ALT Overview
ALT (accelerated life testing) is a technique for speeding up product testing so that you can find the reliability characteristics of a product more quickly. To apply ALT techniques, you test products in high-stress situations that typically shorten product life or hasten degradation of product performance. With such tests, “stress” refers to any accelerating variable.
By statistically fitting an acceleration model to the test data and extrapolating the fitted model, you can estimate the life or degradation of the product at the lower stress levels encountered in normal use. Thus, accelerated testing plays a key role in meeting the requirements of shorter product development cycles.
This technique is especially useful for determining or demonstrating the field reliability performance of highly reliable products. When tested under stress levels encountered during normal use, products with high reliability last so long, or degrade so little, that their performance and lifetime are very difficult to estimate. Accelerated testing yields reliability estimates of such products in a much shorter time.
In a simple situation, the components under test are divided into multiple groups containing one or more components. Each component is tested at specific combinations of stress levels. You enter the failure and suspended times corresponding to the stress levels in the software.
If the stresses on the component are constant with respect to time, the tests are known as constant stress tests. In some cases, the stresses are increased with time so that the test can be completed within a predetermined time with some failures. You record variations in the stresses in stress profiles.
For further information, see the following references.
Meeker, William Q., and Escobar, Luis A., Statistical Methods for Reliability Data, John Wiley & Sons, Inc., New York, 1998
Meeker, William Q., A Comparison of Accelerated Life Test Plans for Weibull and Lognormal Distributions and Type I Censoring, Technometrics, Vol. 26, No. 2, pp. 157-171, 1984
J. E. Dennis, Jr. and Robert B. Schnabel, Numerical Methods for Unconstrained Optimization and Nonlinear Equations, Classics in Applied Mathematics 16
Numerical Recipes: C++ Version (2007)