Power Electronics-Enabled Battery Management Systems for E-Mobility Applications: Energy Storage Systems
Translated title
Power Electronics-Enabled Battery Management Systems for E-Mobility Applications
Author
Larsen, Frederik Holt
Term
4. term
Education
Publication year
2025
Submitted on
2025-05-27
Pages
29
Abstract
Efterhånden som elbiler bliver mere udbredte, stiger behovet for litium-ion batteripakker med høj ydeevne og lang levetid. Dette projekt udvikler og afprøver et batteristyringssystem (BMS), der aktivt balancerer de enkelte celler ved hjælp af effektelektroniske kontakter (switches). Aktiv balancering gør det muligt at omfordele energi og styre varme, så cellerne arbejder mere ensartet, hvilket kan øge effektivitet, sikkerhed og levetid. Styringen bruger et kunstigt neuralt netværk (ANN) til at generere pulsbreddemodulation (PWM), som hurtigt tænder og slukker kontakterne for at regulere energiflowet. På den måde kan BMS’et løbende styre hver celles ladetilstand (State of Charge, SOC) og temperatur. Der er udviklet en detaljeret cellemodel i MATLAB/Simulink med SOC-estimering, estimering af tomgangsspænding (open-circuit voltage), en ækvivalent kredsløbsmodel og en termisk model. Disse dele er samlet i en modulær batteripakke med seks celler. Det neurale netværk er trænet til at minimere forskelle i SOC og temperatur mellem celler og beregner modulationssignaler til at styre bypass-kontakterne. I simulering konvergerede både SOC og temperatur inden for 2500 sekunder, hvilket viser effektiv balancering. En robusthedstest med sensorstøj op til ±10 % bekræftede stabil styring. Derudover blev en virtuel platform brugt til at validere bypass-metoden uden brug af fysiske batterier. Samlet peger resultaterne på, at kombinationen af effektelektroniske kontakter og et neuralt netværksbaseret BMS kan forbedre batteriers ydeevne, levetid og effektivitet i elbiler.
As electric vehicles become more common, there is a growing need for lithium-ion battery packs that deliver high performance and long service life. This project develops and evaluates a battery management system (BMS) that actively balances individual cells using power electronic switches. Active balancing allows the system to redistribute energy and manage heat so cells operate more evenly, which can improve efficiency, safety, and longevity. The control method uses an artificial neural network (ANN) to generate pulse-width modulation (PWM) signals that rapidly switch the devices on and off to control power flow. This enables the BMS to continuously manage each cell’s state of charge (SOC) and temperature. A detailed cell model was built in MATLAB/Simulink, including SOC estimation, open-circuit voltage estimation, an equivalent circuit model, and a thermal model. These components were assembled into a modular six-cell battery pack. The ANN controller was trained to minimize differences in SOC and temperature across cells and to compute modulation signals for the bypass switches. In simulation, both SOC and temperature converged within 2500 seconds, demonstrating effective balancing. A robustness test with sensor noise up to ±10% confirmed stable control. A virtual platform was also used to validate the bypass method without using real batteries. Overall, the results indicate that combining power electronic switches with a neural-network-based BMS can improve battery performance, longevity, and efficiency in electric vehicle applications.
[This summary has been rewritten with the help of AI based on the project's original abstract]
Documents