The IEC 61124 standard (as well as the older MIL-HBDK-781-A guideline) describes test plans for demonstrating from field data that the failure rate of a product or system complies with given targets. Those plans, valid when the failure rate or failure intensity is constant in time, guarantee that the supplier risk (or Type I error) and the client risk (or Type II error) are less than pre-specified levels. Usually, those maximum acceptable risks are taken to be equal.
In practical industrial applications though, it is sometimes difficult to apply those test plans because the minimum test duration T exceeds the reasonably available test time: often it is necessary to complete the test while the products are still under warranty, usually within 1 to 2 years. Economic constraints also tend to reduce the test duration.
To overcome this difficulty, an alternative approach is proposed here: the cumulated test duration T is taken as an input, along with the maximum acceptable failure rate (or minimum acceptable MTTF) and design failure rate. The outputs are now, not only the maximum acceptable number of failures but also the supplier risk and client risk. The algorithm minimizes the absolute difference between those risks. The method is explained and real-life cases are discussed with numerical comparisons between the two methods, as well as with a test method based only on point estimates.