Research and Technical Focus

Converter topology design has been a sustained focus of my research, from the way energy moves through an inductor to the power paths available in a multiport system. Across my publications, I have designed, modeled, and evaluated circuits with different priorities: high voltage gain, low input current ripple, bidirectional power flow, reduced device stress, and efficient switching. I am particularly interested in understanding what each circuit gains and what it asks of the rest of the system.

In high step-up DC–DC conversion, I have worked with current-fed structures, coupled and center-tapped inductors, interleaving, voltage multiplier cells, and resonant soft switching. My approach combines switching-interval analysis and design equations with simulation, comparison, and experimental evaluation. This has given me experience assessing a topology beyond its ideal voltage-gain expression, including the effects of magnetic behavior, semiconductor stress, losses, and component count.

My research also covers multiport and bidirectional converters for EV charging, hybrid powertrains, and AC/LVDC microgrids. I have studied how a topology can coordinate renewable sources, batteries, ultracapacitors, the grid, and a vehicle across charging, regenerative braking, and vehicle-to-grid modes. In integrated on-board charging, I have explored the use of existing motor windings as part of the charging circuit.

What motivates me is the design process itself: understanding why a circuit works, testing where its advantages hold, and finding ways to improve it for a real application. Each project has added a different perspective to how I select and develop power converter topologies.

Power semiconductor selection is where analytical design meets hardware reality. A MOSFET or diode may satisfy its nominal voltage and current ratings yet perform poorly once switching transients, temperature, reverse recovery, and changing power-flow directions are considered. I am interested in that gap between the datasheet and the operating circuit. My research has taught me to select devices by examining the stresses they experience throughout a switching cycle and then checking those predictions against measured behavior.

Across my high step-up converter studies, I have derived the blocking voltage and current stress of switches and diodes for different operating intervals. I examine how duty cycle, coupled-inductor winding ratios, voltage multiplier cells, and clamp circuits redistribute those stresses. Reducing the voltage across a main switch can create an opportunity to use a lower-voltage MOSFET with lower on-state resistance, but I also check its peak and RMS current, switching behavior, and rating margin. An improvement in one device is valuable only if it does not create a larger problem elsewhere in the circuit.

Loss analysis is another part of my approach. I estimate switch conduction and switching losses, diode forward losses, and their contribution to the converter’s overall efficiency. I then compare those estimates with measured switch and diode voltage and current waveforms where hardware results are available. This helps me understand whether a proposed topology has reduced total loss or shifted it among devices. It has also shaped how I compare converters: switch count alone cannot describe device ratings, current utilization, gate-drive needs, or practical cost.

I am particularly curious about switching transitions. In resonant and coupled-inductor converters, leakage and resonant energy can help reduce switching loss or ease diode reverse recovery. I study the conditions under which those benefits occur, as well as the extra circulating current or component stress they may introduce. Soft switching is most useful when its operating range and tradeoffs are understood, not merely identified at one operating point.

In EV chargers and multiport converters, I extend the same analysis across charging, discharging, regenerative, and bidirectional modes. A semiconductor that is well chosen for one power path may face its highest stress in another. My goal is to bring stress equations, loss models, gate-drive considerations, and experimental waveforms together to make informed device choices that support efficient and reliable hardware.

Modeling is how I determine whether a promising converter remains useful beyond a single operating point. I use it to trace energy through each switching interval, predict how currents and voltages change with load and control, and identify the conditions under which an assumed operating mode no longer holds. Across high-gain DC–DC converters, multiport chargers, microgrids, and electric drives, I have used mathematical analysis to turn circuit concepts into predictions that can be tested.

For a new converter topology, I begin with the equivalent circuits of its operating intervals. Volt-second and charge balance lead to expressions for voltage gain, capacitor voltages, and semiconductor stress. I then examine how winding ratios, leakage inductance, resonant paths, and switching frequency affect those results. In my high step-up converter research, I have also analyzed current ripple and the boundaries between continuous and discontinuous conduction. These models help me choose an operating range and component values while revealing when higher gain simply transfers stress or loss elsewhere.

