Available is not the same as reliable

Availability is the share of time equipment could do its job. Reliability is how likely it is to keep working without failing. A machine can be highly available and deeply unreliable at the same time, and the two measures pull maintenance in different directions.

How they come apart

Take two pumps. One fails twice a year and takes two days to fix. The other fails fifty times a year and is back in five minutes. Both land near the same availability. Nobody would call the second one reliable, and nobody running that plant would rather have it.

Why the difference matters for what you buy

High availability can be bought with spares, standby equipment and fast response. That is expensive to run forever and it does not make anything better. Reliability has to be engineered in: better installation, better operation, better preventive work, or different equipment.

When a site is quietly spending a fortune on rapid response, it is usually because availability is the only number being reported.

Frequency matters even when time does not

Every failure carries costs that never appear in downtime hours: the startup scrap, the operator’s attention, the risk of doing it badly under pressure, and the erosion of anybody’s belief that the plan is real. Fifty short failures cost more than the hours suggest.

Which number to report where

• •To operations and to finance: availability, and the cost of the unavailable time
• •Inside maintenance: failure count and MTBF, per asset
• •To a capital committee: both, with the trend, because that is the replace-or-repair argument

The pairing to watch

Availability flat and failure count rising is the early signal of an ageing asset being propped up by fast response. It is the point at which the replacement conversation should start, and it is invisible if only one of the two numbers is on the report.