Transient analysis of maintenance intervention of continuous multi-unit systems

The requirement for self-diagnosing and fault-tolerant systems and the need to conserve dwindling natural resources have led to an increased emphasis on design for reusability, survivability and maintainability. Existing research on failure and maintenance processes focus mostly on steady-state operation. A methodology for intervention analysis of a continuous multi-unit system under transient response is now presented. For modeling convenience, a continuous process without buffers is considered. A delay function is used in the determination of the optimum preventive maintenance policy for the individual units. The cost saving which arises when an opportunistic maintenance policy is adopted is investigated. The approach has three major thrusts. First, the development of analytical models based on the renewal theoretic framework that incorporate maintenance costs in formulating the Maintenance Decision Problem (MDP). Second, numerical methods for solving the results from the proposed formulation using differential and integral equations. Third, the embedded Binary Integer Programming (BIP) formulation that is used to optimize the resulting solutions because of the strong dynamic and unimodularity feature of the MDP problem. Data from an Aluminum Hot Roll Line consisting of three component parts is used to demonstrate proof of concept and to verify the problem formulation. Sensitivity analyses are performed on the resulting BIP model to provide optimum arrangement for the opportunistic maintenance problem.