A Decision Analysis Model for Maintenance Policy Selection in Multi-unit Systems  
Author Mahmood Shafiee




Abstract Selection of an appropriate maintenance policy for multi-unit systems is a very complex and critical task. An improper selection of maintenance policy for each unit may adversely affect the system availability and thereby reducing overall productivity as well as profitability. From the latter half of the 20th century, the multiple-criteria decision making (MCDM) approach has gained momentum in this field. In the current paper, we propose a decision analysis model based on analytic network process (ANP) to select the most suitable maintenance policy for different sets of units associated with the system. In the proposed model, the decision-maker (i.e. system owner or service agent) must choose the best maintenance policy among a set of possible alternatives according to two main criteria: maintenance implementation costs and failure criticality. The former criterion includes the maintenance expenditure required for hardware, software, and personnel training, while the latter focuses on the capability of maintenance in reducing the failure vulnerability and enhancing the reliability and resilience. The possible "correlations" among selection criteria as well as between component failures are also taken into consideration. The model is applied to determine a suitable maintenance policy for a new configuration of wind turbines consisting of several mechanical, electrical and auxiliary subassemblies.


Keywords Maintenance policy selection; multi-unit system; multiple-criteria decision making (MCDM); risk assessment; Analytic network process (ANP); system interdependence
    Article #:  22388
Proceedings of the 22nd ISSAT International Conference on Reliability and Quality in Design
August 4-6, 2016 - Los Angeles, California, U.S.A.