Dynamic modeling becomes especially valuable in multiport systems, where a change at one port can affect the others. I have worked with bilinear, linearized, and small-signal models to study charger behavior and design current and voltage control loops. My research has also explored state-space, Euler–Lagrange, and energy-based descriptions for converter control. I enjoy these methods because each offers a different way to understand the same physical system and exposes different design limits.

The applications have broadened the questions I ask of a model. I have studied current transients during faults in AC/LVDC microgrids, converter dynamics and energy management in fuel-cell hybrid vehicles, and winding reconfiguration in electric drives. Across these projects, I compare analytical predictions with simulations and, where available, measured waveforms and operating results.

The most useful part of modeling is often finding where the first prediction falls short. A mismatch can reveal leakage effects, losses, a control limit, or an operating boundary that deserves closer attention. That process of deriving, testing, and refining a model keeps me curious and helps me make better engineering decisions.

A converter’s topology defines what it can do; its control system determines how well it performs as voltages, loads, and power requests change. I develop control strategies alongside the power circuit, starting with the operating modes and dynamic model, then selecting feedback signals, designing regulation loops, and testing responses to disturbances. This approach has shaped my research on high-gain converters, bidirectional chargers, microgrids, hybrid vehicles, and electric drives.

For DC–DC and multiport converters, I have worked with cascaded current and voltage loops, small-signal models, and frequency-response analysis. I use these tools to assess regulation, stability, and the interaction between ports. Load changes are particularly informative: they show whether a controller designed around one operating point can maintain the required behavior as the system moves away from it. I am interested in the connection between a topology’s physical limits and the performance a controller can realistically achieve.

In grid-connected chargers, my work has addressed grid-current shaping, DC-link voltage regulation, and active and reactive power control. Different charging modes call for different control structures, including current and voltage loops adapted to continuous or discontinuous conduction. Beyond individual loops, I have studied how charging systems select power sources and operating modes according to renewable generation, battery state of charge, grid availability, and vehicle demand.

I have also explored nonlinear control and energy-management methods for systems with strongly coupled states. These studies have given me experience comparing control approaches under changing loads and disturbances. In electric-drive research, control extended beyond feedback equations to a carefully timed sequence for changing motor-winding connections, updating controller settings, and resuming operation safely.

What motivates me is seeing a control concept survive contact with the full system. I enjoy moving from a dynamic model to a design that respects device and converter limits, responds well to disturbances, and can be assessed through simulation and measured behavior.

Magnetic components are central to how a power converter stores energy, controls current, and reaches its intended voltage range. I treat their design as part of the circuit architecture: inductance affects ripple and control dynamics, winding ratio affects gain and semiconductor stress, and core and copper losses influence efficiency. Across my research, I have worked with coupled inductors, multi-winding structures, input filters, and motor windings in high-gain converters and EV charging systems.

In high step-up DC–DC converters, I have explored two-, three-, and multi-winding coupled inductors, including center-tapped and interleaved arrangements. I analyze switching-interval currents, volt-second balance, ripple, and conduction-mode boundaries to understand how a winding choice changes the whole converter. These structures can provide high gain without relying only on an extreme duty cycle, but I also examine their effects on input current, device stress, and magnetic complexity.

I am particularly interested in magnetic behavior that an ideal circuit model can hide. Leakage inductance can create unwanted voltage stress, yet in a suitable resonant or clamp structure it can also contribute to the switching transition. I have studied both possibilities, along with core and winding losses, and compared calculated behavior with prototype results where available. That experience has made me careful about claims that a design is smaller or more efficient simply because it uses fewer magnetic components.

My work on integrated EV chargers and electric drives has broadened how I think about magnetics. I have investigated the use of motor stator windings as charging-circuit inductors and studied winding reconfiguration for different drive operating modes. In these systems, the machine, converter, and control strategy must be considered together.

What keeps me curious is how much a winding arrangement can change a converter’s behavior. I enjoy moving between magnetic models, circuit analysis, and measured waveforms to find a design that performs well across the operating range.

My broader goal is to use power electronics to advance technologies beyond the converter itself: efficient power delivery for AI data centers, more capable electric vehicles and charging systems, and cleaner energy systems that waste less power. I am actively looking for new applications where my experience in converter design, modeling, control, and hardware can make a measurable difference to computing, mobility, and the environment.

