Parts 1 and 2 presented a stochastic model of the cumulative damage process that for the first time makes it possible to take a comprehensive view of the entire process. Some of the properties of the model were discussed and some fatigue and wear data were considered. It also was shown that data on time to failure is not sufficient to characterize the process and that prediction of life under changed conditions is of necessity subject to error which cannot be removed based only on life time data. The purpose of this part is to demonstrate that since it is now possible to take a comprehensive view of the entire process accuracy in life prediction under changed conditions can be improved by the acquisition of additional data of a specific type. This data provides increasing knowledge of member (sample) function behavior that is essential for improvement in predictive accuracy.

This content is only available via PDF.
You do not currently have access to this content.