桥梁风荷载的非高斯特性研究
发布时间：2018-08-07 21:16
【摘要】：随着建造技术的逐渐进步,现代桥梁向着“长大轻柔”的方向不断发展。结构刚度及阻尼随之降低,使得桥梁对风荷载的作用更加敏感。风荷载作为典型的随机过程,其作用极值一直是影响桥梁结构设计的重要因素,传统计算方法一般将其假设为高斯过程来估计极值。越来越多的试验及现场观测结果表明,桥梁所受风荷载有时呈现出明显的非高斯特性,若根据高斯假定计算极值,结果可能偏于危险。在此背景下,本文针对桥梁风荷载的非高斯特性展开研究,主要工作包括以下几方面：(1)介绍风荷载非高斯特性研究现状及荷载极值的主要算法,结合理论分析和实际应用总结各种算法的优势及缺点。经过计算验证,发现改进峰值因子法计算结果偏保守；Gumbel法受短时距下极值间相关性影响,其计算结果稳定性不佳;Sadek-Simiu法仅适用于偏度值较大的一类非高斯过程极值计算。(2)总结了非高斯风荷载极值计算的几大难题：A.如何借助前四阶统计矩尽量全面地描述无穷多样的非高斯分布?B.如何由少量试验样本提取更多信息,从而更好地描述随机过程?C.如何在解决好相关性影响的同时保证极值样本数量,从而保障极值分布的拟合精度?D.如何实现一种计算方法适用范围的最大化,即同时适用于软、硬响应过程,同时适用于最大、最小峰值的计算?(3)针对性提出一种基于仿真技术的风荷载过程极值算法：利用仿真技术生成多条风压时程来模拟时程样本空间,更好地描述了随机过程,且满足了极值分布拟合的样本数量；利用Johnson变换生成具有任意前四阶统计矩的非高斯序列,来模拟无穷多样的非高斯分布；利用AR自回归模型完成对风压过程相关性的仿真；本文算法对非高斯过程的概率特性和相关特性均具有自适应性,针对不同原始测量结果,产生不同概率密度函数及AR模型完成仿真,最终利用经典极值理论完成对风压极值的计算,因此本文算法适用于全部类型的平稳风压过程极值计算。(4)基于桥梁主梁节段模型测压试验结果,总结主梁表面非高斯风压区域分布特点。采用几种已有风荷载极值计算方法及本文提出的算法分别计算典型测点处风压极值,并与真实测量结果进行对比分析,证明了本文算法可适用于全部类型的风荷载过程,且与其他算法相比具有更高的计算精度。
[Abstract]:With the gradual progress of construction technology, modern bridges are developing in the direction of "growing up and soft". The stiffness and damping of the bridge are reduced, which makes the bridge more sensitive to wind load. As a typical stochastic process, the action extremum of wind load is always an important factor that affects the design of bridge structure. The traditional calculation method usually assumes the extreme value as Gao Si process to estimate the extreme value. More and more experiments and field observations show that the wind loads of bridges sometimes show obvious non-Gao Si characteristics. If the extreme value is calculated according to the Gao Si assumption, the results may be partial to the danger. In this context, the non-Gao Si characteristics of wind load of bridges are studied in this paper. The main work includes the following aspects: (1) introduce the research status of non-Gao Si characteristics of wind load and the main algorithms of load extremum. Combined with theoretical analysis and practical application, the advantages and disadvantages of various algorithms are summarized. It is found that the modified peak factor method is more conservative than the Gumbel method, which is influenced by the correlation between the extreme values at short time intervals. The Sadek-Simiu method is only suitable for the extreme value calculation of a class of non-Gao Si processes with large deviation. (2) several difficult problems in calculating the extreme value of non-Gao Si wind load are summarized. With the help of the first four order statistical moments, how to describe the infinite variety of non-Gao Si distributions as comprehensively as possible. How to extract more information from a small number of experimental samples, so as to better describe the stochastic process. How to solve the influence of correlation and ensure the number of extreme samples so as to ensure the fitting accuracy of extreme value distribution? How to maximize the scope of application of a calculation method, that is, to apply to both soft and hard response processes, and at the same time to maximize, (3) A wind load process extremum algorithm based on simulation technology is proposed. The simulation technique is used to generate multiple wind pressure time history to simulate the time history sample space, which better describes the stochastic process. Johnson transform is used to generate the non-Gao Si sequence with arbitrary first four order moments to simulate the infinite variety of non-Gao Si distribution, and AR autoregressive model is used to simulate the correlation of wind pressure process. This algorithm is adaptive to the probability and correlation characteristics of non-Gao Si processes. Different probability density functions and AR models are generated for different original measurement results. Finally, the classical extreme value theory is used to calculate the extreme value of wind pressure, so this algorithm is suitable for all types of extreme value calculation of stationary wind pressure process. (4) based on the test results of bridge main girder segment model, The distribution characteristics of non-Gao Si wind pressure on the main beam surface are summarized. Several existing wind load extremum calculation methods and the algorithm proposed in this paper are used to calculate the wind pressure extremum at typical measuring points, and the results are compared with the real measurement results. It is proved that the proposed algorithm is suitable for all types of wind load processes. Compared with other algorithms, the algorithm has higher accuracy.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U441.2