The applications that interest me most are those where a converter must do more than regulate a single output. It may need to coordinate several energy sources, respond to a changing load, reverse power flow, and remain efficient across very different operating modes. My published research has addressed these challenges in EV charging, renewable-energy systems, hybrid microgrids, fuel-cell vehicles, and electric drives. I am now looking to extend that experience into newer power systems, particularly AI-assisted converter design and high-density data-center power delivery.

Building on Electrification Research

Across my work on EV chargers and multiport converters, I have studied how the grid, photovoltaic generation, batteries, ultracapacitors, and vehicles can exchange power through flexible converter architectures. I have investigated bidirectional charging, vehicle-to-grid and vehicle-to-home operation, grid-current control, and energy-management decisions that account for source availability and battery state of charge. Other projects have examined AC/LVDC microgrid interfaces, fuel-cell hybrid vehicles, and the use of motor windings in integrated chargers and electric drives.

These applications have given me a system-level view of power electronics. An efficient topology is only a starting point: port interactions, mode transitions, current quality, protection, and control determine whether it will be useful in practice. I bring together circuit analysis, mathematical models, simulation, and experimental evaluation to understand those interactions.

AI-Assisted Modeling and Control

I am particularly motivated to explore how AI can help engineers model and optimize converters without losing the physics that makes a prediction trustworthy. My published work provides a foundation in switching-state equations, operating boundaries, loss and stress models, and measured behavior. I want to build on it with physics-informed models that combine those relationships with simulation and experimental data.

Isolated bidirectional converters such as the dual-active bridge and multi-active-bridge are especially interesting to me. Their modulation choices affect power flow, RMS current, efficiency, and soft-switching range at the same time. I want to investigate whether a learned model can predict those outcomes quickly enough to support modulation decisions while respecting energy balance, current limits, and zero-voltage-switching constraints. I am also curious about using such models for condition monitoring, where the challenge is to distinguish a developing fault from normal changes in voltage, load, or temperature. These are research directions I intend to develop, rather than results I claim from my existing publications.

Power Delivery for AI Data Centers

High-density AI computing is creating another application for power converters. The Open Compute Project and industry contributors are developing 800 V DC approaches for future data centers, including side power racks that can work with existing AC infrastructure and longer-term direct conversion from medium-voltage AC. Higher-voltage distribution can deliver the same power at lower current, reducing the burden on conductors as rack power increases.

I see several compelling research problems along that power path: efficient front-end conversion, high-density isolated DC–DC stages, DC-bus regulation, energy-storage integration, and rapid fault isolation. AI workloads can also produce fast power changes, making converter dynamics and coordinated storage important alongside peak efficiency. Proposed architectures place short-duration storage near compute racks and larger battery systems at the facility level to respond at different time scales.

I am actively looking for ways to apply my experience in resonant conversion, magnetics, semiconductor stress, multiport power flow, and microgrid control to these systems. An 800 V-to-low-voltage converter, for example, raises questions about switching loss, transformer design, output current, thermal density, and efficiency over a wide load range. At the facility level, I am interested in how converters and storage can stabilize a DC bus during rapid load changes without allowing a fault to spread through the distribution system.

My aim is to keep moving from a promising topology to a complete, testable power system. EV charging and renewable-energy research established that approach for me; AI-assisted design and data-center power delivery are the next applications I am eager to investigate.

When I design a power converter for an EV charger, electrical performance is only one part of the job. The product must also protect people during normal operation and foreseeable faults, coexist with nearby equipment, and remain safe across its intended input voltage, charging range, and environment. In our lab, I work on these requirements alongside converter topology, control, magnetics, and PCB design so that safety is built into the hardware from the beginning rather than treated as a final test.

From a Standard to a Design Decision

I begin by identifying the product boundary: whether the equipment delivers AC or DC to the vehicle, whether it can export power, which market it is intended for, and which parts of the system are accessible. Those choices determine the applicable IEC, UL, and, where grid interconnection or power quality is relevant, IEEE requirements. For North America, I distinguish UL 2202 for DC charging equipment from UL 2594 for AC EV supply equipment. I then turn the requirements into a working design checklist for the power stage, control board, enclosure, cabling, protective earth, and charging interface. The important work is in the details: a requirement for insulation or fault protection has to become a trace spacing, a component rating, a sensing path, a shutdown response, and a way to verify the result.

