MTBF is a Starting Point, Only

MTBF is not meant to be used for anything other than teaching someone new to reliability how the various functions and tasks work.

Using MTBF in the real world is an oversimplification to the point of being less then useful. Possibly even harmful.

You see MTBF is books, articles, and papers, often with the caveat of the assumption to simplify the math to illustrate the process or concept. Hence, does not apply for actual use.

I’ve been told by professors that they use MTBF and the exponential distribution to get the math out of the way when they teach reliability concepts. Unfortunately, they run out of time to discuss how poor MTBF actually is in the real world. They don’t get to the useful stuff, thus leaving the students with only the basics, based on MTBF and the exponential distribution.

MTBF is Not Useful

I’ve said this many times, as have others. Here’s a run down of a few articles that highlight just how useless MTBF is in the real world:

Illuminating MTBF’s Lack of Information

Yet Another Confused MTBF Definition

Replace After MTTF Time To Avoid Failures – Right?

Another Way to Spot Someone Confusing MTBF

5 Clues Using MTBF is Not Helping

The Magic Math of Meeting MTBF Requirements

and, more. Just sear site for MTBF.

Keep Learning and Avoid Using MTBF

You should move quickly beyond using MTBF and the associated assumptions. Do so, limits your effectiveness and ability to add value to your organization and your career.

Maybe start using Weibull analysis to determine the rate of change of failures rates over time. As you know, when beta is not equal to 1, you really shouldn’t use those assumptions nor MTBF anyway.

Check out Mean Cumulative Function plotting for repairable systems. Another great way to determine how the failure rate is changing over time.

You should explore the range of distributions available. Learn about the various models, including physics of failure models, you have available.

Understand how to fully detail a reliability goal, it is not just a large MTBF value.

Read, ask questions, explore. Continue to learn and distance yourself from MTBF.

Then help others do the same.

Summary

MTBF is a useful tool to teach basic reliability concepts. I would prefer using Weibull or some other distribution as the math is really not that much more difficult. It is what it is and we should work to change our basic reliability eduction material.

The issue really is folks seem to stop learning once they get the intro course which include only MTBF and Exponential. That sucks.

Keep learning. Push toward using the right tools and techniques to best support your organization. MTBF is not it. MTBF is not reliability.