节点文献

汶川地震区崔家油房岩质斜坡的地震响应

Seismic Response of Cuijiayoufang Rock Slope in Wenchuan

【作者】 黄磊

【导师】 唐辉明;

【作者基本信息】 中国地质大学 , 地质工程, 2011, 硕士

【摘要】 汶川地震区大型边坡的复杂表现,难以用传统的滑坡机理来解释,同一区域的岩体在汶川地震中表现出不同的破坏形式,有些产生拉伸破坏有些产生剪切破坏,断层穿越崔家油房斜坡,有必要研究断层对其地震响应的影响,为工程地质条件与之相似的斜坡提供类比意义。2008年05月12日汶川大地震诱发北川陈家坝崔家油房岩质斜坡发生破坏,类型为溃喷型滑坡,滑坡过程为斜坡坡腰岩体产生拉裂,扩展并加深,而坡脚不断的震裂松弛,发生剪切变形,强度不断降低,最终与坡腰拉裂面交会,瞬间坡体的摩阻力降到最低,坡体最终溃喷下来且赋予运动初速度抛射,落地时受到右侧山体阻挡,与之发生碰撞运动方向往左偏转,而后继续运动,前方地形近于水平,运动时与地面摩擦增大且整体过程中散体岩块间发生碰撞消耗能量,最终滑坡前缘停止形成堆积物。从工程地质条件出发,此坡破坏的原因是斜坡地形高差达160m,呈两面临空的孤立山峰,高陡和突兀地形对地震波的放大效应显著,岩体裂隙发育,岩块强度低,偶见灰岩溶孔,岩体质量差,断层穿越坡体,断层带及附近岩体破碎,位于汶川大地震极震区,地面峰值加速度局部地段达1.5~2.0g。在对岩体结构分析中需求取结构面空间分布,其中一项是野外统计结构面产状,而目前多采用地质罗盘测量产状,对广泛使用的传统地质罗盘测量误差、工作效率、操作难度、读数难度等进行分析发现,传统地质罗盘操作繁琐、工作效率低、在线状结构倾伏角测量中误差大、在低倾岩层顶板产状测量中读数困难。为弥补传统地质罗盘的这些不足,发明一种新型地质罗盘,结构包括底座、瞄准器、方位刻度盘、测角器、半球形的透明罩和浮液等,使用方法为当测量线状结构的倾伏向和倾伏角时使用瞄准器2的两个尖端瞄准待测线状结构,再以瞄准器2尖端所在直线为转轴,使整个装置绕其转动,至测角器5的中线5.1与透明罩8上的对准线6重合,此时测角器5浅绿色部分中线指示的方位刻度盘4上读数为磁方位角a。,浮液3液面指示的测角器5上读数为b。,则倾伏向=a。一磁偏角,倾伏角=b。;当测量面状结构的倾向和倾角时将本装置底面紧贴岩层底板或顶板,测角器5浅绿色部分的中线5.1指示的方位刻度盘4上的读数为磁方位角a。,浮液3的液面所指示的测角器5上的读数为b。,则倾向=a。一磁偏角,倾角=b。。与传统地质罗盘相比具有如下优越性:1、测量线状结构时能够消除对倾伏角的测量误差;2、测量面状结构时无需手动调整水平,免除了两次调整水平和减小了误差传播,提高测量的准确性和效率;3、当测量低倾岩层顶板的产状时,有足够的空间和明亮光线进行读数。当假定结构面形状为圆盘,直径成为岩体结构面空间分布的重要参数,须由实测迹长推求。目前:(1)对于结构面与露头面边界交切概率较大的岩体要求取直径概率分布形式,不宜采用窗口法和全迹长测线法而宜采用半迹长测线法;(2)在半迹长测线法中末删截半迹长与直径概率关系是复杂的二重积分隐式函数关系,求解直径分布精确解析解困难,前人算法鲜见;(3) Priests、Zhang成果结合虽可得到直径估算公式,但计算结果准确性不高,相对误差高达7.1%-23.32%;(4)黄磊前期提出的直径算法只适用于直径服从均匀分布,而直径可能服从其它分布形式。鉴于对适用于半迹长测线法的精确算法的迫切需要,提出新试算法,以Monte-Carlo法模拟结构面几何要素分布、解析几何学推导分布关系、Excel或C++等编程试算直径分布使试验未删截半迹长分布逼近实测未删截半迹长分布至要求精度,此时试算直径分布为最佳直径分布。由于结果具有离散性,进行降低离散性试验,发现在直径服从均匀和负指数分布时10次取平均值结果的置信度是增加2倍样本容量的1.93倍和2.00倍、单次结果的2.19倍和2.13倍,说明多次结果取平均值能有效降低离散性。较Priest-Zhang法,在准确性上新试算法结果的相对误差降低一个数量级,在容错性上更好;较实体法,在计算速度、内存要求、主频要求、文件大小、内存占用量上新试算法是实体法的432倍、1/6、1/5、11/12、1/6,编程更简便、运行更稳定;较黄磊前期算法,新试算法适用分布形式更广。在汶川得到验证,值得推广。采用半迹长测线法野外统计崔家油房斜坡断层上盘寒武系(Cq)岩体和下盘志留系(Sh)板岩结构面,室内依据产状、由走向玫瑰花图和等密度图分组,由概率图法和K-S检验法获取各组倾向、倾角、隙宽最佳概率分布,由半迹长分布推算各组直径分布,由间距分布推算各组线密度,调整、校核各组几何要素概率分布参数并最终得到结构面空间几何分布,依据规范进而判断岩体结构类型,为碎裂状或薄层状,宜采用有限差分法模拟斜坡地震响应。其中:1、寒武系(Cq)岩体结构面被划分为2组、志留系(Sh)板岩结构面被划分为3组;2、CqⅠ组倾向服从正态分布、期望为311。、标准差为29。,CqⅡ组倾向服从正态分布、期望为160。、标准差为17。,ShⅠ组倾向服从对数正态分布、期望为3.23。、标准差为0.822。,ShⅡ组倾向服从正态分布、期望为331。、标准差为16。,ShⅢ组倾向服从正态分布、期望为148。、标准差为6。3、CqⅠ组倾角服从对数正态分布、期望为3.27。、标准差为0.557。,CqⅡ组倾角服从正态分布、期望为74。、标准差为8。,ShⅠ组倾角服从对数正态分布、期望为3.75。标准差为0.372。,ShⅡ组倾角服从正态分布、期望为66。、标准差为11°,ShⅢ组倾角服从正态分布、期望为63。、标准差为16。4、CqⅠ组隙宽服从负指数分布、期望为0.06mm,CqⅡ组隙宽服从负指数分布、期望为0.02mm,ShⅠ组隙宽服从负指数分布、期望为0.5191mm,ShⅡ组隙宽服从负指数分布、期望为0.409mmm,ShⅢ组隙宽服从负指数分布、期望为0.313mm;5、采用提出的结构面直径新试算法估算直径,CqⅠ组直径服从正态分布、期望为0.4m、标准差为0.6m,CqⅡ组直径服从负指数分布、期望为0.3m,ShⅠ组直径服从负指数分布、期望为3.4m,ShⅡ组直径服从负指数分布、期望为0.4m,ShⅢ组直径服从负指数分布、期望为0.9m;6、CqⅠ组线密度为21.8m-1,CqⅡ组线密度为8.9m-1,ShⅠ组线密度为14.7m-1,ShⅡ组线密度为8m-1,ShⅢ组线密度为1.8m-1。断层经过崔家油坊斜坡,有必要研究断层对其地震响应特性的影响,取有断层(实际)和无断层(假设)两种情况进行响应特性对比,大致步骤为:(1)由地震波截断频率计算允许最大网格尺寸,划分符合最大尺寸要求的网格,建立两种情况的数值模型。(2)赋予物理力学参数,采用水平和竖直方向固定约束边界条件、摩尔-库仑本构模型与屈服判剧,设置重力加速度,采用改变参数的弹塑性求解法生成坡体初始地应力场。(3)对已去仪器化地震波加速度时程进行10Hz滤波和基线校正,由于模型底部岩体为软岩且为粘滞边界,将加速度时程转换为应力时程。(4)生成初始地应力后,将模型底部设置为粘滞边界、四周设为自由场边界,采用局部阻尼,在模型底部施加地震应力时程进行动力计算,采纳郑颖人最近提出的地震破坏判剧判断岩体稳定性。崔家油房有断层和无断层两种情况斜坡地震响应特征如下:1、地震5s时,无断层,斜坡表面出现贯通的剪切塑性区,停止地震后5s坡腰和坡脚监测点位移尤其水平位移突变不收敛,说明地震5s时斜坡已产生破坏,破坏形式为剪切破坏;地震5s时有断层坡脚出现局部剪切塑性带而剪切塑性带上覆岩体完好,坡腰出现局部拉张塑性区,两塑性区没有贯通,最终坡体不产生破坏,变形形式为坡腰的拉张变形和坡脚的剪切变形。2、地震5s时坡体最大主应力对比为有断层<无断层,最小主应力对比为有断层>无断层,由莫尔判剧可知有断层坡体应力状态较无断层坡体应力状态对坡体稳定性有利,这也可从两种情况变形破坏程度的数值模拟结果验证。断层存在提高了崔家油房斜坡稳定性,可能是由于在地震波传播过程中断层减弱前方岩土体地震波强度。3、地震10s时有断层坡腰拉张塑性区与坡脚剪切塑性带贯通,停止地震后5s坡腰和坡脚监测点位移尤其水平位移突变且不收敛,从而最终判断斜坡发生破坏,破坏形式为贯通的坡腰拉张破坏与坡脚剪切破坏,而无断层时地震5s已经发生破坏,形式为整体剪切破坏。对比崔家油房有无断层斜坡地震响应得到以下结论:可能是在地震波传播过程中断层减弱前方岩体地震波强度,从而使岩体应力状态对稳定性有利,延缓破坏时间,由无断层时的整体剪切破坏改变为有断层时的坡腰拉张破坏与坡脚剪切破坏。在对崔家油房斜坡有无断层的地震响应对比中发现,断层是影响破坏时间和形式的重要因素,在对工程地质条件与之相似的斜坡或工程边坡地震分析中要重视断层对破坏时间和形式的影响。

