Availability Modelling of Dynamic Load Sharing Competing Risk System Based on Continuous Monitoring Incorporating Copula and Hybrid Hazard Rate Model

ABSTRACT

Present study explores the reliability, availability, and maintenance costs of a load‐sharing competing risk system with three primary failure modes: inherent, induced, and damage‐induced. The system operates on a two‐stage failure process—fault occurrence followed by complete failure, where faults progress to failures after a random delay, enabling corrective maintenance. The load‐sharing among identical, interdependent components mitigates sudden breakdowns, with load distribution influenced by component health, environmental factors, and inter‐component interactions. A combined hazard rate model, incorporating inherent, induced, and damage‐induced hazards, along with periodic inspections, governs this mechanism. Additionally, a hybrid hazard rate model with copula accounts for dependencies among failure modes, providing an accurate assessment of component health and age. Post‐inspection hazard rates and reliability expressions facilitate the analysis of both instantaneous and steady‐state availability. Sensitivity analysis of inspection intervals, coupled with a comprehensive evaluation of maintenance costs and steady‐state availability, offers critical insights. A (M/M/1):(∞/FCFS) queuing model simulates real‐world conditions during corrective repairs, triggered by system failures or threshold breaches in the combined hazard rate. The methodology is validated with an example of a Data Centre Power Supply System, demonstrating its practical applicability.