Spares sizing
[Placeholder copy] How many spares should be on hand? Enough to cover the failures you expect while a replacement is on its way, at the service level you choose.
The method
SpareKeep assumes failures arrive at a constant rate. Over the lead time, the expected number of failures is:
λ = installed count × lead time (days) ÷ MTTF (days)
The recommended stock is the smallest number k for which the chance of no more than k failures during the lead time (the Poisson cumulative probability) reaches the target service level. The default target is 90%.
A worked example
| Pumps installed | 30 |
|---|---|
| Mean time to failure | 900 days |
| Lead time | 60 days |
| Target service level | 90% |
λ = 30 × 60 ÷ 900 = 2.0 expected failures over the lead time.
| k | Cumulative probability |
|---|---|
| 0 | 13.5% |
| 1 | 40.6% |
| 2 | 67.7% |
| 3 | 85.7% |
| 4 | 94.7% |
k = 3 falls short of 90%; k = 4 is the first to reach it. Recommended stock: 4.
What it does not do
- It is a recommendation next to the contractual minimum, not a replacement for it.
- If MTTF or lead time is missing, SpareKeep shows "insufficient data" rather than a guess.
- The result is only as good as the failure data behind it.