【Abstract】 The performance of large landslides in Wuchuan Earthquake Area is complex, and it is difficult to be explained with traditional mechanisms. The failure modes of rocks in the same region are different from each other, either tensile failure or shear failure. The fault stretches acrosses Cuijiayoufang Slope. And it is necessary to study the fault influence on its seismic response, which offers reference to the study of slopes with similar engineering geological conditions.On 12 May 2008 Cuijiayoufang Rock Slope was triggered by Wenchuan Earthquake in Beichuan Chenjiaba. The type is spray collapse. Its destructive process is this:First, the cracks in tension are generated in the upper part and post-edge of the slope, then extends, while the cracks in shear are generated in the lower part. Second, the strength decreases, and both kinds of cracks intersect. Instantaneously the friction of the slope drops to the minimum, and the slope sprays down. Third, the slip mass has the initial velocity, and flies. When it lands, it is blocked by the hill on the right side. The slip mass turns left and continues moving. As the front land is near horizontal, friction with the ground increases and rock blocks crashes each other. The energy is consumed and the slide mass stops on the frontal edge. From the engineering geological conditions, there are three main causes of Cuijiayoufang Rock Slope. First, height difference of the slope is 160m, and the slope with two empty faces is isolated. This steep and lofty topography leads to significant amplification effect of seismic wave. Second, fracture is developed in rock, and the strength of rock block is low. Gray karsts holes are seen occasionally. The rock is crumbling and of poor quality. The fault crosses the slope, and the slope is located in the meizoseismal area. The peak acceleration on the ground in some areas is high of 1.5~2.0g.The spatial distribution of discontinuities is calculated in the analysis of rock structure, and one of its previous work is measuring the orientations of discontinuities in the field. The geological compasses are used to measure the orientations generally. From the study of traditional geological compasses, the disadvantages are found, that is, complex operation, low efficiency, high bias in plunge angle of linear structure, and reading difficulty in the orientations measurement of the low pour rock roof. Due to these shortcomings of traditional geological compasses, a new type of geological compasses is invented, including pedestal, collimator, azimuth dial, goniometer, hemispherical transparent cover and liquid. When measuring plunge direction and angle of line, aim at the linear structure with the two tips of the collimators, then rotate the compass on the line which passes by the two tips, till the