Reliability prediction is a quantitative analysis technique used to predict the failure rate of a system based on the components of the system and its operating conditions. A reliability prediction is typically performed using an accepted, published handbook, which defines failure rate equations for various components used in the system design. These equations were developed by using statistical techniques to analyze failure data gathered on actual operating equipment. The equations take into account various parameters, such as part quality and operating stresses, which have an effect on component reliability.

To begin a reliability prediction analysis, you must first define your system and all of its component parts. You then use the model equations to determine the failure rate of each particular component in your system. To get the overall system failure rate, you add up all the component failure rates.

For example, consider a system that is comprised of subsystem A and subsystem B. Assume subsystem A has 5 parts, each of which has a predicted failure rate of 1 failure/million hours obtained from using failure rate equations from a selected handbook. Assume subsystem B has 5 parts, each of which has a failure rate of 3 failures/million hours, computed using the same handbook. The failure rate of subsystem A is 5 failures/million hours and the failure rate of subsystem B is 15 failures/million hours. Thus, the failure rate of the overall system is 20 failures/million hours.

Doing this analysis by hand can be tedious, time-consuming, and error prone. Using a software application such as Windchill Prediction to perform this work can dramatically increase the efficiency and accuracy of analyses.

If you are unfamiliar with reliability prediction analysis and would like to study this topic in more detail, the following selections are excellent references: