It is important in todays global economy for companies to be able to develop, test and manufacture products/processes in a short amount of time, efficiently, and with high reliability. As a result, more practical methodologies are needed for designing-in reliability at the design stage of a product or process. This project will develop experimental design theories and techniques for product/process reliability responses that will fill the gap that currently exists in design for reliability. Traditional reliability testing, such as life testing or accelerated life testing, considers only the effects of environmental and operational conditions on product life time; thus, they are mainly for the purpose of reliability demonstration or verification. Direct application of traditional DOE tools on reliability experiments is inappropriate in most cases because (1) typical responses from reliability experiments are not normally distributed, (2) censoring is common in life testing, (3) the statistical quantity of interest in life testing is different from other experiments, and (4) life acceleration and model extrapolation are unavoidable in most reliability experiments. Experimental designs have been used for accelerated life testing, but preliminary studies indicate that our proposed methodology will result in designs superior to those encountered in practice.
|Effective start/end date||10/1/09 → 9/30/13|
- National Science Foundation (NSF): $348,315.00