Underground Structural Reliability Of Interval Analysis Method

In recent years, many more and more underground have been calling for structural reliability analysis home and abroad. But the existing methods—probabilistic method and fuzzy method are mainly based on probabilistic distributions of the variables and a large number of statistical data of load and structural resistance are needed. However, there are so many factors influencing underground project's reliability (e.g. variables that hold randomicity, fuzzy or uncertainty) that it is almost impossible to develop a mathematical or dynamical model. So, the traditional methods can hardly satisfy the need in current reliability research.Out of the above-mentioned situation, a new method—non-probabilistic reliability measurement and analysis method based on interval model is proposed. This method only needs the bounds of variables, but their probability distributions. Consequently, the amount of the original data and calculation workload are reduced. And because probabilistic concept is not required, it bypasses some of the drawbacks of the traditional probabilistic reliability method. In condition of scarcity of original statistical data, the model provides us a good alternative for conducting structural reliability evaluation.The main wok of this paper include an introduction to the concepts of interval mathematics and computing method of interval analysis, an summarization on the non-probabilistic reliability analysis methods for different situations, and a comparison of interval analysis method with probabilistic method and fuzzy method. Analysis shows that the non-probabilistic model gives a result in consistence with other two models. Then a case study is done on a tunnel from Xiamen East Channel, which indicates that the interval analysis method is applicable in reliability analysis. Finally conclusions and remarks on future research are given.