Join the crewMy track

Classes / Power & Facilities class / Life Support Department

Life Support Track: HVAC, Cooling, Water and Power · Rank 6 coursework · Module 1 of 3 · about 40 min

Maintenance strategy and reliability

Maintenance approaches include reactive work after failure, preventive work by time or usage and predictive work informed by actual condition. Reliability-centered planning combines approaches for the operating context. Condition evidence supports decisions but cannot guarantee an exact future failure moment.

The right mix depends on consequence. A cooling pump that keeps a data hall alive gets redundancy, condition monitoring and spares on the shelf. A break room fan does not.

Metrics keep the program honest: planned versus unplanned work, backlog, mean time between failures and how often preventive tasks get done on time. Leaders read these weekly, not just after an outage.

In space, this thinking is mandatory. ISS life support relies on spares, redundancy and careful records because help is far away, and NASA’s life support work for the Moon and Mars puts reliability first.

Before using a condition trend, we check what was measured, when and under which operating conditions. A change in load or measurement basis can make two records unsuitable for direct comparison. We then connect the evidence to the asset's failure consequences, recovery needs and operating context. More monitoring is not automatically the best answer for every asset.

Worked example, fictional comparison: two pumps have the same recorded mean time between failures, but one supports essential cooling while the other serves a noncritical demonstration. Their consequences and recovery needs differ, so the matching average does not justify identical spare and backup arrangements. A metric also needs a defined observation period, operating exposure and failure definition before comparisons are meaningful.

Practice, on paper: we compare a critical pump, a low-consequence fan and two backups sharing one unassessed dependency. Our self-check separates observed trends from guaranteed forecasts, and equipment count from independent capability. We take no real vibration measurement or maintenance action.

A short observation period with no failures does not demonstrate an infinite reliable life or guarantee the next operating interval.

Sources

Module quiz

10 questions, 8 correct to pass. You'll see what you missed and why, and you can retake it.

Next: Leading a life support operation →