All sub-models are combined into MATLAB for system-level optimization, prioritizing power density, efficiency, and cost. Two PFC and LLC systems are proposed and prototypes are designed to validate the accuracy of the optimization tool. This integrated approach ensures comprehensive optimization, resulting in high-performance converter designs that meet the project's objectives.
System Level Optimization
FET & Heatsink Selection Algorithm
EMI Filter Optimization & Design
PFC Inductor Optimization & Design
An optimal design using off-the-shelf wound inductors is obtained through a proposed algorithm that considers the total volume, total cost, and total efficiency of candidate designs. Different weighting coefficients are assigned to these factors, and the design with the lowest total score is considered optimal. After optimization, the 4-level TP PFC converter is chosen as the most suitable for the 3.7 kW output power condition with off-the-shelf inductors.
The resonant frequency of an LLC converter influences the resonant tank design, which in turn affects the converter's working conditions and soft switching state. To achieve zero-voltage switching (ZVS), resonance elements must be selected systematically rather than relying on a designer's experience. Lithium-ion battery chargers using LLC converters have different design requirements due to the battery's nonlinear load profile, making the design process more complex. The charging process typically involves constant-current (CC) and constant-voltage (CV) stages, with different design requirements for each stage.
To optimize the design, four operating points of the battery are considered, allowing for a more comprehensive evaluation of all possibilities. Sweeping the resonance frequency, quality factor (Q), and inductance ratio (Ln) parameters, the optimal values for each parameter combination are determined. The researcher also designed the transformer and inductor based on these parameters, using a multicore approach to minimize size, cost, and power losses. The final design optimization took into account the volume, power loss, and cost of the system, ultimately selecting the most suitable resonance frequency using a penalty function.
