钢管混凝土拱桥正常使用极限状态可靠度分析

发布时间：2018-02-11 20:08

本文关键词： 钢管混凝土拱桥可靠度分析正常使用极限状态 BP神经网络 JC法　出处：《长安大学》2015年硕士论文　论文类型：学位论文

【摘要】：本文基于桥梁结构可靠度基本理论,针对现有可靠度分析方法在工程应用中存在的问题,将BP神经网络与改进JC算法相结合,对钢管混凝土拱桥正常使用极限状态可靠度进行计算,为钢管混凝土拱桥可靠度评估提供理论参考。主要研究内容如下:(1)根据极限状态基本定义,明确钢管混凝土拱桥正常使用极限状态下作用及抗力概率模型,讨论传统可靠度计算方法在钢管混凝土拱桥结构分析中存在的缺陷,并结合梯度计算法,得出钢管混凝土拱桥极限状态方程参数灵敏性分析方法。(2)基于桥梁结构可靠度基本理论及钢管混凝土拱桥结构特点,结合神经网络理论,采用均匀设计法,将结构功能函数显式化。通过改进JC算法迭代方式,求解由BP神经网络拟合出的非线性功能函数,提高结构可靠度分析精度,实现从结构体系层面对钢管混凝土拱桥正常使用极限状态挠度、应力可靠度的评估,提出完整的可靠度分析、计算方法。并以MATLAB计算程序为平台,采用系统化、模块化设计思路编制可靠度分析程序,并结合简单算例对程序可行性及准确性进行验证。(3)建立实桥有限元模型,利用BP神经网络与改进JC算法,从挠度及应力两方面对实桥工程可靠性进行评估,研究钢管混凝土拱桥正常使用极限状态的主要失效模式。并基于主要受力构件应力可靠度指标,建立钢管混凝土拱桥的结构等效体系及简化等效体系,计算了等效串联体系模式下结构应力可靠度指标。分析可靠度指标对正常使用极限状态方程参数的灵敏性,并结合灵敏度计算结果,减少随机变量个数,简化可靠度计算过程。
[Abstract]:In this paper, based on the basic theory of bridge structure reliability, the BP neural network and the improved JC algorithm are combined to solve the problems existing in the existing reliability analysis methods. The limit state reliability of CFST arch bridge is calculated, which provides a theoretical reference for the reliability evaluation of CFST arch bridge. The main research contents are as follows: 1) according to the basic definition of limit state, The function and resistance probability model of CFST arch bridge under the limit state of normal use is defined. The defects of traditional reliability calculation method in structural analysis of CFST arch bridge are discussed, and the gradient calculation method is combined. Based on the basic theory of reliability of bridge structure and the structural characteristics of concrete-filled steel tube arch bridge, the uniform design method is adopted in combination with the neural network theory. By improving the iterative method of JC algorithm, the nonlinear functional function fitted by BP neural network is solved, and the precision of structural reliability analysis is improved. In this paper, the deflection and stress reliability of concrete-filled steel tubular arch bridge under normal use are evaluated from the structural system level, and a complete reliability analysis and calculation method is put forward. Based on the MATLAB calculation program, the system is adopted. The reliability analysis program is compiled by modularization design idea, and the feasibility and accuracy of the program are verified by a simple example. The finite element model of the real bridge is established, and the BP neural network and the improved JC algorithm are used. In this paper, the reliability of real bridge is evaluated from deflection and stress, and the main failure modes of concrete filled steel tube arch bridge under normal service limit state are studied, and based on the stress reliability index of main stress components, the failure mode of concrete filled steel tube arch bridge is studied. The structural equivalent system and simplified equivalent system of concrete-filled steel tubular arch bridge are established, and the structural stress reliability index under the equivalent series system mode is calculated. The sensitivity of the reliability index to the parameters of the limit state equation of normal use is analyzed. Combined with the result of sensitivity calculation, the number of random variables is reduced and the process of reliability calculation is simplified.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U448.22;U441

【参考文献】