BibTex format
@inproceedings{Todd:2018,
author = {Todd, MD and Leung, M and Corcoran, J},
title = {A probability density function for uncertainty quantification in the failure forecast method},
year = {2018}
}
Publications
Citation
RIS format (EndNote, RefMan)
TY - CPAPER
AB - It has been observed that many material failure modes follow empirically similar trends across many types of materials and load states. This commonality, rooted in the underlying hypothesis in a positive feedback mechanism, has led to several generic models that exhibit this mechanism, generally now known as the “failure forecast method”. This method essentially links the rate of change in structural health monitoring (SHM) data (features), which indicate something about current structural state or performance, to a prediction of when such data are representative of failure (characterized by an infinite data rate of change). Given inevitable noise in data, this paper will derive an uncertainty model in a practical implementation of the classic failure forecast method, where the inverse rate of change of the feature is linearly related to the time of expected failure. A probability density function (PDF) is proposed for the estimation of that failure time from updated linear regressions of data obtained during a simulated fatigue experiment. The mean, median, and mode of the data will be compared as predictors of the time of failure, and the effects of regression selection parameters (e.g., regression time frame, regression block overlap, etc.) will be explored.
AU - Todd,MD
AU - Leung,M
AU - Corcoran,J
PY - 2018///
TI - A probability density function for uncertainty quantification in the failure forecast method
ER -