How often it fails

MTBF and MTTF: mean time between failures and mean time to failure, side by side

MTBF and MTTF for plant maintenance: both formulas, a free calculator, a worked example, and how MTBF and MTTR together set availability.

Plant Floor · Practical maintenance management

MTBF, mean time between failures, is the average time a repairable asset runs between breakdowns. MTTF, mean time to failure, is the average time a non-repairable item lasts before it fails and gets thrown away. Both are operating time divided by the number of failures. The difference is what happens after the failure: you fix a pump, you replace a seal.

Both formulas, and a calculator you can use with your own numbers.
MTBF, MTTF and MTTR side by side.
A worked example for one pump and one batch of seals.
What counts as a failure, and the mistakes that make MTBF lie.

MTBF, MTTF and MTTR side by side

	MTBF	MTTF	MTTR
Question it answers	How long does it run between breakdowns?	How long does it last before it fails for good?	How long is it down each time?
Applies to	Repairable assets: pumps, motors, gearboxes, conveyors	Non-repairable items: seals, bearings, fuses, lamps	Repairable assets
Formula	Operating time ÷ number of failures	Total run time of the items ÷ number of items failed	Repair downtime ÷ number of repairs
You want it	Higher	Higher	Lower
Owned by	Reliability, with the crew	Reliability and stores	Maintenance crew and planning

MTTR has its own page with the five clocks inside every repair: mean time to repair (MTTR). This page stays with the other two.

The MTBF formula

MTBF = total operating time ÷ number of failures

Operating time is the time the asset was running, or meant to be running, minus the time it was down for those failures. Count unplanned failures that needed a repair. Leave out planned stops: PMs, changeovers, no orders. They are not failures, and counting them makes a well-maintained asset look unreliable.

Some textbooks measure MTBF from one failure to the next, repair time included, so MTBF = MTTF + MTTR. In a plant the repair is usually small next to the run time, so the two versions land close together. Pick one, write it down, and use it every month.

The MTTF formula

MTTF = total run time of all items ÷ number of items that failed

Use MTTF for parts you replace instead of repair. Add up how long each one ran before it failed, then divide by how many failed. The answer tells stores and planning how often to expect a replacement, and whether a new supplier or a new install method made the part last longer.

MTBF and MTTF calculator

The fields start with the numbers from the worked example below. Change them to your own.

Repairable asset: MTBF, MTTR and availability

Period (hours)Hours the asset was meant to run. 720 is 30 days around the clock.Failure downtime (hours)Total hours down for unplanned repairs.Number of failuresUnplanned stops that needed a repair.

Operating time: 702 h
MTBF: 175.5 h
MTTR: 4.5 h
Availability: 97.5%

Non-repairable item: MTTF

Hours each item ran before it failedOne number per failed item, separated by commas. Seals, bearings, fuses, lamps.

Items: 6
Total run hours: 15,200 h
MTTF: 2,533 h

The numbers stay in your browser. Nothing is sent or saved.

A worked example: one pump, one batch of seals

Original fictional scenario. Created for this page to show the arithmetic. It is not a real plant or a real result.

MTBF for a transfer pump

A transfer pump is meant to run around the clock: 720 hours in a 30-day month. It fails 4 times. The four repairs keep it down for 18 hours in total.

Operating time = 720 minus 18 = 702 hours. MTBF = 702 ÷ 4 = 175.5 hours. MTTR = 18 ÷ 4 = 4.5 hours. Availability = 175.5 ÷ (175.5 + 4.5) = 97.5%.

One month is a small sample. Four failures can be four different problems or the same one four times. Before anyone celebrates or worries, the team opens the four work orders and checks the failure codes.

MTTF for mechanical seals

Six seals on the same pump model failed after 1,800, 2,100, 2,400, 2,600, 3,000 and 3,300 hours. Total run time = 15,200 hours. MTTF = 15,200 ÷ 6, about 2,533 hours.

The spread matters as much as the average. If the short-lived seals were all fitted by the same crew or came from the same batch, that is the next question to ask.

What counts as a failure

MTBF is only as good as the failure definition behind it. Write one down before you calculate anything. A workable one: the asset stopped doing its job, or had to be stopped, and needed a repair to get back. Then decide the edge cases once for the whole plant:

A trip that an operator resets in two minutes. Many plants log these as minor stops, not failures.
A failure found during a PM. It counts if the asset would have failed in service.
A failure caused by upstream equipment or bad material. Log it, and code it to the real source.

Consistent failure codes make this possible. The failure code cheat sheet gives the crew a short list to use at close-out.

Five mistakes that make MTBF lie

One average across unlike assets. A plant MTBF that mixes pumps, conveyors and robots describes no machine you own. Calculate it per asset or per asset class.
Counting planned stops as failures. PMs and changeovers drag MTBF down and punish the plants that do their PMs.
Calendar time instead of run time. An asset that runs one shift and one that runs three get the same MTBF on the calendar and very different ones on the hour meter.
Trusting the datasheet. A manufacturer's MTBF comes from test conditions. Your load, dust, heat and operators are different. Use your own history.
Reading MTBF as a life expectancy. An MTBF of 1,000 hours does not mean the asset runs 1,000 hours and then fails. If failures are random, only about 37% of assets make it to the MTBF without a failure.

How to raise MTBF

MTBF goes up when the same failure stops coming back. The order that works:

Rank the repeat failures. List the assets and failure modes that keep returning. The repeat failure watch list is built for this.
Find the root cause of the top one. Use the RCA template, and see a full root cause analysis example.
Prove it worked. Close the RCA only when the MTBF on that asset has moved. The RCA close-loop checklist asks for the trend, not a feeling.
Watch condition where failures are random. Fix the System's 30-day quick start points critical assets with an MTBF under 500 hours to the condition-based monitoring playbook. Start with the CBM pilot setup.

Questions planners ask

What is the difference between MTBF and MTTF?

MTBF is for things you repair and put back in service. MTTF is for things you replace when they fail. The arithmetic is the same: operating time divided by the number of failures.

Is a higher MTBF better?

Yes. A higher MTBF means longer runs between breakdowns. Watch the trend on each asset class rather than one number for the plant.

Does MTBF include repair time?

In the operating-time version on this page, no: repair time is taken out and reported as MTTR. In the failure-to-failure version, it is in. Both are in use. What matters is that your plant uses one definition every month.

What is a good MTBF?

There is no single good number. A conveyor and a compressor fail at different rates by design. Set a baseline per asset class from your own history, then aim to move it up. For the plant-level benchmarks that do travel between plants, see the world class maintenance metrics.

How do you convert MTBF to a failure rate?

The failure rate is 1 ÷ MTBF. An MTBF of 175.5 hours is about 0.0057 failures per operating hour. For an asset that runs all year around the clock, 8,760 ÷ 175.5 is about 50 failures a year.

How do MTBF and MTTR set availability?

Availability = MTBF ÷ (MTBF + MTTR). Raise MTBF and the asset fails less often. Lower MTTR and each failure costs less time. The maintenance KPI examples show where both sit in a KPI set.