Life-Testing Results Based on a Few Heterogeneous Lognormal Observations

摘要

During some future period, a number of similar items are to be independently life-tested under 'typical' conditions. The observation for each item is its time to failure. The interest is in determining a length of time such that the probability of all the items surviving this long is at least a specified amount. The desired value is to be determined from the results of independent tests on a few items of this type, from supplementary information (tests on similar items), and from a combination of experience and judgment. All observations on items of the type to be life-tested are assumed to have lognormal distributions. The testing conditions are considered to be the same for each of the items to be life-tested later. However, each of the few items that are first considered may be subjected to different testing conditions. A consensus of these heterogeneous conditions furnishes the 'typical' conditions that are used for the items tested later. The statistical method consists in utilizing the few observations to allow for mean effects, and in evaluating an upper bound for the variation effects from supplementary information, experience, and judgment. This approach can have strong advantages when there are only two or three observations. (Author)