Mean Time Between Outages

Reliability is described by how often a unit fails and how long it takes to restore - the drivers behind the availability number.

Behind the availability number

Availability is not a primitive quantity; it emerges from two others: how often a unit fails (mean time between failures) and how long it takes to restore (mean time to repair). The design-stage availability of 0.86–0.995 is the ratio these produce. Improving availability means failing less often, restoring faster, or both.

The two levers

Failing less often comes from robust components, condition monitoring, and avoiding common-cause faults. Restoring faster comes from modular design, on-site spares, and remote-handling access to the parts most likely to need service. A remote or austere site lengthens repair time, which is why such sites lean harder on redundancy.

• MTBF up: robust parts, monitoring, common-cause avoidance
• MTTR down: modularity, on-site spares, remote handling
• Availability = uptime / (uptime + downtime)
• Remote sites: longer repair, more redundancy needed

Why the range is wide

The estimate spans 0.86–0.995 because the burner is at the design stage: the plug operating regime is 166–830× beyond any device, so failure rates cannot yet be measured, only bounded. The wide range is honest uncertainty, and narrowing it is a core objective of the test-burner program.

These reliability drivers are physics-and-engineering quantities: raising how rarely a unit fails and lowering how long it takes to restore are the only two ways to move the availability figure, and both are explicit development targets for the burner program. Because both levers act on the same availability figure, a remote site with slow repair logistics can reach the same system availability as a well-connected one by trading additional redundancy for its longer repair times.