ABSTRACT. Power electronics is a foundational technology that drives a wide range of important and emerging applications including cloud computing, wireless communications, robotics, and smart energy systems. By systematically managing the increased complexity in materials, circuits, and systems, new opportunities are created to greatly advance the functionality and performance of power electronics systems. This speech provides a few examples to illustrate the potential of managed complexity in power electronics design. These include: 1) artificial intelligence and machine learning for managed complexity in passive component modeling; 2) modular and scalable architecture for managed complexity in high performance circuits; and 3) integrated electrical-mechanical co-design for managed complexity in robotics and metamaterials. This managed complexity approach addresses key challenges in emerging applications by overcoming traditional design barriers from new angles and redefining how power electronics are conceived and implemented in complex systems.

ABSTRACT. High-frequency, high-power, and high-efficiency multi-phase (MP) power amplifiers are needed for industrial, scientific, and medical (ISM) applications. MP systems can provide more power than their single-phase counterparts while maintaining the same single-switch stress or achieving the same power output with reduced single-switch stress. However, existing research lacks the design method for three-phase current-mode power amplifiers (TP-CMPAs). We investigate modeling and analysis of TP-CMPAs by differential-mode and common-mode circuit decomposition and Y-Δ transformation. A systematic design method for the TP-CMPAs under an infinite quality factor (Q) with zero-voltage switching (ZVS) is introduced and a corresponding tuning method to maintain ZVS under both low Q and high Q conditions is developed. Thus, ZVS can be achieved under a wide resistive load range. Different resonant networks driven by TP-CMPAs are also analyzed. All TP-CMPAs achieve ZVS under a 5X resistive load range.

ABSTRACT. Accurate, rapid, and dynamically-controlled impedance matching is critical for many radio-frequency (RF) power applications. This work presents a solid-state electronically-variable reactance (eVX) actuator capable of continuous tuning with response times on the order of a single RF cycle. The actuator is based on Phase-Switched Impedance Modulation (PSIM), which switches passive elements at the RF frequency to modulate their impedances. Prior PSIM-based implementations have demonstrated the feasibility of fast, solid-state tunable matching networks (TMNs) for high-frequency applications at kilowatt power levels. This work introduces a PSIM architecture using a differential push-pull topology, significantly reducing switching harmonics, enhancing power handling, and improving design flexibility. This architecture is employed in the design of a 13.56 MHz eVX actuator targeting 0–5 Ω reactance modulation range and its performance is demonstrated in a 3.5 kW TMN for RF plasma applications, achieving microsecond-scale system tuning.

ABSTRACT. This paper presents a systematic design approach for high frequency switched-mode broadband converters. Closed-form design equations are derived, and a 200 W class-Φ2 prototype achieves 90% efficiency across a 1.6 MHz bandwidth centered at 13.56 MHz. Experimental results validate the methodology, offering a practical solution for efficient broadband high frequency power supplies.

ABSTRACT. Electromagnets in low-field portable MRI systems offer significant advantages due to their adjustable field strength and superior portability. Nonetheless, the development of the main field amplifier for such systems remains an area with scant research. This digest presents a pulsed magnet system designed for low-field portable MRI applications. Preliminary experiments of the prototype demonstrate the generation of consistent main magnetic field pulses of 27.44 mT by ramping the coil current to 26.5 A within 50 μs.

ABSTRACT. The presence of right-half-plane (RHP) zero and sampling delay in a digital current-mode controlled (DCMC) boost converter restricts the achievable closed-loop bandwidth (BW) to ${\omega _c} = {\omega _{\rm{rhp}}}/3$, making it impractical for high BW and efficient envelope tracking applications. State-feedback-based mixed-signal hysteresis current control (SFMSHCC) enhances the converter's dynamic performance and damping at higher BW without exhibiting any overshoot/undershoot. The resistive characteristics of fixed-gain power amplifiers offer the potential to further improve performance by incorporating reference feedforward. Steady-state frequency regulation is realized through real-time band adaptation. The performance improvements, supported by analytical predictions, are validated using simulation and experimental results.

ABSTRACT. This digest investigates a new integrated control method to enhance the transient response of variable-frequency, multiphase series-capacitor buck voltage regulation modules. The disturbance is first estimated by a trajectory-based controller, which explicitly computes the time-optimal recovery switching sequence for each phase operating in the MHz range. Controlled de-phasing bypasses the series capacitors and leverages the high input voltage to accelerate current delivery to the load. Any remaining error is treated as a small-signal perturbation and is rapidly rejected by a cycle-by-cycle digital controller designed in a switching-synchronized sampled state-space. Together, these strategies allow for the sub-microsecond load-step transient response. The proposed strategies are validated through simulation, demonstrating significant improvements in dynamic response. Ongoing hardware results support the control framework.

ABSTRACT. In this paper, our objective is to use notions of system energy to formulate closed-loop dc-dc converter controls that are nonlinear and satisfy passivity properties that guarantee stability. Port-Hamiltonian models are a particular form of models which can be used to describe the total energy in a converter, much like a Lyapunov function. In our approach, we first formulate a port-Hamiltonian model that represents the desired closed-loop dynamics we seek. However, the solution to this model is generally quite difficult even for the simplest of converters. To bypass this challenge, we offer a framework where a neural-network-based controller is trained to estimate the solution to this design problem. Essentially, our objective is to ensure that the energy dynamics of the dc-dc converter with a neural network as a controller closely matches that of the target port-Hamiltonian model. This method circumvents the mathematical difficulties encountered when attempting to solve the closed-loop port-Hamiltonian model directly and gives a generalized framework. Our paper illustrates this approach and its versatile application towards boost, buck, buck-boost, and Cuk converters.