Maintenance issues

There are many theoretical models of component failure. The most probable is the 'bath tub' curve. There is evidence from the housing insurance experience of some 100,000 dwellings that the model has validity in practice (Mayer 2003).

This model has 3 distinct phases:

• Phase 1: early failure. Also known as 'burn-in' failures. The rate of failures decreases.
• Phase 2: random failure. The rate of failure is constant. Failure is random and therefore unpredictable.
• Phase 3: age-related failure. Components have reached the end of their effective life and fail due to effects of deterioration agents. The failure pattern at the end of life follows a normal distribution (bell-shaped curve) around a mean (or average) service life.

Figure 1: 'Bathtub' model of component failure and maintenance response

The question from a maintenance perspective is how to manage this pattern of component failure. The challenge is compounded for components with different service lives; the pattern of failure may be similar but the period of time in any one phase will vary.

A maintenance strategy to deal with the different phases may follow this sequence:

• Phase 1: early failure. Failures may be covered by manufacturer's guarantees. Otherwise corrective maintenance would be required.
• Phase 2: random failure. Generally failure during this phase of a component's life is managed by corrective maintenance. Regular inspection or some form of condition monitoring may pick up deterioration in performance prior to failure so that maintenance can be planned.
• Phase 3: age-related failure. Planned maintenance typically targets age-related failure or age-related deterioration of performance. The aim of planned maintenance is to intervene at some stage on the boundary between phase 2 and phase 3, before the rate of component failure increases to an unacceptable level.