Data-Driven RUL Estimation for DC-Link Capacitor in Ultra-Fast EV Charging Systems

The reliability of ultra-fast electric vehicle charging systems heavily depends on the performance of critical components, notably the DC-link capacitor. This paper introduces a data driven approach for estimating the remaining useful life of DC-link capacitors in such systems. A virtual model of the system, developed as a digital twin, integrates real-time operational data with machine learning algorithms trained on historical degradation patterns. Key health indicators, such as a 30% reduction in capacitance (from 20,000 μ𝐹) and a 100% increase in Equivalent Series Resistance (from 0.4 𝑚𝛺), define the failure thresholds. Simulation-based validation demonstrates the effectiveness of the proposed methodology in accurately predicting capacitor degradation trends. This approach enhances the reliability of high-power EV charging systems and highlights the role of digital twins in predictive maintenance.