基于随机响应面的PC连续刚构桥施工控制可靠度与灵敏度分析

发布时间：2019-03-30 16:17

【摘要】：在预应力混凝土连续刚构桥的施工过程中,结构尺寸、材料性能和荷载存在随机性,使主梁的挠度和应力难以与设计的理想状态吻合。因此,本文将随机响应面应用于可靠度和灵敏度分析中,对桥梁的施工控制进行研究。首先,提出了针对桥梁施工控制的随机响应面法,并解决了非标准正态施工参数的输入问题。通过一个悬臂梁的数值算例,对比三种配点方法的优缺点。分析表明:相比于拉丁超立方抽样法(Latin Hypercube Sampling,LHS)和蒙特卡洛抽样法(Monte Carlo Sampling,MCS),基于线性无关原则的概率配点法效率最高,并可改变阶次和增加配点数量,从而提高随机响应面的精度。其次,结合随机响应面与基于广义λ分布(General Lambda Distribution,GLD)的高阶矩法,拟合状态变量的概率密度函数,从而建立了基于随机响应面的主梁挠度控制可靠度分析法。此外,以多重随机响应面为工具求解施工全过程的失效概率,可以避免复杂的积分,提高计算效率。再次,将随机响应面与灵敏度分析结合,分别发展了准全局灵敏度分析法和方差灵敏度分析法。分析表明:两种方法的计算精度均取决于随机响应面,并且可以得到相同的参数灵敏度排序。方差灵敏度分析法可以考虑施工参数之间的交叉作用;而准全局灵敏度分析法计算量少,且简单实用。最后,以李家洼大桥为工程背景,在确定主要施工参数概率分布的基础上,利用基于随机响应面的可靠度与灵敏度方法进行分析。研究表明:主梁挠度控制的失效概率随着悬臂的伸长而升高,并在合龙后迅速降低。施工参数灵敏度在施工期内也发生显著变化,其中预应力束张拉控制应力对主梁挠度和应力的灵敏度随着悬臂的伸长而降低,恒载的灵敏度反而升高。此外,混凝土抗压强度与弹性模量只对挠度产生较小影响。
[Abstract]:In the construction process of prestressed concrete continuous rigid frame bridge, the structural size, material performance and load are random, which makes the deflection and stress of the main beam difficult to fit with the ideal state of the design. Therefore, the random response surface is applied to the reliability and sensitivity analysis, and the construction control of the bridge is studied in this paper. Firstly, a random response surface method for bridge construction control is proposed, and the input problem of non-standard normal construction parameters is solved. A numerical example of cantilever beam is given to compare the advantages and disadvantages of the three methods. The analysis shows that compared with Latin hypercube sampling (Latin Hypercube Sampling,LHS) and Monte Carlo sampling (Monte Carlo Sampling,MCS), the probabilistic collocation method based on linear independent principle has the highest efficiency, and can change the order and increase the number of collocation points. Thus, the precision of random response surface is improved. Secondly, the probability density function of the state variable is fitted by combining the stochastic response surface with the higher order moment method based on the generalized 位 distribution (General Lambda Distribution,GLD, and the reliability analysis method for deflection control of the main beam based on the stochastic response surface is established. In addition, using multi-random response surface as a tool to solve the failure probability of the whole construction process, the complex integration can be avoided and the calculation efficiency can be improved. Thirdly, the quasi-global sensitivity analysis method and variance sensitivity analysis method are developed by combining random response surface with sensitivity analysis. The analysis shows that the calculation accuracy of the two methods depends on the random response surface, and the same order of parameter sensitivity can be obtained. The variance sensitivity analysis method can take into account the cross-action between construction parameters, and the quasi-global sensitivity analysis method is simple and practical because of its less computation. Finally, on the basis of determining the probability distribution of the main construction parameters, the reliability and sensitivity method based on random response surface is used to analyze the Lijiawa Bridge. The results show that the failure probability of deflection control increases with the cantilever elongation and decreases rapidly after closure. The sensitivity of construction parameters also changes significantly during the construction period, in which the sensitivity of prestressed beam tension control stress to the deflection and stress of the main beam decreases with the cantilever elongation, while the sensitivity of dead load increases. In addition, the compressive strength and elastic modulus of concrete have little effect on deflection.
【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】：U445.4

【参考文献】