Han Cui

This article summarizes the discussions and outcomes of the ”Power for AI and AI for Power” session at the IEEE Power Electronics Society (PELS) Future of Electronic Power Processing and Conversion (FEPPCON XII). Power electronics embeds intelligence into energy systems. Modern artificial intelligence (AI) systems place massive demand on electrical power, and in return, introduce new opportunities in improving energy systems design and implementation. Power electronics, as sensors and actuators for all types of energy systems, need to embrace AI to address large-scale societal challenges such as industrial decarbonization, transportation, and climate change. The ”Power for AI and AI for Power” session generated lots of excitement in FEPPCON XII and will have a continuous impact on the future development of power electronics.

Nowadays, interleaved boost converter (IBC) is widely used in electric vehicle on-board power supplies, which puts forward high requirements for the efficiency and power density of the converter due to high power and limited interior space. Consequently, a three-phase coupled inductor with ladder magnetic core is adopted in this paper and a multi-objective optimization approach based on genetic algorithm (GA) is proposed for converter efficiency and coupled inductor volume. Firstly, the comprehensive objective function and corresponding constraint condition are formulated for the coupled inductor structure to facilitate global optimization. During the modeling process, a source controlled by the inductor voltage is introduced into the magnetic reluctance model to analyze AC magnetic flux. On this foundation, GA is employed to optimize the objective function, which improves the optimization efficiency significantly and the Pareto front of converter efficiency and coupled inductor volume can be obtained according to the optimization results. Finally, an experimental prototype of 30kW is established with a power density of 22kW/L and a maximum efficiency of 98.6%. Compared with non-coupled inductor, the volume is reduced by 43.8% and the power density is increased by 12.8%.

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.

The wide applications of wide-bandgap (WBG) devices in power electronics systems bring benefits in higher switching speed, frequency, and power density, but also cause severe electromagnetic interference (EMI) issues. While many EMI mitigation methods are available to reduce the noise generated from the source (i.e. switching WBG devices), it is equally important to analyze and control the EMI propagation paths. In WBG systems, the parasitic and coupling effects become dominant due to the high switching frequency and the compact system size, which creates uncontrolled propagation paths that degrade the systems' EMI performance significantly. Therefore, it is crucial to investigate the parasitic and coupling effects and develop corresponding mitigation techniques. This review focuses on analyzing the EMI propagation paths to provide a comprehensive guideline to identify possible parasitics and couplings in the power module package level, EMI filter level, and the power electronics system level. The corresponding mitigation techniques for common-mode and differential-mode conducted EMI as well as radiated EMI are summarized, and the remaining challenges and future research topics are also discussed for all three levels in the conclusion.

Accurately measuring the switching current behavior of wide bandgap devices requires a current sensor with a wide bandwidth covering dc to high frequency ranges. However, traditional Rogowski coils (RCs) for current sensing either lack the dc measurement capability or the inclusion of dc sensor largely increases the insertion inductance. This article presents a concept of combining a shielded RC with a dc sensor to improve the dc measurement capability without interfering with the power loop area of a double pulse test (DPT) circuit. The RC and dc sensor are positioned at different locations to capture the high-frequency and low-frequency components, respectively. An analog circuit combines the output of both sensors to recreate the signal. Frequency domain analysis of the prototype confirms the improved dc measurement of the probe. This article also presents the results of a DPT of a half-bridge SiC power module to demonstrate the enhanced dc measurement of the shielded RC, without any parasitic interference in the power loop area.

The progress made in wide bandgap (WBG) semiconductors has resulted in rapid miniaturization and increased efficiency of power converters. However, the improvements in magnetic components, such as inductors and transformers, have not kept pace with these advancements [1] , [2] , [3] . Although advances in WBG devices, novel topologies, control schemes, and hardware fabrications have greatly improved circuit efficiency and power density, the bottleneck now lies with magnetic components [4] , with magnetics accounting for more than 30% of the cost and more than 30% of the loss in almost all power converters [5] . Magnetics design has become a critical issue for power electronics as trends towards high efficiency and high power-density.

