Optimizing burn-in and predictive maintenance for enhanced reliability in manufacturing systems: A two-unit series system approach

Abstract

In daily life, the reliability of manufacturing systems is critical, influencing everything from consumer goods availability to global supply chain stability. As manufacturing reliability increasingly dictates market leadership, ensuring system dependability and efficiency has become crucial. Despite extensive research, the integration of burn-in processes and Predictive Maintenance (PdM) within operational frameworks remains inadequately explored, especially in the dynamics of a Two-Unit Series Manufacturing System (TUMS). This research addresses this gap by developing an advanced Semi-Markov Decision Process (SMDP) model that synergistically optimizes burn-in and PdM strategies. This model minimizes downtime and operational costs while maximizing system reliability. Employing a combination of theoretical modeling and empirical validation, the study introduces novel algorithms that optimize maintenance schedules and predict system degradation effectively. The robustness of our approach is validated through comprehensive comparison analysis, which highlights the superior performance of our predictive maintenance model over traditional methods. A practical case study involving an EV battery system demonstrates the real-world applicability and significant improvements in battery reliability and operational lifespan. Sensitivity analysis and Monte Carlo simulations further substantiate the model’s effectiveness, showing resilience to parameter variations and consistent performance benefits under varied scenarios. In the specific context of our EV battery case study, implementing the proposed strategies initially resulted in a notable increase in EV battery lifespan and a significant reduction in maintenance costs. Ultimately, this study not only advances the theoretical framework of maintenance optimization but also equips industrial applications with a robust, scalable model, marking a significant step forward in the PdM of complex manufacturing systems.