工程用地震动模拟随机性方法研究

发布时间：2018-01-17 00:15

  本文关键词：工程用地震动模拟随机性方法研究　出处：《大连理工大学》2016年博士论文　论文类型：学位论文

  更多相关文章： 土木工程 地震动模拟 随机性方法 小波包分解 设计谱拟合

【摘要】：地震动模拟可以为工程界提供大量的结构抗震设计所需的地震动数据。其中,随机性方法因其概念清晰、简单易用且模拟效果较好,受到了地震学界和工程界的普遍关注。目前使用比较广泛的地震动模拟随机性方法是基于点源法的有限断层法。然而,点源法和有限断层法在参数取值完全一致时,可能导致结果相差较大。有限断层法在大量数据的模拟效率上仍然较低。另外,基于傅里叶变换的点源法和有限断层法在模拟地震动时对于具有频率非平稳性的地震动模拟效果较差,在时频非平稳性方面具有一定的局限性。本文从工程用地震动的角度,对随机点源法和有限断层法之间的异同性进行了研究,同时研究了小波包方法在地震动非平稳性模拟方面的应用,主要研究内容如下：(1) 针对点源法与有限断层法之间的差异,提出了一种基于等效距离的随机点源法模型用来考虑地震断层面上的非一致滑动分布,并分析了该模型与有限断层法的一致性。利用该模型,对发生在中国的汶川地震(2008年)和鲁甸地震(2014年)分别进行了地震动加速度时程和反应谱模拟,以及地震等震线模拟。用以验证基于等效距离的随机点源法的正确性和高效性,以及与有限断层法的一致性。(2) 基于小波包分解和重构方法,提出了一种时频非平稳地震动模拟方法。根据实际地震动小波包系数矩阵的特点,采用6个模型参数将地震动时程的时间和频率轴展开,研究了时域和频域的联合非平稳性。对实际地震动加速度时程和已有非平稳随机模型进行了对比,对比结果证明了模型的有效性。(3) 进一步基于小波包模型,提出了一种“先模拟后拟合”的拟合工程抗震设计谱的随机性方法。利用小波包在频域上分辨率较高的特点,根据设计谱幅值对小波包系数矩阵进行调整。拟合算例证明本方法具有较好的灵活性和模拟精度。(4) 依据小波包模型,对包含有1222条实际地震加速度时程的数据库进行了两阶段回归分析,将地震参数与小波包参数联系起来,提出了本文带有地震间和地震内残差的模拟回归模型。通过对属于地震数据库中的地震进行模拟和不属于数据库中的地震进行的“预测”,说明回归模型同时具有确定性和随机性的特点。与已有地震动衰减关系进行的对比,进一步证明了回归模型的有效性。(5) 对本文提出的基于等效距离的点源法和小波包方法进行了模拟地震动的有效性检验。利用编制的EQSIM和WPSIM程序对4次历史地震进行了模拟,计算并对比了4个工程常用的结构反应指标。证明了本文提出的两种随机性方法是有效的。
[Abstract]:Ground motion simulation can provide a large amount of seismic data for seismic design of structures. Among them, the randomness method is simple and easy to use because of its clear concept and good simulation results. At present, the widely used stochastic method of ground motion simulation is the finite fault method based on point source method. Point source method and finite fault method may lead to a large difference in the results when the parameter values are completely consistent. The simulation efficiency of finite fault method in a large number of data is still relatively low. The point source method and finite fault method based on Fourier transform are not effective in simulating ground motion with frequency nonstationary. There are some limitations in time-frequency nonstationarity. In this paper, the similarities and differences between random point source method and finite fault method are studied from the point of view of engineering land motion. At the same time, the application of wavelet packet method in the simulation of non-stationary ground motion is studied. The main contents are as follows: 1) aiming at the difference between point source method and finite fault method. A stochastic point source model based on equivalent distance is proposed to consider the non-uniform sliding distribution on seismic fault plane, and the consistency between the model and the finite fault method is analyzed. Ground motion acceleration time history and response spectrum of Wenchuan earthquake (2008) and Ludian earthquake (2014) occurred in China were simulated respectively. To verify the validity and efficiency of the random point source method based on equivalent distance, and the consistency with the finite fault method. 2) the wavelet packet decomposition and reconstruction method is used to verify the validity of the random point source method based on the equivalent distance and the consistency with the finite fault method. A time-frequency non-stationary ground motion simulation method is proposed. According to the characteristics of the wavelet packet coefficient matrix of the actual ground motion, the time-frequency axis of the time-history of the ground motion is expanded by using six model parameters. The joint nonstationarity of time domain and frequency domain is studied, and the time history of actual ground motion acceleration is compared with the existing non-stationary stochastic model. The results show that the model is effective and based on the wavelet packet model. In this paper, a random method of fitting the seismic design spectrum of engineering is presented, which is "first simulated and then fitted". The wavelet packet has high resolution in frequency domain. The wavelet packet coefficient matrix is adjusted according to the design spectrum amplitude. The fitting example shows that this method has good flexibility and simulation accuracy. A two-stage regression analysis is carried out on a database containing 1222 actual seismic acceleration time histories, which links the seismic parameters with wavelet packet parameters. In this paper, a simulated regression model with inter-earthquake and intra-seismic residuals is put forward. Through the simulation of earthquakes in seismic database and the prediction of earthquakes not in database. The results show that the regression model has the characteristics of certainty and randomness, and compared with the existing attenuation relation of ground motion. The validity of the regression model is further proved. The point source method based on equivalent distance and wavelet packet method proposed in this paper are used to verify the validity of the simulated ground motion. Four historical earthquakes are simulated by using the EQSIM and WPSIM programs developed in this paper. Four commonly used structural response indexes are calculated and compared. It is proved that the two randomness methods proposed in this paper are effective.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P315.9
，

本文编号：1435457