高层建筑风荷载识别研究

发布时间：2018-01-07 21:42

本文关键词：高层建筑风荷载识别研究　出处：《武汉理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：近一百年来,高层建筑飞速发展,高层建筑越来越高,结构的柔度越来越大,尽管至今还没有因为风荷载导致高层建筑倒塌的事故发生,但是在强风的过程中超高层建筑因为振动过大致使使用者不适的现象却常有发生。在这种情况下风荷载应该成为超高层建筑设计安全性和实用性的控制荷载;所以,我们需要了解风荷载。但是因为风荷载的很多数据都是来自气象部门或者是军事部门,并且与风工程研究的方向不同,因此关于风荷载的记录比较少;直接去测量风荷载是很难的,而对响应的测量会简单很多,所以近些年来,通过测量出结构的响应再来反演出风荷载成为获得风荷载数据的新方法,本文就将详细介绍风荷载复合反演的内容。本文运用的方法是荷载归一化统计平均算法,这种方法是在脉动风相关性的假定下,运用动力系统状态方程的经典形式和最小二乘准则,在成功的推导出荷载归一化统计平均算法以后,为后面的实例分析提供了方法上的支持;然后对此方法进行了收敛性证明。在实际工程中,分为下面几种情况:1.有噪声和无噪声情况的区别,2.不同的噪声水平对反演结果的影响,3.不同脉动风相关性假定对反演结果的影响等。无噪声的情况下进行结构的物理参数识别,取不同的初值,经计算最大误差不超过0.99%;同样,反演出的风荷载与真实值的误差也较小,在1.25%以内说明此方法在刚度识别方面的准确性与假设的初始刚度无关,结果令人满意。我们在噪声水平为15%的情况下观察刚度识别结果和风荷载反演结果在脉动风相关性系数ρ分别为4、8、12的三种情况;结果发现,不同脉动风相关性系数对刚度的识别结果和反演结果影响不大,但是在相关性系数ρ最小为4的时候,刚度识别结果风荷载反演结果较好,更符合实际情况。为了验证该算法的抗噪声性能,我们加入不同噪声水平的高斯白噪声到理论计算中,结果说明,在低噪声情况下(5%,15%),由于刚度识别的准确度很高,所以反演得到的风荷载也具有较好的准确性;可是如果噪声水平达到30%,风荷载的识别结果出现的误差较大达到9.3%。整体来看,大量的理论分析和数值分析表明了本文提出的荷载归一化复合反演算法的有效性和实用性。
[Abstract]:In the past one hundred years, the rapid development of high-rise buildings, high-rise building more and more high, more and more flexibility structure, although has not occurred because of wind load to high-rise building collapse accidents, but in the process of high-rise building in strong winds because the vibration will be over user discomfort but often occur in this case. Wind load should be the design of super high-rise building safety and practicality of the control load; therefore, we need to understand the wind load. But because of a lot of data of wind load are from the meteorological department or military departments, and research and wind engineering in different directions, so less about the wind load directly to the measurement of wind records; the load is very difficult, but the response measurement will be much simpler, so in recent years, by measuring the response of the structure and performance of wind load for anti wind load data The new method, this paper introduces the wind load composite inversion content. This paper uses the method of load is the normalized statistical average algorithm, this method is the assumption that the correlation of fluctuating wind power system, using the classical form of state equation and least square criterion, a normalized load statistical average algorithm in the derivation of success, provided a method for the back of the case analysis; then this method of convergence is proved. In practical engineering, divided into the following situations: the difference between the 1. noise and no noise. The effect of 2. different noise levels on the inversion results, 3. different wind correlation assumptions of the inversion results. To identify the physical parameters of the structure without noise under the condition of different initial value, the calculation of the maximum error is less than 0.99%; the same error, the wind load and the value of the real. The smaller, less than 1.25% in the initial description and accuracy of this method in the assumption of stiffness identification of stiffness has nothing to do with satisfactory results. We are in the noise level of 15% cases observed stiffness identification results and wind load inversion results in fluctuating wind relativity coefficient for each of the three 4,8,12 the results showed that different; the fluctuating wind correlation coefficient has little influence on the stiffness of the identification results and the inversion results, but the minimum relativity coefficient is 4 when the stiffness identification results of wind load inversion results is better, more in line with the actual situation. In order to verify the anti noise performance of the algorithm, we added different noise levels of Gauss white noise to the theoretical calculation the results shows that, in low noise conditions (5%, 15%), because the stiffness identification accuracy is very high, so the wind load obtained has good accuracy; but if the noise level of Up to 30%, the error of wind load identification results is larger than that of 9.3%.. As a whole, a large number of theoretical analysis and numerical analysis show that the proposed load normalization composite inversion algorithm is effective and practical.

【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU973.213

【参考文献】