PREDICTIVE MODELING OF THE SERVICE LIFE OF ELECTRONIC COMPONENTS IN PREMIUM AUTOMOTIVE SYSTEMS

The relevance of this study is driven by the rapid increase in the architectural complexity of electronic systems in premium automotive platforms, the growing integration density of ECUs and ADAS, and the heightened requirements for functional safety and resource predictability.Under multifactor thermal, electrical, and mechanical loading conditions, conventional reliability indicators such as MTBF do not provide sufficient accuracy for individualized prediction of the remaining useful life, which necessitates the development of adaptive models for RUL assessment.The purpose of this article is the development and experimental validation of a hybrid model for predicting the service life of electronic components in premium automotive systems, aimed at improving the accuracy of RUL estimation under realworld multifactor operating conditions.The study applies a systems approach, elements of reliability theory, parametric modeling of degradation processes, and ML algorithms for the analysis of telemetry time series.Simulation modeling of operational loads was conducted, and the model was experimentally verified using the MAE and RMSE metrics.The results identify patterns of electronic component degradation under variable loading conditions and establish the nonlinear and cumulative nature of resource depletion.A hybrid model was developed that integrates a parametric description of aging kinetics with adaptive correction based on ML.Experimental findings demonstrate a reduction in RUL prediction error compared with standalone parametric and ML approaches, confirming improved accuracy of resource assessment under variable operating profiles.