Condition-Based Maintenance Decision Making: A practical approach for Marine Vessels

Abstract

Condition monitoring and predictive maintenance are gaining importance amongst researchers and practitioners ... The reason lies in the lack of high quality monitoring data, largely due to the reluctance of system owners to invest in monitoring equipment and data processing. This leads to a vicious circle that limits the applicability of condition based maintenance. The need for simple yet effective schemes for data-driven maintenance optimization motivated the present research. This paper demonstrates the development of such a scheme for maintenance optimization of marine vessels based on the readily available data, thereby lowering the entry barrier for ship operators and promoting future improvements of the data quality. Our approach combines data-driven and physical modelling. It then uses the models as an input for a constrained optimization algorithm based on linear programming and a Monte-Carlo-based uncertainty analysis. The algorithm allows for a gradual enhancement of the data-driven component upon acquisition of more and better data.