本文编号：2171367
[Abstract]:With the gradual progress of construction technology, modern bridges are developing in the direction of "growing up and soft". The stiffness and damping of the bridge are reduced, which makes the bridge more sensitive to wind load. As a typical stochastic process, the action extremum of wind load is always an important factor that affects the design of bridge structure. The traditional calculation method usually assumes the extreme value as Gao Si process to estimate the extreme value. More and more experiments and field observations show that the wind loads of bridges sometimes show obvious non-Gao Si characteristics. If the extreme value is calculated according to the Gao Si assumption, the results may be partial to the danger. In this context, the non-Gao Si characteristics of wind load of bridges are studied in this paper. The main work includes the following aspects: (1) introduce the research status of non-Gao Si characteristics of wind load and the main algorithms of load extremum. Combined with theoretical analysis and practical application, the advantages and disadvantages of various algorithms are summarized. It is found that the modified peak factor method is more conservative than the Gumbel method, which is influenced by the correlation between the extreme values at short time intervals. The Sadek-Simiu method is only suitable for the extreme value calculation of a class of non-Gao Si processes with large deviation. (2) several difficult problems in calculating the extreme value of non-Gao Si wind load are summarized. With the help of the first four order statistical moments, how to describe the infinite variety of non-Gao Si distributions as comprehensively as possible. How to extract more information from a small number of experimental samples, so as to better describe the stochastic process. How to solve the influence of correlation and ensure the number of extreme samples so as to ensure the fitting accuracy of extreme value distribution? How to maximize the scope of application of a calculation method, that is, to apply to both soft and hard response processes, and at the same time to maximize, (3) A wind load process extremum algorithm based on simulation technology is proposed. The simulation technique is used to generate multiple wind pressure time history to simulate the time history sample space, which better describes the stochastic process. Johnson transform is used to generate the non-Gao Si sequence with arbitrary first four order moments to simulate the infinite variety of non-Gao Si distribution, and AR autoregressive model is used to simulate the correlation of wind pressure process. This algorithm is adaptive to the probability and correlation characteristics of non-Gao Si processes. Different probability density functions and AR models are generated for different original measurement results. Finally, the classical extreme value theory is used to calculate the extreme value of wind pressure, so this algorithm is suitable for all types of extreme value calculation of stationary wind pressure process. (4) based on the test results of bridge main girder segment model, The distribution characteristics of non-Gao Si wind pressure on the main beam surface are summarized. Several existing wind load extremum calculation methods and the algorithm proposed in this paper are used to calculate the wind pressure extremum at typical measuring points, and the results are compared with the real measurement results. It is proved that the proposed algorithm is suitable for all types of wind load processes. Compared with other algorithms, the algorithm has higher accuracy.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U441.2
【参考文献】
相关期刊论文 前1条
1 周群艳,田澎,田志友;基于Johnson转换体系的非正态过程能力指数估计[J];系统工程;2004年05期
，本文编号：2171367
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2171367.html