superposition of the goniometer center line and the cover alignment. The reading of azimuth dial which the center line of green part of goniometer indicates is the magnetic azimuthα°, and the reading of goniometer which liquid level indicates is the plunge angle b°.The plunge direction is the difference betweenα°and magnetic declination. When measuring the orientation of face, the compass bottom is taken to stick to rock bottom or top, and the reading of azimuth dial which the center line of green part of goniometer indicates is the magnetic azimuthα°. The reading of goniometer which liquid level indicates is the dip angle b°.The dip direction is the difference betweenα°and magnetic declination. Compared with traditional geological compass the new kind has the following advantages:1) the plunge angle bias of line measurement can be eliminated;2) Bias can be minimized and efficiency can be improved in the face measurement without manually adjusting the horizon;3) There can be plenty of space and bright light for reading in the orientation measurement of low pour rock roof.When assuming the discontinuity is circular, diameter becoming an important parameter of the spatial distribution of rockmass discontinuities should be deduced by the observed trace length. Nowadays:(1) Window sampling and scanline sampling for complete trace are mostly used for trace field survey, however, due to outcrop condition or time limit, both the sampling methods cannot be adopted, while it is suitable to adopt scanline sampling for semi-trace. (2) The probabilistic relation between untruncated uncensored semi-trace length and diameter in the scanline sampling for semi-trace is complex implicit function, and it is difficult to acquire exact analytical solution, previous algorithms of it are rare. (3) Though the diameter estimation formula applicable for this sampling method can be derived from the integration of Priest’s and Zhang’s achievements, the result is inaccurate, that is, the relative error is up to 7.1%-23.32%. Considering that an accurate algorithm applicable for scanline sampling for semi-trace is badly needed, a new trial algorithm is proposed, which takes Monte-Carlo method and analytic geometry as its theory, and Excel, C++, etc. as tools, to estimate the best diameter distribution, making the trial untruncated uncensored semi-trace length approximate the observed one. Due to the dispersion of its result, experiments to reduce the dispersion are conducted, it is found that, the result confidence degree of taking mean of ten trials is respectively 1.93 or 2 times as that of doubling the trial sample size, and 2.19 times or 2.13 times as that of single trial, when the diameter follows uniform or negative exponential distribution. So it is illustrated that taking mean of many trials can reduce the dispersion effectively. Compared to Priest-Zhang Algorithm, on accuracy, the new trial algorithm can reduce the error for an order of magnitude, and is better in fault tolerance. On calculation speed, memory requirement, dominant frequency requirement, file size and memory occupancy, the new trial algorithm is respectively 432 times,1/6,1/5,11/12 and 1/6 as that of the entity algorithm, and more convenient to program, more stable to operate. It