ABSTRACT

The aim of this research is to develop a probability method that can identify the flexure, shear or flexural-shear failure of RC columns. A total of 336 columns of cyclic load test results were collected, and a probabilistic model was constructed using the Bayesian procedure. Then, the model was simplified by Bayesian updating. The boundaries for each failure mode were selected through an internal search optimization. The results show that the Bayesian methodology can accurately predict the seismic failure modes of RC columns and the comparison of the identification with other methods underscore the validity of Bayesian method.

中文翻译：

---

基于贝叶斯理论的钢筋混凝土柱地震破坏模式识别

摘要

本研究的目的是开发一种概率方法，可以识别 RC 柱的弯曲、剪切或弯曲剪切破坏。共收集336列循环载荷试验结果，采用贝叶斯程序构建概率模型。然后，通过贝叶斯更新简化模型。每个故障模式的边界是通过内部搜索优化选择的。结果表明，贝叶斯方法可以准确预测钢筋混凝土柱的地震破坏模式，与其他方法的识别比较强调了贝叶斯方法的有效性。