Lateral gallium nitride (GaN) high-electron-mobility transistors present better electrical characteristics compared to silicon or silicon carbide devices such as high switching speed and low gate charge, but also present additional challenges on the module design. This article discusses a high-density GaN power module with double-sided cooling, low inductance, on-package decoupling capacitors, and integrated gate drivers. The GaN dies as well as the gate drive are sandwiched between the printed circuit board and direct bonded copper substrate to achieve compact loop and double-sided cooling effect. Design considerations and thermal performance are analyzed. A module assembly procedure is presented utilizing the layer-by-layer attachment process. Finally, a 2.7 cm × 1.8 cm half-bridge GaN power module is fabricated and tested, achieving a low power-loop inductance of 1.03 nH, and the overshoot voltage of the switching waveform is less than 5% under a 400 V/25 A double-pulse test. The thermal resistance is 0.32 K/W, verified by simulation and experimental results. The design and assembly process can be generalized and applied to high power applications to achieve high power density and high performance.

A compact nonlinear circuit model is developed for frequency-selective limiters (FSLs) based on thin film magnetic materials. The model contains a three-port coupled inductor circuit called a spin unit that represents the dynamic spin precession behaviors of a single spin, which is derived rigorously from the Landau-Lifshitz–Gilbert (LLG) equation. The spin units are then combined with external circuitry to represent the coupling between the RF signal in the form of electromagnetic (EM) waves and the spin waves. Specifically, two types of coupling are considered, parallel pumping and perpendicular pumping. The former contributes to the power-limiting effects of FSLs, while the latter accounts for the small signal insertion loss. In addition, the RF magnetic fields provided by coplanar waveguide (CPW)-FSLs are spatially nonuniform. Such nonuniformity can be modeled by creating an array of inductors based on the field distribution in both the width and the thickness directions. The circuit model is implemented in the advanced design system (ADS) and is capable of predicting the small signal $S$ -parameters, large signal insertion loss, power threshold, time delay, intermodulation (IM) spectrum, and frequency selectivity for CPW-FSLs. The model may also be used for modeling other RF magnetic devices based on similar physics.

A uniform-flux inductor design strategy is described in this paper to improve the gravimetric power density of the common inductor of a 32 kW, 350 V dual-interleaved Buck-Boost DC-DC converter. The strategy consists of an unusual, enclosed winding arrangement that virtually distributes the core magnetic flux density outcoming in a more compact inductor. The resultant common inductor weight and volume are reduced by 41 % and 24 % respectively, the power density being increased to 7.9 kW/kg. A COMSOL numerical prediction confirms almost uniform flux density distribution along the magnetic core axes. Estimation of the inductance, power losses and thermal behaviour are also analysed. Construction assessment is described for a 7 μH, 150 kHz, 180 A inductor, which is experimentally verified in a 350 V, 32 kW prototype. The experiments reveal that the built uniform-flux inductor properly operates ensuring both electrical and thermal stability.

While the employment of wide bandgap (WBG) devices in high-frequency and high-voltage applications brings benefits such as reduced system size and improved efficiency, it aggravates the electromagnetic interference (EMI) issue due to fast switching. High-frequency EMI noise suppression relies mainly on the filter design, where the filter's performance is strongly affected by parasitics. Through analyzing the common-mode (CM) equivalent circuit of a half-bridge power module, this letter identifies the key parasitics that dominate the performance of a common-mode filter (CMF) at high frequencies. To minimize the parasitics, the concept of integrating the CMF inside the WBG power module package is developed to improve the noise attenuation. A π-type CMF is integrated with a half-bridge GaN-based power module as a prototype to validate the concept. Experiments are conducted by measuring the CM noise spectrum received by the line impedance stabilization networks (LISNs) from the hard switching of the designed power module under 70 V and 80 kHz. Comparing the measured results of the integrated CMF to the externally-added CMF, up to 50 dBμV more attenuation is achieved by the integrated CMF in the frequency range of 10 MHz to 100 MHz, verifying the theoretical analysis and the established CM equivalent circuit.