Xiaobing Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xiaobing Shen has a total of 142 co-authors. Here are the top co-authors by number of shared publications:
- Wilmar Martinez
Department of Electrical Engineering (ESAT), KU Leuven and EnergyVille, Leuven, Belgium
7 shared publications - Yu Zuo
Department of Electrical Engineering (ESAT), KU Leuven and EnergyVille, Leuven, Belgium
5 shared publications - Bangli Du
Department of Electrical Engineering, KU Leuven - Energyville, Leuven, Belgium
2 shared publications - Jiaze Kong
Department of Electrical Engineering, KU Leuven - Energyville, Leuven, Belgium
2 shared publications Alessio Giuffrida
Politecnico di Torino, Torino, Italy
1 shared publicationAlfonso Martinez
Wrth Elektronik, Waldenburg, Germany
1 shared publication
Latest Publications
- 16 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 41 (9) , 16164-16177
Integration of a Hybrid POD-CGaNet Model and NSGA-II for Efficient PCB Thermal Analysis and Multiobjective Heatsink Optimization
This paper proposes a combination of Proper Or thogonal Decomposition (POD) and Conditional Generative Ad versarial Neural Networks (CGaNet) to predict the steady-state temperature distribution of PCB and optimize heatsink design under forced convection conditions. In comparison to supervised models, the POD-CGaNet offers the advantages of accurate predictions over a wide data range while relying on a small range of training data. In contrast to an unsupervised temper ature prediction model, the POD-CGaNet features a few model parameters, low hardware requirements, and short training time. Furthermore, to explore how input dimensionality within the extrapolation range influences prediction accuracy, this work demonstrates that POD-CGaNet can achieve an accuracy of R2 = 0.9 even when all input parameters fall within extrapolated ranges, outperforming the other models in comparison. The prediction results are consistent with numerical analysis and ex perimental results under two different cases. Due to its low error and fast prediction, POD-CGaNet can be integrated into the Non dominated Sorting Genetic Algorithm II (NSGA-II) framework as a surrogate model instead of commercial software in heatsink design process. Within large design space, the algorithm can reduce 95% optimization time compared to FEM methods.
- 16 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 41 (9) , 16164-16177
Integration of a Hybrid POD-CGaNet Model and NSGA-II for Efficient PCB Thermal Analysis and Multiobjective Heatsink Optimization
This paper proposes a combination of Proper Or thogonal Decomposition (POD) and Conditional Generative Ad versarial Neural Networks (CGaNet) to predict the steady-state temperature distribution of PCB and optimize heatsink design under forced convection conditions. In comparison to supervised models, the POD-CGaNet offers the advantages of accurate predictions over a wide data range while relying on a small range of training data. In contrast to an unsupervised temper ature prediction model, the POD-CGaNet features a few model parameters, low hardware requirements, and short training time. Furthermore, to explore how input dimensionality within the extrapolation range influences prediction accuracy, this work demonstrates that POD-CGaNet can achieve an accuracy of R2 = 0.9 even when all input parameters fall within extrapolated ranges, outperforming the other models in comparison. The prediction results are consistent with numerical analysis and ex perimental results under two different cases. Due to its low error and fast prediction, POD-CGaNet can be integrated into the Non dominated Sorting Genetic Algorithm II (NSGA-II) framework as a surrogate model instead of commercial software in heatsink design process. Within large design space, the algorithm can reduce 95% optimization time compared to FEM methods.