has been applied in Wenchuan, so it is awarding to popularize it.The discontinuities developed in the upper plate rock of Cambrian ((?)q) and the lower plate rock of Silurian (Sh) of Cuijiayoufang Slope are measured with semi-trace line sampling. Grouping is taken by Strike Rose Diagram and Orientation Isopycnal Diagram. The optimal probability distribution of each group’s dip direction, dip angle and gap width is derived from Probability Graph and K-S Test. Diameter distribution is calculated from semi-trace, and linear density from spacing. Finally the spatial distribution of discontinuities is obtained by adjusting the probability parameters observed. Then the type of its structure is judged as cataclastic or thin-layer. And it is better to simulate its seismic response with FLAC.Where,1) Group. The discontinuities in Cambrian ((?)q) rock mass are divided into 2 groups, and that in Silurian (Sh) slate is divided into 3 groups.2) Dip direction. GroupⅠof (?)q follows normal distribution. Its mean is 311°, and standard deviation is 29°. GroupⅡof (?)q follows normal distribution. Its mean is 160°, and standard deviation is 17°. GroupⅠof Sh follows lognormal distribution. Its mean is 3.23°, and standard deviation is 0.822°. GroupⅡof Sh follows normal distribution. Its mean is 331°, and standard deviation is 16°. GroupⅢof Sh follows normal distribution. Its mean is 148°, and standard deviation is 6°.3) Dip angle. GroupⅠof (?)q follows lognormal distribution. Its mean is 3.27°, and standard deviation is 0.557°. GroupⅡof Cq follows normal distribution. Its mean is 74°, and standard deviation is 8°. GroupⅠof Sh follows lognormal distribution. Its mean is 3.75°, and standard deviation is 0.372°. GroupⅡof Sh follows normal distribution. Its mean is 66°, and standard deviation is 11°. GroupⅢof Sh follows normal distribution. Its mean is 63°, and standard deviation is 16°.4) Gap width. The five groups follows negative exponential distribution. The means are: GroupⅠof (?)q 0.06mm, GroupⅡof (?)q 0.02mm, GroupⅠof Sh 0.5191 mm, GroupⅡof Sh 0.409mm, GroupⅢof Sh 0.313mm.5) Diameter. Diameter is estimated with the new algorithm proposed in this paper. The diameter of GroupⅠof (?)q follows normal distribution. Its mean is 0.4m, and standard deviation is 0.6m. The other four groups follow negative exponential distribution. The means of them are:GroupⅡof (?)q 0.3m, GroupⅠof Sh 3.4m, GroupⅡof Sh 0.4m and GroupⅢof Sh 0.9m.6) Linear density. The linear density of five groups is:GroupⅠof (?)q 21.8m-1, GroupⅡof (?)q 8.9m-1, GroupⅠof Sh 14.7m-1, GroupⅡof Sh 8m-1, GroupⅢof Sh 1.8m-1.The fault stretches across Cuijiayoufang Slope, and it is necessary to study the fault influence on its seismic response. The two cases i.e. existence (actual) and inexistence (assumptive) of fault, are taken to the comparison of seismic response. Its general steps are like this:(1) the maximum mesh size is calculated from the censoring limit of seismic wave. Then the meshes whose size is small than calculated maximum size are generated and the model is erected. (2) the physical and mechanical parameters, horizontal and vertical boundary constraints, Mohr-Coulomb constitutive model and yield criterion, gravity acceleration are set. The initial ground stress is generated