Hands-on projects from graduate coursework

1

Project 1 of 5. ECEN 828 Power Electronics, Prof. Jun Wang. Term project, team of two.

A dual-output flyback converter, from the magnetics up

Design, PLECS simulation, transformer fabrication, PCB assembly, and hardware qualification. 2021.

The problem

A flyback converter is the smallest way to build an isolated power supply: one switch, one coupled inductor, one diode per output. The specification was an 18 to 30 V input and a regulated ±15 V dual output at 0.3 A per channel, 9 W in total, switching at 100 kHz in discontinuous conduction across the whole input range, with better than 80% efficiency and no forced cooling. The controller, a peak-current-mode LM2587, and every semiconductor on the board were fixed by the project. The coupled inductor was not. It had to be designed, wound, characterized, and made to work.

What I built

The design started with two operating points. For continuous and for discontinuous conduction I derived the turns ratio, magnetizing inductance, and duty cycle that meet the specification at rated load, bounded by the 65 V switch rating, the 40 V output diodes, and the 30 V clamp on the primary, then checked both in a Level-1 PLECS model with an ideal coupled inductor. The discontinuous design carried forward: magnetizing inductance set at the minimum input voltage so the converter never crosses into continuous conduction, peak flux density held below half of saturation, and the maximum duty cycle capped at 0.6 so that peak-current-mode control stays stable. A Level-2 PLECS model then added the core, the air gap, the leakage inductance, and the winding resistance that the real part would have.

Then the part itself. A Ferroxcube E25/13/7 core in 3C94 ferrite, Litz-wire windings, and a Kapton-tape air gap, wound by hand and measured on an LCR meter for magnetizing inductance, leakage inductance, and dc and ac winding resistance. Iterations went back to the bobbin until the measured values sat within tolerance of the design. Only then were the transformer and the rest of the bill of materials soldered onto the PCB.

On the bench

Qualification meant running the converter into 50 Ω per channel and holding both outputs within ±4% of 15 V for three minutes each at 24 V, 18 V, and 30 V input, while input and output power were logged to compute the efficiency. The prototype passed at all three input voltages.

Converter specification the prototype was qualified against

Property	Requirement
Input voltage	18 to 30 V
Output	±15 V, within ±4%
Rated output	0.3 A per channel, 9 W total
Conduction mode at rated load	Discontinuous over the full input range
Switching frequency	100 kHz
Efficiency, natural convection	Above 80%

Mathematical background

• Steady-state converter analysis by inductor volt-second and capacitor charge balance; the boundary between continuous and discontinuous conduction
• Flyback transformer design: turns ratio from device voltage ratings, magnetizing inductance from the conduction-mode requirement, air gap from peak flux density
• Core loss and winding loss estimation, Litz wire and skin depth at 100 kHz
• Peak-current-mode control and its subharmonic instability above 50% duty cycle

2

Project 2 of 5. ECEN 898 Modeling and Control of Three-Phase Power Converters, Prof. Jun Wang. Term project, team of three.

A permanent-magnet machine drive, emulated with two inverters and controlled on a DSP

dq-frame current control, phase-locked loop design, PLECS switching and processor-in-the-loop simulation, deployment on TI C2000 hardware. 2022.

The problem

Testing a motor drive on a real motor means a shaft, a load, and many ways to break something. This platform replaces the machine with a second inverter. VSI-2 synthesizes the back-EMF of a permanent-magnet machine whose electrical frequency sweeps between 30 and 60 Hz every few seconds, with voltage proportional to frequency as in a real machine. VSI-1 is the drive under test. Its job is to hold the dq-frame phase current at a constant 1 A on the q axis and zero on the d axis, within ±0.05 A of switching-cycle-averaged variation, while the frequency, the back-EMF amplitude, and the unregulated dc bus on the far side all move. The only measurements available were the three phase currents, the line-to-line back-EMF voltages, and the bus voltage.

What I did

My share of the design was the steady-state operating point of the drive, the phase-locked loop, and the full switching simulation. From the dq averaged model of the inverter and the RL link I computed the phase-to-neutral and line-to-line duty cycles that produce the commanded current at each back-EMF frequency. The phase-locked loop locks to the filtered line-to-line back-EMF so that the current controller can work in a rotating frame aligned with the machine; it had to be fast enough to track the frequency sweep and keep at least 45 degrees of phase margin, with the RC sensing filters, the half-period PWM delay, and the one-period computation delay all modeled in the loop. The switching model of the complete system, both inverters, both modulators (third-harmonic-injection PWM for the machine emulator and minimum-loss center-aligned space-vector modulation for the drive), the phase-locked loop, the current controllers, and the time-varying back-EMF, was my simulation to build and run before anything touched hardware.

The last steps were shared across the team. Controller code was generated from the PLECS model with PLECS Coder and checked processor-in-the-loop against the simulation, the models were revised with the measured sensor gains and filter values, and the code was loaded onto two TI LAUNCHXL-F28069M boards driving a pair of BOOSTXL-DRV8301 inverter stages, with dead time and overcurrent protection configured in the firmware.