- 22 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Emerging and Selected Topics in Power Electronics
- Vol. 14 (2) , 1473-1483
Dual-Mode Interleaved IPOS LLC Converter With 200–900 V Output for EV DC Charging
Wide-voltage and high-power operation is increasingly required in grid-tied renewable-energy interfaces such as electric-vehicle charging systems, where LLC resonant converters are widely adopted due to their high efficiency and soft-switching capability. However, conventional LLC converters require wide frequency variation for voltage regulation, complicating magnetics and reducing efficiency. While single-module modifications and two-stage approaches can extend the gain range within a narrower frequency band, this comes at the cost of control complexity, and efficiency, while the converters must still handle the full voltage and power burden. Different from existing approaches, this work presents an interleaved IPOS LLC resonant converter that introduces module-level structural reconfigurability, enabling high power output and wide-voltage operation without requiring extreme switching-frequency range. Unlike existing single-module or two-stage approaches, the proposed architecture distributes the voltage span across two LLC modules, providing both scalable high-power capability and lower secondary-side voltage stress. In addition, the control strategy is simplified by employing a secondary-side on-off switch to transition between two operation modes, while pulse-frequency modulation regulates the output voltage in each mode, ensuring soft switching across the voltage range Moreover, automatic voltage balancing and current sharing are achieved through coupled inductors. A quantitative design methodology is introduced that considers both voltage gain characteristics and the current balancing ratio, ensuring reliable operation of all modules and power devices. Finally, the proposed innovations are experimentally verified on a 1 kW prototype with 400 V input and 200–900 V output demonstrates a peak efficiency of 96.16%.
- 22 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Emerging and Selected Topics in Power Electronics
- Vol. 14 (2) , 1473-1483
Dual-Mode Interleaved IPOS LLC Converter With 200–900 V Output for EV DC Charging
Wide-voltage and high-power operation is increasingly required in grid-tied renewable-energy interfaces such as electric-vehicle charging systems, where LLC resonant converters are widely adopted due to their high efficiency and soft-switching capability. However, conventional LLC converters require wide frequency variation for voltage regulation, complicating magnetics and reducing efficiency. While single-module modifications and two-stage approaches can extend the gain range within a narrower frequency band, this comes at the cost of control complexity, and efficiency, while the converters must still handle the full voltage and power burden. Different from existing approaches, this work presents an interleaved IPOS LLC resonant converter that introduces module-level structural reconfigurability, enabling high power output and wide-voltage operation without requiring extreme switching-frequency range. Unlike existing single-module or two-stage approaches, the proposed architecture distributes the voltage span across two LLC modules, providing both scalable high-power capability and lower secondary-side voltage stress. In addition, the control strategy is simplified by employing a secondary-side on-off switch to transition between two operation modes, while pulse-frequency modulation regulates the output voltage in each mode, ensuring soft switching across the voltage range Moreover, automatic voltage balancing and current sharing are achieved through coupled inductors. A quantitative design methodology is introduced that considers both voltage gain characteristics and the current balancing ratio, ensuring reliable operation of all modules and power devices. Finally, the proposed innovations are experimentally verified on a 1 kW prototype with 400 V input and 200–900 V output demonstrates a peak efficiency of 96.16%.
- Open Access
- 10 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 194304-194313
Extrapolation Ability of Denoising Diffusion Probabilistic Models for High-Frequency Inductor Core Loss Predictions With DC Bias
Accurate prediction of high-frequency magnetic core losses under DC bias conditions remains a critical challenge in power electronics design. Conventional models, such as empirical curve-fitting approaches and neural networks, often struggle with extrapolation beyond training data, particularly under varying DC bias levels. This paper proposes a Denoising Diffusion Probabilistic Model (DDPM) to enhance the extrapolation capability for core loss predictions in ferrite materials, N87 and 3C90, under DC bias. The proposed DDPM leverages a two-phase noise injection and denoising process to improve prediction stability and generalization. Experimental validation demonstrates that the proposed model achieves an R2 score of 0.9999 for interpolation and maintains high accuracy (R2 = 0.9996)in extrapolation tasks, significantly outperforming conventional CNN (R2 = 0.1856) and GANET (R2 = 0.7974) models. Additionally, the DDPM model enables high-fidelity data augmentation, generating realistic core loss datasets for unseen DC bias conditions. These results highlight the DDPM’s potential as a robust tool for magnetic component design, offering improved accuracy and reliability in both interpolation and extrapolation scenarios.
