What are MTBF and MTTR, and how are they calculated?

Definitions and data

MTBF and MTTR are averages computed from failure and repair records. First, define the system boundary: a single pump, or a pump group? Then collect two things: the operating time, with planned stops excluded, and the repair time for each failure. The definition of repair time matters. Some organizations count only hands-on work; others include waiting for spare parts or a crew. Write the choice down, because two MTTR values computed with different definitions cannot be compared.

Worked example: one line over a period (hypothetical)

Suppose a packaging line runs for 8 000 operating hours and has 4 failures. MTBF = 8 000 / 4 = 2 000 hours. Suppose the four repairs take 24 hours in total. MTTR = 24 / 4 = 6 hours. Availability A = 2 000 / (2 000 + 6) = 2 000 / 2 006 = about 0.997, or 99.7 percent. This describes that period only. It says nothing about the next one, and it excludes setup, material shortages and quality stops, which need their own accounting.

Common mistakes

The most common mistake is reading MTBF as a part's life. MTBF is an average time between failures of a repairable system, not the expected life of one component. If the failure rate is assumed constant, the chance of surviving one MTBF is e to the minus one, about 36.8 percent, which shows MTBF is an average rather than a lifetime. The second mistake is computing MTTR without waiting time, which understates real downtime. The third is trusting a value built on very few failures. Four events give a wide uncertainty range, and one unusual repair can move the average a lot.

Follow-up questions