Operating point and control requirements

Quantity	Value
dc bus	20 V
Interface inductance and resistance per phase	3.44 mH, 3.4 Ω
Emulated back-EMF	30 to 60 Hz, 3 to 6 V on the q axis
Current command, dq frame	0 A, 1 A
Allowed current variation	±0.05 A, switching-cycle averaged
Switching and sampling frequency	12 kHz
Phase margin, PLL and current loops	At least 45°

Mathematical background

• Power-invariant Clarke and Park transformations, phase and line-to-line representations of three-phase variables
• Switching and averaged models of the voltage source inverter in stationary and rotating frames
• Small-signal loop design with sampling, PWM, and computation delays; discrete-time implementation of the compensators
• Phase-locked loop as a feedback loop: bandwidth against phase margin, and the effect of the sensing filters
• Carrier-based space-vector modulation and third-harmonic injection

3

Project 3 of 5. MECH 892 Special Topics in Robotics, Prof. Carl Nelson. Coding assignments and term report, individual.

Parallel robot kinematics and the dynamics of a four-axis arm

Workspace and singularity analysis in code, configuration-space search, and a report on manipulator dynamics and control. 2023.

The course was a tour of robot architectures that are not serial arms: parallel and cable-driven mechanisms, compliant and soft actuators, origami-inspired structures, and the screw-theory language that describes all of them. The assignments were code. I derived the inverse kinematics of a 3-RPR planar parallel robot and swept the plane to map its reachable workspace, then found its singularity loci graphically and confirmed them by deriving the Jacobian and the conditions under which it loses rank. I sketched the configuration space of a planar RP arm around a workspace obstacle and wrote a search over it, and I built a heat map of a kinematic performance index across the workspace of a two-link arm. Earlier in the term I modeled McKibben pneumatic muscles from their ideal contraction and force-pressure relations to size an actuator against a spring load.

The term report took the modeling side further: Euler-Lagrange dynamics of a four-degree-of-freedom arm, and a comparison of the control strategies the literature has applied to it, from PID through differential-evolution-tuned and fuzzy-logic PID.

Mathematical background

• Forward and inverse kinematics of parallel mechanisms; workspace mapping through the inverse solution
• Jacobians, singularity classification, and manipulability and condition-number indices
• Configuration-space topology and obstacle representation
• Screw theory: twists, wrenches, pitch, and reciprocity
• Euler-Lagrange dynamics of open kinematic chains

4

Project 4 of 5. ECEN 935 Computational Intelligence, Prof. Wei Qiao, individual project; and CSCE 878 Introduction to Machine Learning, Prof. Stephen Scott, team assignments.

Learning systems, from backpropagation to swarm optimization

Particle swarm optimization of a wireless power link, and six machine learning notebooks trained on a GPU cluster. 2021 and 2023.

Two courses, two years apart, covered the learning methods that later fed into my work on physics-informed observers. In Computational Intelligence the emphasis was on the algorithms themselves: multilayer networks trained by backpropagation, radial basis function and echo state networks, particle swarm optimization, evolutionary computation, fuzzy inference and fuzzy control, support vector machines, and reinforcement learning. My project applied particle swarm optimization to a hybrid inductive and capacitive wireless power transfer link, searching the design parameters to maximize link efficiency.

Introduction to Machine Learning was practice. Working in a small team on the university's Swan cluster, submitting Jupyter sessions as SLURM jobs and requesting GPU nodes when a model needed one, we completed six notebooks: exploratory analysis and decision trees on the Iris and heart-disease data sets, naive Bayes sentiment classification of IMDB reviews, a convolutional network for CIFAR-10 image classification, k-means and hierarchical clustering for image segmentation, hidden Markov models for sequence data, and random forests for heart-disease prediction. The course project was a proposal built around a Kaggle competition: problem definition, literature review, and a proposed learning solution.

Mathematical background

• Gradient descent and backpropagation; radial basis and recurrent reservoir networks
• Bayes' rule and generative classifiers; maximum-margin classification and kernels
• Bias and variance, bagging, boosting, and random forests
• Expectation-maximization, and the forward-backward and Viterbi algorithms for hidden Markov models
• Population-based stochastic optimization: particle swarms and evolutionary algorithms

5

Project 5 of 5. ECEN 932 Advanced Power Electronics and Applications, Prof. Wei Qiao. Individual project.

A turn-off snubber for the switch in a photovoltaic battery charger

Snubber sizing and PLECS comparison of switching trajectories with and without the circuit. 2022.

A buck converter stepping a 240 V, 2 A photovoltaic string down to 48 V at 10 A for battery charging puts its MOSFET through a hard turn-off every cycle, with the voltage rising across the device while it is still carrying current. The project designed a polarized RC turn-off snubber to shape that trajectory: the capacitor takes over the current as the switch opens and limits the rate of voltage rise, the resistor discharges the capacitor when the switch closes again, and the diode makes the two paths asymmetric so the resistor stays out of the way during turn-off. I sized the snubber from the switching transition, built the converter in PLECS with and without it, and compared the switching waveforms and device stresses between the two.

Mathematical background

• Switching loci of a hard-switched transistor and the safe operating area
• RC snubber sizing from the turn-off current and the allowable rate of voltage rise
• Snubber loss and its trade against switching loss in the device