Reliability Analysis of Critical Components in Electric Vehicles Considering Environmental Covariates

Sammanfattning

Traditional reliability assessments of automotive components are usually conducted based on operating time or mileage and do not consider the variation of operating environmental factors. Consequently, identical components operating under different environmental conditions may have different reliability behavior which is not reflected in traditional reliability analysis. This study focuses on how operating environmental conditions affects the reliability of the high voltage contactor box installed within the AKASOL battery system used in Volvo electric bus fleets. Survival analysis methods were applied to 10 years of field replacement data having 730 unique chassis to identify and quantify the environmental covariates that are associated with the component replacement behavior of the buses. The effects of temperature and humidity on replacement events were evaluated and compared the reliability between different operating cities using a structured modelling framework that incorporates Kaplan–Meier estimation, Weibull analysis, LASSO based variable selectionand Cox proportional hazards modelling. The results identified 14-day maximum temperature and 1-month maximum relative humidity were the most strongly associated environmental covariates with replacement events. Humidity exhibited the strongest association with replacement events, while temperature contributed to measurable differences in predicted reliability between different operating locations. The distribution of observed replacement events was found to be consistent with humidity conditions referenced in relevant automotive qualification standards. However, it is important to note that these findings should be considered statistical associations and not direct evidence of the physical failure mechanisms. The study demonstrates that incorporating operating environmental covariates substantially improves field reliability assessment across electric bus fleets operating in different geographic regions. The resulting framework supports geographically differentiated fleet reliability management and provides valuable insights for risk-based maintenance decision making and warranty period optimization.