The importance of maintenance is industry is unquestionable. Accordingly, maintenance should be scheduled and performed such that the benefits outweigh the costs. Maintenance and reliability theory is a necessary part of such an analysis.
Industry failure data is typically “incomplete” in the sense that times to failure cannot always be correlated with the underlying physics of failure. In many cases it is not simply a case of an organisation not collecting adequate data, but rather that access to that data is restricted by service contracts.
Maintenance performed by a service agent can provide a means for assuring skilled repairpersons. Such practice is becoming commonplace in industry, as businesses become more competitive and service-oriented. Knowledge of failure mechanisms should be specified in the maintenance service contract to allow for “white-box” failure modeling.
Optimal component maintenance is determined after ‘optimal’ is defined. Numerous objective functions can be used to determine optimal component maintenance. Two of the most pertinent are total cost and availability. The steps in determining optimal component maintenance are:
1. Develop a suitable failure model from reliability theory
2. Apply the model to maintenance theory
3. Determine operational parameters to use in the maintenance model
Operational parameters used in the maintenance model are cost of preventative and corrective replacement, time for corrective replacement and time for preventative replacement. Accurate estimation of these parameters is crucial, since they have a profound effect on the optimal replacement interval.
This report discusses the theory needed to determine the optimal age-based replacement policy for a component. The “black-box” approach to failure modeling is used, where there is no knowledge of the failure mechanism.
As an illustration of reliability and maintenance theory, a case-study of optimal agebased replacement for hydraulic pumps on Liebherr 996 excavators is undertaken. The case-study brings to attention salient points for consideration when writing a service contract with a service agent. Optimal replacement for the pumps is discussed with attention drawn to the stochastic nature of failure modeling and the effect this has on changing maintenance actions.
Ultimately, the best failure models result from a complete knowledge of the mechanisms responsible for component failure. This is reliant upon negotiations with the service agent and is the best step in a transition from the stochastic to the deterministic.