Influence of tested samples on acceleration factors variation

In this paper, investigations were made to visualize the influence of tested samples on acceleration factors variation. For this, failure times were simulated on test stress levels, based on a basic population. Afterwards the p-value is calculated. In the next step, a constant shape parameter model and a non-constant shape parameter model were used to obtain failure times at various failure probabilities at field stress level. In the last step, the variation regarding failure time between the two used models, as well as between the models and the basic population, has been determined at the field stress level. The results show that high p-values will lead to smaller variations regarding failure time at field stress level than low p-values. Large variation between the two used models can also be used as an indicator that the tested samples do not stem from the same population. With a high p-value, the variations between the two used models to the basic population are similar. If the p-value is low, the constant model also yields appropriate extrapolation to the field stress level. Furthermore, a tendency towards decreasing variation can be noticed with an increasing failure probability, as well as with an increasing number of tested samples within a p-value range.