Improved Analysis Methods of Multilevel Accelerated Life Tests

摘要

The development of multilevel accelerated life stress tests, ALT, has been hindered by the difficulty of performing adequate and satisfactory data analysis especially when 3 or 4 levels of stress exist. Even seven levels of a single stress can become a challenge to analyze. This paper shows computerized software approach that has recently become available. The use of software makes the analysis easier, but sets limits on the models that the analyst must handle. Suspended interval data with multiple stresses can now be analyzed employing an improved Maximum Likelihood Estimator to achieve best fit solutions. The data analysis solutions permit a wider latitude in selecting test stresses and test conditions.