Predictive Reliability Modeling in Distributed EV Charging Systems
Keywords:
electric vehicle charging systems, reliability modeling, failure prediction, distributed charging networks, degradation forecasting, service stability.Abstract
Distributed EV charging systems are becoming reliability-sensitive infrastructure in which charger availability, service continuity, and failure behavior vary across nodes under different load and stress conditions. Existing studies have addressed charging-behavior prediction, duration forecasting, flexibility estimation, charging-performance modeling, and infrastructure reliability assessment, showing that data-driven methods are valuable for interpreting heterogeneous charger operation. However, the literature remains fragmented across usage prediction, performance forecasting, and reliability analysis, leaving a gap in unified frameworks that connect distributed charger-state variables, degradation behavior, and predictive failure evolution. To address this issue, this article presents a reliability-modeling and failure-prediction framework based on rolling reliability state vectors, degradation-index forecasting, distributed node coupling, behavioral reliability classification, and failure-probability estimation. The results show distinct reliability trajectories under varying stress conditions, progressive failure-probability growth in vulnerable nodes, and declining service stability under degradation-sensitive behavior. Overall, the study demonstrates that predictive reliability modeling can support reliability-aware operation and maintenance planning in distributed EV charging networks.