Condition-based maintenance planning for two-phase deterioration considering inspection errors with non-constant probabilities

Maintenance planning is a significant channel to balance the requirements of operational availability and profitability. Inspection, as a predominant maintenance action, is frequently confronted with inevitable errors due to management issues, tools and health state, whose probabilities can be deemed as time-related variables. Few of related literatures consider that imperfect effect, this paper investigates an innovative condition-based maintenance planning oriented to two-stage deterioration from a more practical view, where the inspection outcomes exist two types of variant-probability errors, false positive and false negative errors. Upon inspection, the asset will be replaced once (1) the defective state is identified, and (2) the inspection outcome has been normal for a pre-set times and then a delayed replacement will be scheduled. Both the inspection-based replacements mitigate the failure risk caused by inspection errors and utilize the remaining useful value more sufficient. The model applicability and superiority are verified via numerical experiments.