Calculation of MTBF using MIL-HDBK-217 and Bellcore TR-332
MTBF

Mean Time Between Failures, MTBF, for an existing product can be found by studying field failure data. For a new product however, or if significant changes are made to the design, it may be required to estimate MTBF before any field data is available. In some cases MTBF for previous products can be used, if changes to a design are unlikely to affect reliability. In the case of a new or significantly changed design, an accepted technique is to estimate MTBF using a recognized model. The most popular models are MIL-HDBK-217 or Bellcore Reliability Prediction Procedure TR-332 version 6 for electronic systems. For mechanical MTBF prediction NSWC-94/L07 from the David Taylor Research Center is the most widely used model.
 

Reliability Engineering

What is reliability engineering? Reliability engineering is the function of analyzing the expected or actual reliability of a product, process or service, and identifying actions to reduce failures or mitigate their effect. Engineers analyzing reliability typically carry out reliability predictions, FMEA or FMECA, design testing programs, monitor and analyze field failures, and suggest design or manufacturing changes. Reliability engineering can be done by reliability engineers, design engineers, quality engineers, or system engineers. The overall goal of reliability engineering is to make your product more reliable in order to reduce repairs, lower costs, and to maintain your company's reputation. To best meet this goal, reliability engineering should be done at all levels of design and production, with all engineers involved.
 

Reliability Engineering and MTBF.

Reliability predictions form the groundwork for reliability analyses. They are used to compute the predicted failure rate or Mean Time Between Failures (MTBF) of your system. MTBF is usually expressed in terms of hours. For example, if your system has a predicted MTBF of 1000 hours, this means that, on average, your system experiences one failure in 1000 hours of operation. By using accepted standards for modeling failure rates of components, you can analyze your system and calculate your system's predicted failure rate or MTBF. The goal is to make sure that your system's predicted MTBF is within acceptable limits. If your prediction analysis shows low MTBFs, this means that you can expect your system to fail more often, and you may need to take steps to improve it. By making changes to your design, for example maybe lowering temperatures or stress levels, your predicted MTBF may improve and you can expect better product reliability. Reliability predictions can be done at any level of the design phase. In the early design stages, the reliability prediction may be more of a rough estimate, but then may be refined as the design becomes more stable. Even once the product is in the field, reliability predictions can take into account actual field data for more accurate predictions. Some of the accepted standards for performing reliability predictions are MIL-HDBK-217 and Bellcore TR-332.

More about reliability software