- 10 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 194304-194313
Open AccessExtrapolation Ability of Denoising Diffusion Probabilistic Models for High-Frequency Inductor Core Loss Predictions With DC Bias
Accurate prediction of high-frequency magnetic core losses under DC bias conditions remains a critical challenge in power electronics design. Conventional models, such as empirical curve-fitting approaches and neural networks, often struggle with extrapolation beyond training data, particularly under varying DC bias levels. This paper proposes a Denoising Diffusion Probabilistic Model (DDPM) to enhance the extrapolation capability for core loss predictions in ferrite materials, N87 and 3C90, under DC bias. The proposed DDPM leverages a two-phase noise injection and denoising process to improve prediction stability and generalization. Experimental validation demonstrates that the proposed model achieves an R2 score of 0.9999 for interpolation and maintains high accuracy (R2 = 0.9996)in extrapolation tasks, significantly outperforming conventional CNN (R2 = 0.1856) and GANET (R2 = 0.7974) models. Additionally, the DDPM model enables high-fidelity data augmentation, generating realistic core loss datasets for unseen DC bias conditions. These results highlight the DDPM’s potential as a robust tool for magnetic component design, offering improved accuracy and reliability in both interpolation and extrapolation scenarios.
- 26 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 40 (10) , 15254-15264
Iron Loss Extrapolation Predictions for High-Frequency Magnetic Components Using Denoising Diffusion Models
This article introduces denoising diffusion probabilistic models to enhance core loss predictions in high-frequency magnetic components. Conventional methods, such as the Steinmetz equation, often fail to accurately capture the nonlinear dynamics and complex waveforms characteristic of high-frequency magnetic core losses. Previous approaches using multilayer perceptron, transfer learning, and generative adversarial networks have encountered issues, such as data boundaries and noise, adversely affecting prediction accuracy. Denoising diffusion probabilistic models overcome these challenges by improving both interpolation and extrapolation capabilities, achieving a minimum average relative error of 0.99% across all datasets provided by the MagNet challenge. Furthermore, denoising diffusion probabilistic models, first introduced into power electronics, support effective data augmentation, validated through the generation of high-quality core loss data for Nanocrystalline ring cores in three different sizes (R16, R25, R50). This robust modeling framework not only reduces prediction errors but also enhances dataset comprehensiveness, thereby facilitating the design and optimization of next-generation power electronic components.
- 26 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 40 (10) , 15254-15264
Iron Loss Extrapolation Predictions for High-Frequency Magnetic Components Using Denoising Diffusion Models
This article introduces denoising diffusion probabilistic models to enhance core loss predictions in high-frequency magnetic components. Conventional methods, such as the Steinmetz equation, often fail to accurately capture the nonlinear dynamics and complex waveforms characteristic of high-frequency magnetic core losses. Previous approaches using multilayer perceptron, transfer learning, and generative adversarial networks have encountered issues, such as data boundaries and noise, adversely affecting prediction accuracy. Denoising diffusion probabilistic models overcome these challenges by improving both interpolation and extrapolation capabilities, achieving a minimum average relative error of 0.99% across all datasets provided by the MagNet challenge. Furthermore, denoising diffusion probabilistic models, first introduced into power electronics, support effective data augmentation, validated through the generation of high-quality core loss data for Nanocrystalline ring cores in three different sizes (R16, R25, R50). This robust modeling framework not only reduces prediction errors but also enhances dataset comprehensiveness, thereby facilitating the design and optimization of next-generation power electronic components.
