Correction of harmonics in electric vehicle li-ion battery charging systems using passive filter structures

In this study, the aim is to analyze and adaptively suppress harmonic distortions occurring during the charging process of Li-ion batteries used in electric vehicles. It is well known that fixed filtering methods are unable to adapt to the dynamic nature of Constant Current–Constant Voltage (CC–CV) charging modes. In this context, a battery and charger system rated at 400 V, 11 kW, and 50 kWh was modeled in MATLAB/Simulink, and harmonic behavior was examined based on the battery's State of Charge (SoC). The analyses revealed that harmonic distortions remained relatively stable within the 0–50% SoC range (CC mode), whereas a significant increase in higher-order harmonic components and total harmonic distortion (THD) values occurred with the transition to CV mode. It was specifically found that THD peaked in the 70–75% SoC range. The dominant components in the harmonic spectrum were identified as the 5th, 7th, 9th, 11th, and 13th harmonics, and passive LC filters were designed to suppress these. For instance, the filter for the 5th harmonic (250 Hz) was designed with an inductance of 10 mH and a capacitance of 40.5 µF. To select the appropriate filter, a K-Nearest Neighbors (KNN) algorithm was implemented using SoC and harmonic distortion factors (DFn) as inputs. Four filter classes were defined, and this structure was tested in real time on an ESP32-based microcontroller. The results demonstrated significant improvements in THD values after adaptive filtering. In terms of average power, the filtering provided an overall efficiency improvement of 1.23%, and 0.73% in the fast-charging region. The study shows that harmonic generation during battery charging is not static but dynamic, and that adaptive algorithms in harmonic management are critically important for both power quality and battery lifespan.