Conventional correlation factors in the safety analysis of stochastic systems

Abstract

The implantation of simple engineering techniques in time-dependent survival probability predictions of structural and technical systems as stochastic systems of events is discussed. A possibility to avoid complicated multidimensional integrations in a probabilistic safety analysis of stochastic systems with perfectly ductile components is based on the approaches of Transformed Conditional Probabilities (TCP) and Conventional Correlation Vectors (CCV). Dynamic (time-dependent) autosystems of extreme events of particular single and mixed ductile elements are treated as correlated components of static stochastic systems with m random failure modes. The unsophisticated and fairly exact prediction of the probability-based reliability of stochastic systems of events is demonstrated by numerical examples and histograms of their reliability indices.