破損確率評価技術
スポンサーリンク
概要
- 論文の詳細を見る
Since risk is defined as combination of the probability of failure and the consequence of failure, these two values should be evaluated for quantitative risk analyses. In this lecture, focus is centered on technique how to evaluate failure probabilities of structural components. The conventional reliability engineering approach, which is applicable when the life distribution or failure rate of the component is available through life tests or a reliability database, is firstly reviewed. Then, structural reliability engineering approach, in which the probability of failure is evaluated based on the failure mechanism and the statistical properties of loads and material strength, etc. , is reviewed. The essence of structural reliability engineering approach is the mathematical formulation of the failure probability using the uncertainty variables and the limit-state function. It is, however, difficult to directly calculate this formula since it is a multi-fold integral of a multi-dimensional joint distribution function. The first-order reliability method (FORM) is widely used technique to approximately calculate this probability integral. The Monte Carlo simulation, a numerical method to solve probabilistic problems, can be also applied to calculate this probability integral. However, the crude Monte Carlo simulation is not realistic because it takes too much time to calculate very low failure probabilities. To conquer this problem, more efficient Monte Carlo simulation technique like importance sampling and stratified sampling is employed. A brief review on technique to evaluate the failure probability of a structural system is also presented. An emphasis on the necessity of reliability data and the usage of Bayesian approach to evaluate failure probabilities of practical structures concludes this paper.
- 一般社団法人 日本高圧力技術協会の論文