基于统计反演理论的既有桥梁冲刷风险评估研究

发布时间：2018-02-12 05:57

本文关键词： 冲刷风险评估桥墩统计反演分析正演模型后验概率　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：桥梁是交通系统中的重要枢纽,近几年桥梁事故的发生,对社会经济发展造成了很大的负面影响,而就当前的风险评估方法还存在着一些不足,如多数风险评估是基于定性分析,对于定量分析的研究较少,评估过程中常带有一定的主观因素,所以要求进一步改进和完善当前的风险评估理论与方法。本文基于统计反演理论,提出一种风险评估方法,对冲刷风险场景进行分析,并对之中的不同影响因素进行了分析,本文的主要工作如下:(1)建立解析模型与有限元模型以计算冲刷场景下的桥墩临界荷载,作为失效功能函数中的桥墩的抗力,根据所建立的计算模型同时对冲刷风险案例的分析,归纳桥墩冲刷过程中的风险源。(2)建立了基于统计反演理论的冲刷风险评估方法,在该方法中对先验信息的分布基于实测值应用贝叶斯更新,得到改进后的后验分布,在正演模型中通过蒙特卡洛抽样即可计算失效概率的大小。(3)应用所建立的风险评估方法,以苏通大桥的北引桥为例,分别用先验信息与使用统计反演更新后的后验信息对桥梁冲刷场景的风险概率进行计算,对比二者的结果,采用后验信息计算后,冲刷深度的变异系数变小,使风险评估的结果更为稳定。(4)通过对比正演模型采用中国水文规范与美国水文规范的评估结果,结果显示采用美国水文规范进行评估的结果较为保守,分析计算了不同的冲刷参数分布类型的评估结果,在采取不同分布形式后,评估结果会有较大偏差,分析了模型参数相关性对结果的影响,分析了实测参数信息对于风险评估结果的影响,结果显示采用不同的实测信息更新后,会得到不同评估结果,这也表明风险概率评估是一个动态过程,应根据具体的环境对桥梁风险加以评估。
[Abstract]:Bridge is an important hub in traffic system. In recent years, bridge accidents have caused a great negative impact on the social and economic development. However, there are still some shortcomings in the current risk assessment methods. For example, most risk assessment is based on qualitative analysis, but there is less research on quantitative analysis, and there are some subjective factors in the evaluation process. In this paper, based on the statistical inversion theory, a risk assessment method is proposed to analyze the scour risk scenarios, and the different influencing factors are analyzed. The main work of this paper is as follows: (1) the analytical model and the finite element model are established to calculate the critical load of the bridge pier in the scouring scenario, which is regarded as the resistance of the pier in the failure function. According to the established calculation model, the case of scour risk is also analyzed. The risk source in the course of bridge pier scouring is summarized. The method of scouring risk assessment based on statistical inversion theory is established. In this method, the distribution of prior information is updated by Bayesian method based on the measured value, and the improved posterior distribution is obtained. In the forward model, the magnitude of failure probability can be calculated by Monte Carlo sampling. The established risk assessment method is applied. Taking the north approach bridge of Sutong Bridge as an example, The risk probability of the bridge scour scene is calculated by using the prior information and the updated posteriori information by statistical inversion, and the coefficient of variation of the scour depth becomes smaller after the use of the posteriori information to calculate the risk probability of the scour scene. The results of risk assessment are more stable. (4) by comparing the results of Chinese hydrological code and American hydrological code, the results show that the evaluation results of American hydrological code are conservative. The evaluation results of different distribution types of scour parameters are analyzed and calculated. After adopting different distribution forms, there will be a large deviation in the evaluation results. The influence of model parameter correlation on the results is analyzed. The effect of measured parameter information on risk assessment results is analyzed. The results show that different evaluation results can be obtained by updating different measured information, which also indicates that risk probability assessment is a dynamic process. Bridge risk should be assessed according to specific environment.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U447;U442.32

【参考文献】