelligent Q&M required by Industry 4.0 and 5.0. Therefore, transformative high-reliability design, optimization, and health monitoring methodologies that are synergized with the upcoming industrial digital technologies will be imperative for next-generation power electronic systems. To fundamentally improve the reliability of next-generation power converters, predictive and diagnostic digital twin methods are required. A digital twin is a digital replica (i.e., interactive digital models) of a physical system which can accurately predict and reflect the real-time health condition of the physical systems (i.e., physical twin), with real-time bi-directional data stream between the physical components or systems and the associated digital twin models. One major difference between a conventional simulation model and a digital twin model is that, a digital twin model is embeddable, dynamic, and interactive, which generally requires real-time sensory data to automatically update the models based on various operating conditions and mission profiles to achieve high accuracy and robustness, while the digital twin model can provide predictive operation command (e.g., fault-tolerant controls) to the physical systems for optimal operation. To promote the related research activities and summon scientific articles from the global professional communities on this emerging topic, this Special Issue on “Digital Twin Driven High-Reliability Power Electronic Systems” was launched in December 2023, based on the encouragement and support from the editorial board of IEEE Journal of Emerging and Selected Topics in Power Electronics (JESTPE). Although the topic of this Special Issue is still relatively new in the area of power electronics, there have been 18 articles submitted to this special issue, and eventually ten articles from various countries or regions were accepted based on a standard IEEE JESTPE peer review process. A brief overview of these ten accepted articles is provided below, for the convenience of researchers and readers in this area. A data-driven noninvasive digital twin approach is introduced in [A1] for estimating critical degradation parameters in the main power components (semiconductor switches, dc capacitors, and inductors) in a dc–dc Buck converter. Specifically, the proposed digital twin models estimate the parasitic resistances of passive elements and on-state resistance ( R ds−on ) of the MOSFET, as a real-time health indicator of converters’ reliability during operation. In [A2], parameter identification methods for digital twins of Buck converters and adaptive compensator parameter tuning techniques are introduced. Compared to the traditional particle swarm optimization (PSO), the proposed two-stage metaheuristic PSO method reduces the parasitic resistance estimation errors of the MOSFETs and inductors in the Buck converter from 31% and 45% to 1.5% and 2.3%, respectively, and reduces computational time by over 60%. In [A3], a PSO-based digital twin estimation method is presented to estimate the on-state resistance of the semiconductor switches and dc-link capacitance in a six-phase ac–dc boost rectifier for wind offshore applications. In [A4], a multitimescale digital twin concept is presented for comprehensive health and fault monitoring of dc–dc boost converter. The main technical aspects of the digital twin-based condition monitoring method are examined, including state-space modeling, optimization algorithms, fault detection approaches, and sensor placement strategies. In [A5], a digital twin-based fault diagnostic technique is introduced for identifying open-switch faults that could occur to a three-phase five-level active neutral point clamped (5L-ANPC) inverter. The digital twin diagnostic model is based on analyzing the existing real-time sensory information on dc capacitor voltages, load currents, and switching patterns, which exhibits fast and accurate detection speed, while requiring no additional sensors or hardware components. In [A6], a voxel-based method for digital twinning the thermal behavior of power converters is introduced, which provides a solution to monitor the 3-D temperature of the power converter in real-time. To more accurately represent the power converter and reduce the computation of digital twin modeling, an adaptive mesh technique is implemented into the voxel-based method which has reduced 75% of resources demanded by the computation of the three-phase LLC converter. In addition, considering that thermal performance, especially the junction temperature profile, is a key reliability indicator used for digital twin modeling of semiconductor