Dynamic scheduling of intelligent group maintenance under adaptive information updating

Intelligent group maintenance management is an effective approach to ensure the operational availability and profitability of large-scale industrial plants. Most current group maintenance models are static, mainly based on setting fixed groups or age thresholds, which cannot fully harness real-time health status. To address this issue, this paper proposes a dynamic intelligent group maintenance policy based on adaptive information prediction. A universal stochastic model is established to capture the non-steady deterioration trends, followed by lifetime prediction through Maximum likelihood estimation (MLE) and Bayesian inference. Then a two-stage maintenance optimization model is established, which combines inspection-based predictive and opportunistic replacement. The penalty functions and cost saving functions are separately calculated based on the real-time state status and historical inspection information. The optimal group sequence and operational time are obtained through sequential dynamic programming. Numerical experiment are provided to demonstrate the feasibility and advantages of the proposed model.