Failure Prediction for Non-Linear Structural Systemsby Jon D. Collins, ACTA Inc, United States,
David A. Evensen, ACTA Inc, United States,
William F. Heubach, ACTA Inc, United States,
Document Type: Proceeding Paper
Part of: Structural Safety and Reliability
A method is described that makes failure predictions for structures which are complex, nonlinear and expensive to evaluate on a computer. The method makes use of a non-linear response surface and a multi-level failure definition. After the development of the response surface, a Monte Carlo simulation is used to predict failure at specific locations and also for the overall system. Physical property uncertainties and correlations and model uncertainties are considered in the analysis. The response levels are defined first linguistically, e.g. cracking, severe distress, complete collapse, etc. and then are correlated to various numerical levels of response. The levels are described by probability distributions to accommodate the uncertainty in the definitions.
Services: Buy this book/Buy this article
Return to search