Postponed maintenance scheduling integrating state variation and environmental impact

Abstract

In asset management, defect signals are critical indicators of upcoming malfunctions, whose initialization and growth are often affected by uncertain environmental stresses. It is cost-effective to postpone the removals of defects, so as to take advantage of the remaining useful lifetime since the defect identification. However, the impacts of environmental damage as well as heath state variation on postponement period are rarely addressed, which may lead to non-negligible malfunction hazard and maintenance resource waste. This study addresses these challenges by designing a state-based postponement maintenance policy. The environmental damage to asset lifetime is characterized by (a) random increment of state transition rates and (b) probabilistic malfunction risk. Two levels of postponed maintenance windows are set according to inspection consequences. Level-1 window is pre-set along with a finite number of inspections, where a normally working asset over the inspection zone is left untouched until the window. Level-2 window is determined upon the defect identification, offering shorter postponed duration due to the vulnerability. The applicability is demonstrated by a numerical experiment on asset management of oil pipelines.