Mean time between failures is one of the most quoted and most misread numbers in maintenance. The arithmetic is simple. The trouble is what people conclude from it, because MTBF is an average across a population and a period, and it tells you very little about when any particular machine will fail.
The formula, and an example
The mtbf formula with example: total operating time divided by the number of failures in that time. Ten identical pumps running a thousand hours each, with five failures between them, gives ten thousand hours divided by five, so an MTBF of two thousand hours. That mtbf calculation example is the whole of the arithmetic. Note what it required: identical assets, a defined period, and an agreed definition of what counts as a failure. Change any of those three and the number moves without anything physical changing.
What MTBF does not tell you
It does not say a machine will run two thousand hours. Failures cluster, wear-out is not random, and an average over a population hides the outlier that is about to fail. It also degrades badly with small samples: one failure on one machine gives a number that looks precise and means almost nothing. Treat MTBF as a fleet-level trend to watch over time rather than a prediction about any individual asset, and it becomes genuinely useful.
MTBF and MTTR together
Mtbf and mttr, sometimes written mttr and mtbf or in Spanish-language sources as mttr y mtbf, are the pair worth reading together: how often it fails, and how long it takes to get back. They pull on different levers. Improving MTBF is maintenance and design work; improving MTTR is spares, access, documentation and skills. A site with acceptable MTBF and terrible MTTR has a logistics problem, not a maintenance one, and reading only the first number hides that entirely.
Questions people ask about mtbf meaning
What is the MTBF meaning?
Mean time between failures: total operating time divided by the number of failures in that time. It is an average across a population and a period, not a prediction for one machine.
How do you calculate MTBF?
Ten pumps running a thousand hours each with five failures gives ten thousand divided by five, so two thousand hours. It needs identical assets, a defined period and an agreed definition of failure.
Why read MTBF and MTTR together?
They pull different levers. MTBF is maintenance and design; MTTR is spares, access, documentation and skills. Acceptable MTBF with terrible MTTR is a logistics problem, not a maintenance one.