Modulation Strategy Optimization for a Dual Active Bridge Converter in EV On-Board Charger Applications: An Investigation of ZVS Boundaries and Backflow Power in DAB Converters During the Battery Charge Cycle
Translated title
Modulation Strategy Optimization for a Dual Active Bridge Converter in EV On-Board Charger Applications
Author
Pappu, Prasanna Shastri
Term
4. semester
Education
Publication year
2024
Submitted on
2024-09-15
Abstract
This thesis examines how to best control a Dual Active Bridge (DAB) converter that serves as the isolated DC‑DC stage in a 3.3 kW on‑board charger for electric vehicles. Instead of comparing control methods at only one fixed operating point, as is common in existing studies, this work continuously tracks both the limits for Zero‑Voltage‑Switching (ZVS, soft switching) and undesired backflow (reactive) power over the entire charging profile of a lithium‑ion battery. The analysis is based on a 40 kWh battery pack whose voltage changes from 250 V to 420 V while connected to a regulated 400 V DC bus. The system model uses a large‑signal averaged representation of the DAB converter, combined with a simple first‑order lumped battery model and a nested voltage and current PI control scheme. This allows realistic simulation of closed‑loop CC/CV (constant current/constant voltage) charging behavior. The thesis first derives the classical equations for power transfer and ZVS boundaries for Single Phase Shift (SPS) modulation directly from fundamental principles. The same general ZVS‑requirement method is then extended to Extended Phase Shift (EPS) and Dual Phase Shift (DPS) modulation. The study investigates how the additional degrees of freedom – the inner phase shifts – can be used to maintain soft switching and reduce circulating reactive current when the voltage conversion ratio (the ratio of input to output voltage) moves away from unity. Simulation studies in MATLAB are used to validate the SPS baseline model and the proposed control loop. EPS and DPS strategies are then evaluated across the full charge cycle. The goal is to produce a practical modulation‑selection map that indicates which strategy the on‑board charger controller should use at each stage of charging to maximize efficiency and maintain ZVS over the entire operating range.
Denne afhandling undersøger, hvordan man bedst styrer en såkaldt Dual Active Bridge (DAB) konverter, som fungerer som den isolerede DC-DC del af en 3,3 kW ombordlader til elbiler. I stedet for kun at sammenligne styringsmetoder ved ét fast driftspunkt, som det ofte ses i litteraturen, følger arbejdet både grænserne for Zero-Voltage-Switching (ZVS, blød omskiftning) og den uønskede tilbageførte (reaktive) effekt gennem hele ladeforløbet for en litium-ion-batteripakke. Der antages en 40 kWh batteripakke, hvis spænding ændrer sig fra 250 V til 420 V, mens den er koblet til en reguleret 400 V DC-forsyning. Modellen bygger på en stort-signal middelværdi-beskrivelse af DAB-konverteren, kombineret med en simpel førsteordens batterimodel og et lagdelt PI-reguleringssystem for strøm og spænding. Dette gør det muligt at simulere et lukket CC/CV-ladeforløb (constant current/constant voltage) på en realistisk måde. Afhandlingen udleder først de klassiske ligninger for effektoverførsel og ZVS-grænser for Single Phase Shift (SPS) styring direkte fra grundligningerne. Derefter udvides den samme ZVS-metode til Extended Phase Shift (EPS) og Dual Phase Shift (DPS) styring. Her undersøges, hvordan de ekstra frihedsgrader – de indre faseskift – kan bruges til at bevare blød omskiftning og begrænse cirkulerende reaktiv strøm, når spændingsomformingsforholdet (forholdet mellem indgangs- og udgangsspænding) afviger fra 1. Simuleringer i MATLAB bruges til at verificere SPS-modellen og det foreslåede reguleringssystem. EPS- og DPS-metoderne analyseres dernæst gennem hele ladeforløbet. Målet er at opnå et praktisk “kort” til valg af modulationsstrategi, som fortæller, hvilken styringsmetode laderens controller bør anvende i hver fase af opladningen for at øge virkningsgraden og bevare ZVS i hele driftsområdet.
[This abstract has been rewritten with the help of AI based on the project's original abstract]