Insulation and High Voltage Boundaries

Much of my recent work has involved mapping clearance and creepage requirements on power and control PCBs. I consider the working voltage, expected transients, overvoltage category, pollution degree, material properties, and altitude before setting spacing rules. I also distinguish between functional, basic, and reinforced insulation, and follow the insulation boundary through connectors, gate drivers, sensors, auxiliary supplies, and the transformer. A board can look well separated in a schematic while a pad, slot, mounting feature, or enclosure interface creates a different physical path. That is why I review both the electrical architecture and the actual PCB geometry. UL 840 provides an insulation coordination approach where the applicable end-product standard specifically invokes it.

For an isolated DC charger, the transformer is only one part of the barrier. I examine the complete path between the grid side, the high voltage DC bus, the vehicle side, and any user-accessible or low voltage circuits. This includes protective-earth continuity, isolation of measurements and controls, stored energy after shutdown, and the behavior of relays or contactors. Dielectric withstand (hipot), insulation resistance, touch-current, and ground-continuity checks are among the verification activities that must be planned against the applicable charger product standard (UL 2202 for DC equipment or UL 2594 for AC EVSE in the North American context); a successful spacing calculation alone cannot establish compliance.

Protection Under Real Operating and Fault Conditions

I treat overvoltage, overcurrent, short circuits, overheating, loss of control power, and abnormal connection states as design cases, not just entries on a compliance list. The protective response has to be coordinated across hardware and firmware: sensing must remain trustworthy, a fault must be detected in time, switching must move to a safe state, and hazardous stored energy must be managed. I pay particular attention to how a charger starts, changes operating mode, shuts down, and responds if a component or communication path fails. These transitions often reveal risks that steady-state efficiency measurements cannot show. UL 2231-1 and UL 2231-2 address personnel-protection systems for EV supply circuits. Where a surge protective device is used, UL 1449 addresses that device; the charger itself still has to satisfy its applicable product standard.

This changes the way I review a circuit. A voltage-sensing divider is not only a measurement input; its resistor ratings, insulation, failure modes, and interface to the controller affect whether an overvoltage fault can be recognized safely. A relay is not only a way to reconfigure a power path; contact ratings, open-contact withstand, control supply behavior, and the timing of its operation matter during faults. Looking at components this way helps me connect a requirement in a standard to a testable hardware decision.

EMI and EMC as Part of Converter Design

Fast switching makes EMI a practical design constraint. In the lab, I use a LISN, spectrum analyzer, and EMCview to establish conducted-emission baselines and investigate noise peaks while changing the power circuit, filtering, grounding, and layout. I look at common-mode and differential-mode current paths, switching-node capacitance, gate drive behavior, cable routing, and the tradeoff between Y-capacitor filtering and leakage or touch current. I also consider radiated emissions and immunity: a charger must avoid disturbing its surroundings and keep its control and protection functions reliable when exposed to electrical disturbances. Pre-compliance work lets us find weaknesses early and prepare the product for formal EMC evaluation. UL 1283 can be relevant to the safety of a selected EMI filter, but it is not a charger emissions or immunity standard; IEC 61851-21-2 addresses off-board charger EMC.

Building Evidence for a Product

My approach is to connect each requirement to evidence: calculations, component specifications, schematic and PCB reviews, test procedures, and measured results across relevant operating corners. For EV charging products, that means considering electrical safety, EMC, thermal behavior, enclosure protection, and the charging and grid interface as one system. Depending on the design, the review also extends to enclosure and coupler requirements associated with UL 50/50E and UL 2251, and the additional power-export requirements of UL 9741 for bidirectional equipment. I value standards work because it changes engineering decisions while they are still affordable to change. The goal in our current lab projects is to develop chargers that are technically strong and ready for the independent testing and certification required in their intended markets.

UL Standards by Safety Topic

The applicable edition and test plan depend on the equipment configuration, destination market, and the requirements invoked by the end-product standard. This map shows which UL document addresses each topic; component standards support, but do not replace, evaluation of the assembled charger.