Methodology to Predict Product Reliability Under Repeated Random Loading

The methodology proposed in this paper is illustrated for a particular failure mode that was observed in a recently developed product during accelerated life testing (ALT). The probabilistic simulation methodology applied in this paper is used to predict and improve the survivability of the product under repeated drop testing. The failure mode encountered is the cracking of magnesium housing under repeated drop testing. A finite element model (FEM) of the entire phone was developed. In this paper, logarithmic and exponential strength degradation models are assumed. Based on baseline drop testing in ALT and the observed probabilities of failure, the strength distribution for the housing was inferred using Monte-Carlo simulations and the stress-strength interference theorem. The inferred strength distribution and degradation models were then combined to develop a "dynamic" probability density function for strength. With Monte-Carlo simulation, probability of housing failure was estimated from the stress and strength distributions under repeated drop loading. Using this method, several designs were evaluated virtually to arrive at a final design version of the housing with significant improvement in drop reliability.