power modules, innovative junction temperature estimation methods and electro-thermal-mechanical modeling of silicon carbide (SiC) MOSFET modules are presented in [A7] and [A8], respectively. In [A9], a digital twin LC parameter identification strategy based on comparing three optimization algorithms, namely, PSO, genetic algorithm (GA), and the simulated annealing (SA), is developed to estimate the characteristics of the L-type ac filter and the dc capacitance for a three-phase boost rectifier, which aims to serve as online health condition indicators of the power converters. Balanced and unbalanced operations of the converter have been tested to demonstrate the robustness and feasibility of the digital twin concept. In [A10], a data specification and cloud-based data streaming method are presented to manage and analyze operational data for implementing cloud-based battery management systems (BMSs), including power converters. The data specification method can accommodate the configuration and control parameters of modular multilevel converter-based reconfigurable battery packs. ACKNOWLEDGMENT We would like to acknowledge the effort from the authors for their contributions to this Special Issue and the numerous reviewers who have voluntarily provided constructive technical feedback to all the submitted papers. Furthermore, sincere thanks are due to the following Guest Associate Editors from various regions for their dedicated support to make this Special Issue possible (in alphabetical order). Mohammed Agamy, SUNY University at Albany, Albany, NY, USA Roberto Cardenas, University of Chile, Santiago, Chile Fang Luo, Stony Brook University, Stony Brook, NY, USA Hong Li, Zhejiang University, Hangzhou, China Margarita Norambuena, Universidad Técnica Federico Santa María, Valparaíso, Chile Deepak Ronanki, Indian Institute of Technology Madras, Chennai, India Kai Sun, Tsinghua University, Beijing, China Luca Solero, Roma Tre University, Rome, Italy Appendix: Related Articles S. Roy, M. Behnamfar, A. Debnath, and A. Sarwat, “Data-driven digital twin for reliability assessment of DC/DC buck converter,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2712--2724, Jun. 2025, doi: 10.1109/JESTPE.2024.3497772. Y. H. Liu, Z. Z. Yang, and M. C. Liu, “Digital twin-based online health monitoring of power electronics systems with self-evolving compensators and improved parameter identification capability,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2725–2737, Jun. 2025, doi: 10.1109/JESTPE.2024.3495017. G. Di Nezio, M. Di Benedetto, A. Lidozzi, and L. Solero, “Digital-twin based health monitoring for multi-phase boost rectifier in wind offshore applications,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2738–2748, Jun. 2025, doi: 10.1109/JESTPE.2024.3503759. K. Choksi, M. Hijikata, A. B. Mirza, A. Zhou, D. Singh, and F. Luo, “Multi-time-scale digital twin for health and fault monitoring of a boost converter,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2749–2765, Jun. 2025, doi: 10.1109/JESTPE.2024.3519255. M. T. Fard, B. J. Luckett, and J. He, “Digital twin enabled open-circuit fault diagnosis for five-level ANPC multilevel converters,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2766–2780, Jun. 2025, doi: 10.1109/JESTPE.2024.3468332. X. Mo, D. R. Linares, R. Ramos, and M. Vasić, “Towards embeddable digital twins: A voxel-based approach for power converter temperature monitoring,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2781–2798, Jun. 2025, doi: 10.1109/JESTPE.2025.3541343. Y. Tang, C. Zhan, L. Zhu, W. Wang, Y. Gou, and S. Ji, “Efficient junction temperature estimation of SiC power modules based on temperature-dependent lumped thermal model,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2799–2810, Jun. 2025, doi: 10.1109/JESTPE.2024.3470907. P. Wu et al., “Investigation on the electrical-thermal-mechanical performance of multi-chip SiC power device with cu clip interconnect,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2811–2819, Jun. 2025, doi: 10.1109/JESTPE.2025.3544806. G. Di Nezio, S. D. L. Diz, M. Di Benedetto, A. Lidozzi, E. J. B. Peña, and L. Solero, “LC parameters identification for a 3-phase AC–DC converter through digital twin modeling technique and optimization algorithms,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2820–2833, Jun. 2025, doi: 10.1109/JESTPE.2025.3534616. D. Karnehm et al., “Universal data specification and real-time data streaming architecture for cloud-based battery management systems,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 13, no. 3, pp. 2834–2844, Jun. 2025, doi: 10.1109/JESTPE.2024.3413163.