with the solution of changing the elastic-plastic parameters. (3) The seismic wave acceleration time history is filtered with censoring limit of 10Hz and baseline-corrected. As the model bottom is soft rock and viscous boundary, the acceleration time history is converted to stress time history. (4) After generation of the initial stress, the model bottom is set as viscous boundary and surrounding as free-field boundary. The local damping is set. And the dynamic stress time history is applied at the model bottom for seismic calculation. At last Zheng Yingren’s seismic failure criterion is applied to judge the slope stability.The seismic response characteristics of two cases, i.e. existence and inexistence of fault in Cuijiayoufang Slope are shown as follows:1) At the end of 5s earthquake, without fault, he shear plastic zones are joined up on the slope surface. During 5s after earthquake, the displacements in the middle and toe of the slope change abruptly and do not converge, especially the horizontal displacements. And it indicates that the slope has failed at the end of 5s earthquake. Its type is shear failure; at the end of 5s earthquake, with fault, the shear plastic belt emerges in the toe, and the rock on it is not destroyed. The tension plastic zone emerges in the middle. The shear plastic belt and tension plastic zone are not joined up. In this case, the slope does not fail, and its deformation type is tension deformation in the middle and shear deformation in the toe.2) At the end of 5s earthquake, the maximum principal stress with fault is smaller than that without fault, while the minimum principal stress with fault is greater than that without fault. The stress state with fault is better for the slope stability than that without fault by Mohr Criterion, which can be verified by the comparision of deformation and failure between two cases. Therefore, the existence of fault improves the stability of Cuijiayoufang Slope, it is probably because that the fault weakens the strength of seismic wave in the ahead rock and soil.3) At the end of 10s earthquake, with fault, the shear plastic belt in the toe and tension plastic zone in the middle are joined up. During 5s after earthquake, the displacements in the middle and toe of the slope change abruptly and do not converge, especially the horizontal displacements. And it indicates that the slope has failed at the end of 10s earthquake. Its type is the perforation of shear failure in the toe and tension failure in the middle, while the type without fault is shear failure, which has occurred at the end of 5s earthquake.The conclusions are drawn from seismic responses of Cuijiayoufang Slope with fault and without fault:The fault may weaken the strength of seismic wave in the front rock. It leads to that the stress state is good for the slope stability and it delays the time of failure. The failure type is shear failure without fault, while that is shear failure in the toe and tension failure in the middle with fault. Therefore, the fault is an important factor in the time and type of failure. The attention should be paid to the fault influence on the time and type of slope failure in the future, whose engineering geological conditions is similar to Cuijiayoufang Slope.

节点文献中: