【摘要】：地震動衰減關系在地震安全性評價、地震危險性分析及為抗震設計提供定量設防標準等工作中都發(fā)揮著重要作用。場地非線性放大模型是地震動衰減關系(GMPEs)的重要組成部分,它可以通過多種方法構造,最好的辦法是直接通過強震數據回歸得到。然而,即使是在強震記錄豐富的日本,其含有有效土體強非線性效應記錄的數量太少,以至于單純依靠強震記錄還不能夠推斷出可靠的非線性場地模型。因此本文用數值模擬方法來獲得場地反應譜放大比。本文選取的基巖強震動記錄是來自日本和PEER數據庫中不同類型的地震。這些記錄的矩震級、震源距和峰值地面加速度分布范圍廣泛。建立的一維模型是基于日本Kik-net強震臺站實測剪切波速資料,臺站的場地周期、場地土體厚度和阻抗比分布范圍也較大。本文提出一個以運用場地分類、并以一維等效線性模型為基礎的非線性場地放大模型。為了克服一維模型的局限性,本文提出了一個新的模型調整方法將非線性放大模型應用于地震動衰減關系中。本文主要工作如下：(1)收集了大量日本Kik-net、K-net和PEER強震數據,這些數據來自不同構造類型的地震,其震級、震源深度、震源距離范圍很廣。以此為基礎來建立本文研究的數據庫。(2)按Kik-net場地實測的剪切波速計算場地周期,并將場地按照計算的場地周期進行分類。依據土層深度、剪切波波速并在一個給定范圍內隨機選取土層非線性模型。用SHAKE程序等效線性分析計算出場地的放大比。(3)首先選取經驗模型擬合計算的反應譜放大比,然后基于上一步的模型來逐步進行殘差分析,確定模型參數。最終選取的模型考慮到了場地阻抗比、地震矩震級和震源距離的影響。(4)結合殘差分析所有的參數,本文建立了一個隨機效應模型,回歸得到了模型參數。在回歸分析過程中,用極大似然估計的改變量并結合各個系數中值和標準差的比值來檢驗各個參數。如果某一項參數回歸系數的中值與標準差之比小于2,表明它在統(tǒng)計上可能不顯著。從比值最小的一項參數開始檢驗,在模型中去掉這一項后,如果在統(tǒng)計上不顯著,極大似然函數值的減小量小于1,就把這一項參數從模型中去掉。從而就得出了一組場地非線性放大模型。(5)考慮到對于世界上許多真實的強震臺站和實際工程場地來說,一維模型是一個簡單的假設,所以本文設計了一個方法來調整一維模型,使它能夠適用于地震動衰減方程。
[Abstract]:The attenuation relationship of ground motion plays an important role in the evaluation of seismic safety, seismic risk analysis and the provision of quantitative fortification standards for seismic design. Site nonlinear amplification model is an important part of ground motion attenuation relation (GMPEs). It can be constructed by a variety of methods, and the best way is to get it directly from strong earthquake data regression. However, even in Japan, where strong earthquake records are abundant, there are so few records of strong nonlinear effects in soil that it is not possible to infer a reliable nonlinear site model by relying solely on strong earthquake records. Therefore, the amplification ratio of site response spectrum is obtained by numerical simulation. The bedrock strong motion records selected in this paper are from different types of earthquakes in Japan and per databases. The moment magnitude, focal distance and peak ground acceleration are widely distributed in these records. The one-dimensional model is based on the measured shear wave velocity data at the Kik-net strong earthquake station in Japan. The site period, the distribution range of soil thickness and impedance ratio are also large. This paper presents a nonlinear site amplification model based on site classification and one-dimensional equivalent linear model. In order to overcome the limitation of one-dimensional model, a new model adjustment method is proposed to apply the nonlinear amplification model to the attenuation relation of ground motion. The main work of this paper is as follows: (1) A large number of data of Kik-net-K-net and per strong earthquakes in Japan are collected. These data come from different tectonic types of earthquakes, whose magnitude, focal depth and focal distance range are very wide. Based on this, the database of this paper is established. (2) the site period is calculated according to the measured shear wave velocity of Kik-net site, and the site is classified according to the calculated site period. According to soil depth, shear wave velocity and a given range of random selection of nonlinear soil model. The amplification ratio of the site is calculated by equivalent linear analysis with Shake program. (3) the amplification ratio of the response spectrum is first selected from the empirical model fitting, and then the residual error analysis is carried out step by step based on the previous model to determine the parameters of the model. Finally, the influence of site impedance ratio, seismic moment magnitude and focal distance is taken into account. (4) combined with the residual analysis of all parameters, a random effect model is established and the model parameters are obtained by regression. In the process of regression analysis, the variables of the maximum likelihood estimation and the ratio of the median value of each coefficient to the standard deviation are used to test each parameter. If the ratio of median to standard deviation of a parameter regression coefficient is less than 2, it may not be statistically significant. When the minimum ratio parameter is removed from the model, if it is not statistically significant and the maximum likelihood function decreases less than 1, the parameter is removed from the model. Therefore, a set of nonlinear amplification models are obtained. (5) considering that the one-dimensional model is a simple hypothesis for many real strong earthquake stations and practical engineering sites in the world, this paper designs a method to adjust the one-dimensional model. So that it can be applied to the equation of ground motion attenuation.
【學位授予單位】：西南交通大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：TU435

【參考文獻】

相關期刊論文 前2條

1 李小軍,彭青;不同類別場地地震動參數的計算分析[J];地震工程與工程振動;2001年01期

2 姜治軍;胡進軍;謝禮立;張齊;;中國西部地區(qū)場地放大模型[J];天津大學學報(自然科學與工程技術版);2013年12期

相關博士學位論文 前1條

1 史大成;區(qū)域性場地地震動放大研究及應用[D];中國地震局工程力學研究所;2013年

相關碩士學位論文 前1條

1 楊俊;集集地震地震動衰減特性和相關性研究[D];哈爾濱工業(yè)大學;2011年

，

本文編號：2128090