Application of design of experiments to accelerated life testing

摘要

Today, there is an increasing demand for improved quality and reliability due to increasing system complexity and increasing demands from customer. Continuous improvement of quality is not only a means of competition but also a matter of staying in the market. Accelerated life testing and statistical design of experiments are two needed methods for improvement of quality. The combined use of them is very advantageous and increases the test efficiency. Accelerated life testing is a quick way to provide information on the life distribution of materials and products. By subjecting the test unit to conditions more severe than those at normal usage, the test time can be highly reduced. Estimates of life at normal stress levels are obtained by extrapolating the available information through a reasonable acceleration model. Accelerated life testing has mostly been used to measure reliability but it is high time to use it for improvement of quality. Design of experiments serves to find out the effect of design parameters and other interesting factors on performance measure and its variability. The obtained information is essential for a continuous improvement of quality. As an illustration, two sets of experiment are designed and performed at highly increased stress levels. The results are analysed and discussed and a time saving alternative is proposed. The combination of experimental design and accelerated life testing is discussed and illustrated. The combined use of these methods can be argued for in two different cases. One is for an exploratory improvement investigation and the other is for verification of reliability. In either case, the combined use is advantageous and improves the testing efficiency. Some general conclusions are drawn to be used for planning and performance of statistically designed accelerated life testing experiments. (70 refs.) (au). (Atomindex citation 24:007028)