- 19 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 40 (3) , 3838-3843
Reconfigurable LLC Resonant Converter for Wide Voltage Range and Reduced Voltage Stress in DC-Connected EV Charging Stations
To meet the charging demands of current electric vehicles (400V and 800V) and next generation powertrain architectures (1200V), this letter introduces a reconfigurable LLC resonant converter connected to DC charging stations, capable of supporting a wide output voltage range for low-, medium-, and high-voltage EV charging by selecting various inverter and rectification methods. Unlike traditional LLC pulse frequency modulation approaches, the proposed converter maintains high efficiency across the entire output range and operates at a narrow switching frequency. The converter design features a stacked structure on both primary and secondary sides, including stacked half-bridges, diodes, and capacitors, which halves the voltage stresses on switches and diodes. This configuration also possesses voltage self-balancing among the stacked capacitors. Additionally, soft-switching is achieved throughout the wide voltage range. Finally, experimental results from a prototype with an 8:1 output voltage ratio validate the feasibility of the proposed scheme.
- 19 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 40 (3) , 3838-3843
Reconfigurable LLC Resonant Converter for Wide Voltage Range and Reduced Voltage Stress in DC-Connected EV Charging Stations
To meet the charging demands of current electric vehicles (400V and 800V) and next generation powertrain architectures (1200V), this letter introduces a reconfigurable LLC resonant converter connected to DC charging stations, capable of supporting a wide output voltage range for low-, medium-, and high-voltage EV charging by selecting various inverter and rectification methods. Unlike traditional LLC pulse frequency modulation approaches, the proposed converter maintains high efficiency across the entire output range and operates at a narrow switching frequency. The converter design features a stacked structure on both primary and secondary sides, including stacked half-bridges, diodes, and capacitors, which halves the voltage stresses on switches and diodes. This configuration also possesses voltage self-balancing among the stacked capacitors. Additionally, soft-switching is achieved throughout the wide voltage range. Finally, experimental results from a prototype with an 8:1 output voltage ratio validate the feasibility of the proposed scheme.
- Open Access
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Power Electronics
- Vol. 6, 883-898
MagNet Challenge for Data-Driven Power Magnetics Modeling
This paper summarizes the main results and contributions of the MagNet Challenge 2023, an open-source research initiative for data-driven modeling of power magnetic materials. The MagNet Challenge has (1) advanced the stateof-the-art in power magnetics modeling; (2) set up examples for fostering an open-source and transparent research community; (3) developed useful guidelines and practical rules for conducting data-driven research in power electronics; and (4) provided a fair performance benchmark leading to insights on the most promising future research directions. The competition yielded a collection of publicly disclosed software algorithms and tools designed to capture the distinct loss characteristics of power magnetic materials, which are mostly open-sourced. We have attempted to bridge power electronics domain knowledge with state-of-the-art advancements in artificial intelligence, machine learning, pattern recognition, and signal processing. The MagNet Challenge has greatly improved the accuracy and reduced the size of data-driven power magnetic material models. The models and tools created for various materials were meticulously documented and shared within the broader power electronics community.
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Power Electronics
- Vol. 6, 883-898
Open AccessMagNet Challenge for Data-Driven Power Magnetics Modeling
This paper summarizes the main results and contributions of the MagNet Challenge 2023, an open-source research initiative for data-driven modeling of power magnetic materials. The MagNet Challenge has (1) advanced the stateof-the-art in power magnetics modeling; (2) set up examples for fostering an open-source and transparent research community; (3) developed useful guidelines and practical rules for conducting data-driven research in power electronics; and (4) provided a fair performance benchmark leading to insights on the most promising future research directions. The competition yielded a collection of publicly disclosed software algorithms and tools designed to capture the distinct loss characteristics of power magnetic materials, which are mostly open-sourced. We have attempted to bridge power electronics domain knowledge with state-of-the-art advancements in artificial intelligence, machine learning, pattern recognition, and signal processing. The MagNet Challenge has greatly improved the accuracy and reduced the size of data-driven power magnetic material models. The models and tools created for various materials were meticulously documented and shared within the broader power electronics community.
