An Intelligent Age Replacement Maintenance Framework for Nonrepairable Systems

Replacement maintenance models with constant-interval preventive replacement times and associated replacement maintenance costs were developed for a radio transmitter system with a sudden but non-constant failure rate. Derived analytical solution enhanced economic outcomes, particularly in optimal replacement time at minimum cost. System parameters included reliability, hazard rate, and availability, useful for adapting the system’s operational conditions. To integrate intelligence, a hybrid framework combining both analytical and intelligent techniques was proposed. In demonstrating its feasibility, feature engineering was employed to derive target variables for constructing a reliable knowledge base. Up to 10,000 unique data points were simulated from the Birnbaum-Saunders or fatigue life distribution, with boundary conditions defined by the operational and service conditions of the original dataset. An exploratory analysis of the simulated knowledge base confirmed a perfectly aligned distribution, sufficient for efficient transmitter systems classification. Our intelligent solution therefore proves the efficacy of classification models over analytical models.