Determination and interpretation of activation energy using accelerated-test data

The activation energy is an indication of effect that the covariate (temperature, humidity, etc.) has on the life of the product. An incorrect assumption on the value of the activation energy can lead to an incorrect deduction regarding the effect that the covariate has on the product, as well as an erroneous calculation of the acceleration factor. Estimating the activation energy based on the accelerated-test data removes all of the guesswork when it comes time to decide on a value. The idea of estimating the activation energy based on data is not exactly a new idea, but now this process can be conducted more easily. In the past, the process was long and tedious and involved a lot of iterative calculations. But now, the ability to estimate the activation energy based on collected accelerated-test data has been implemented within a software package, ALTA^TM, which is expressly designed for accelerated life testing analysis. The activation energy can be determined by simply making it one of the unknown parameters that need to be estimated. In doing so, the problems which arise from selecting an activation energy have been removed, and the data alone dictates the value. To illustrate this, accelerated life test data was collected on thermally stressed hard drive systems and analyzed using the Arrhenius-Weibull model. The parameters of the Arrhenius model are then estimated using maximum likelihood theory. Once the parameters of the life-stress relationship are obtained, the activation energy can then be easily estimated, along with the acceleration factor