- 6 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 39 (8) , 9953-9964
Conditional Generative Adversarial Network Aided Iron Loss Prediction for High-Frequency Magnetic Components
This article tackles the complex challenge of predicting magnetic iron losses in high-frequency magnetic components by introducing a novel conditional generative adversarial network model. Diverging from traditional loss prediction methodologies that often overlook intricate interactions of factors, our conditional generative adversarial network framework is designed to comprehensively incorporate diverse aspects such as material properties, geometrical variations, and environmental conditions. To facilitate this advanced approach, a specialized four-wire measurement kit was employed, which significantly enriched the training dataset with a wide range of measurements. When benchmarked against conventional deep neural network models, the conditional generative adversarial network not only achieves faster convergence but also demonstrates markedly superior accuracy in predicting iron losses. This superiority is particularly notable in scenarios that extend beyond the training data's range, underscoring the model's robustness and adaptability. Such advancements in predictive accuracy and efficiency represent a significant leap forward in the design and optimization of high-frequency magnetic components.
- 6 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 39 (8) , 9953-9964
Conditional Generative Adversarial Network Aided Iron Loss Prediction for High-Frequency Magnetic Components
This article tackles the complex challenge of predicting magnetic iron losses in high-frequency magnetic components by introducing a novel conditional generative adversarial network model. Diverging from traditional loss prediction methodologies that often overlook intricate interactions of factors, our conditional generative adversarial network framework is designed to comprehensively incorporate diverse aspects such as material properties, geometrical variations, and environmental conditions. To facilitate this advanced approach, a specialized four-wire measurement kit was employed, which significantly enriched the training dataset with a wide range of measurements. When benchmarked against conventional deep neural network models, the conditional generative adversarial network not only achieves faster convergence but also demonstrates markedly superior accuracy in predicting iron losses. This superiority is particularly notable in scenarios that extend beyond the training data's range, underscoring the model's robustness and adaptability. Such advancements in predictive accuracy and efficiency represent a significant leap forward in the design and optimization of high-frequency magnetic components.
- 25 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 39 (7) , 8478-8496
Artificial Intelligence Applications in High-Frequency Magnetic Components Design for Power Electronics Systems: An Overview
The paper provides an overview of how Artificial Intelligence (AI) is applied in designing highfrequency magnetic components, primarily high-frequency inductors and transformers, for power electronics systems. Four categories of AI, including expert systems, fuzzy logic, metaheuristic methods, and machine learning techniques, are addressed. Firstly, AI models for estimating losses in high-frequency magnetic components are discussed. Subsequently, AI-based design methods in high-frequency inductors and transformers are observed. Then, AI tools applied to the automatic design of high-frequency magnetic components are introduced and compared. Drawing insights from an analysis of over 200 publications, the paper highlights significant advancements: the development of AI-driven models for precise loss estimation in highfrequency magnetic components, the application of AI in optimizing design configurations for the components, and the automation of design processes. These achievements demonstrate AI's capability to enhance the efficiency, performance, and innovation in high-frequency magnetic component design, offering a roadmap for future research in power electronics systems.
- 25 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 39 (7) , 8478-8496
Artificial Intelligence Applications in High-Frequency Magnetic Components Design for Power Electronics Systems: An Overview
The paper provides an overview of how Artificial Intelligence (AI) is applied in designing highfrequency magnetic components, primarily high-frequency inductors and transformers, for power electronics systems. Four categories of AI, including expert systems, fuzzy logic, metaheuristic methods, and machine learning techniques, are addressed. Firstly, AI models for estimating losses in high-frequency magnetic components are discussed. Subsequently, AI-based design methods in high-frequency inductors and transformers are observed. Then, AI tools applied to the automatic design of high-frequency magnetic components are introduced and compared. Drawing insights from an analysis of over 200 publications, the paper highlights significant advancements: the development of AI-driven models for precise loss estimation in highfrequency magnetic components, the application of AI in optimizing design configurations for the components, and the automation of design processes. These achievements demonstrate AI's capability to enhance the efficiency, performance, and innovation in high-frequency magnetic component design, offering a roadmap for future research in power electronics systems.