基于逆可靠度理论的边坡稳定分析方法研究

发布时间：2018-08-14 20:19

【摘要】：边坡工程是岩土工程界的经典研究领域之一,其对象是经历了亿万年漫长地质史时期的岩土体,其赋存环境、组成成分、物理、力学特征等因素存在不同程度的非确定性。在边坡工程的分析和设计中,为刻画这些因素不确定性程度的影响,常采用安全系数或可靠度指标来进行表征。但安全系数侧重揭示平均抗力与平均荷载的相对数量关系,而可靠度指标提示的是在既定条件下存在需要被动承受失效风险的概率。对于存在不确定性影响因素的工程而言,完全消除失效风险是不可能的,重点在于明确相关因素对风险的影响程度,依据具有针对性的方法,通过相应的设计分析,为制订合理的工程处治措施指明方向,主动将风险控制在可接受的范围之内。而目前的相关分析方法基本上只着重地体现了其中的某个方面,没有达到完备地能同时实现上述功能的程度。为此,本文将在现有边坡稳定分析方法的研究基础上,结合边坡工程自身的特征开展一系列的研究工作,进而建立起一种基于风险主动控制理念的边坡稳定评估新方法。首先,根据边坡破坏时沿规则滑面滑动和任意滑面滑动的两类破坏形式,相应地从边坡极限平衡理论体系中挑选出毕肖普模式和简布模式作为代表,基于风险控制理念,利用一次逆可靠度原理,演绎了边坡稳定逆可靠度算法,导出了两类滑动模式的相关计算公式,形成了具有可靠度和安全系数双重控制指标的能完成简单情况边坡计算的边坡稳定逆可靠度分析方法框架。该方法基于事先可接受的失效风险(目标可靠度指标)设计影响边坡稳定的各类参数,使可靠度指标与稳定性系数相互对应,达到主动控制边坡失效风险的目的。其次,岩土参数的非确定性描述和边坡稳定状态的函数表达形式决定着逆可靠度设计分析方法的实用性及其计算结果的有效性,影响边坡稳定的众多参数在不确定特征上往往又体现为随机性、区间性、灰色性等多样化现象,同时还表现为各参数对边坡稳定状态影响程度的非均等性。因此,在建立基于充满空间、多水平、稳健性的均匀设计进行试验样本提取方法的基础上,将灰色关联理论嵌入到区间分析模型之中,求解各因素与响应值间的区间关联度,并通过结合相对优势度理论对区间关联度进行排序,构建出基于不确定性理论的边坡稳定影响因素显著性综合评估方法。这一方法既为建立边坡稳定力学状态描述函数时,合理剔除某些因素、约简边坡状态模型提供了筛选工具,同时为下一步提高风险控制评估方法的实用性准备了相应手段。结合边坡参数分布形态各异及参数间存在相关性的实际工程特征,从响应面函数模型自身构建条件入手,继续展开深入的研究与探讨。首先,通过引入Nataf变换把边坡结构中具有多种分布形态及有相关性的参数进行独立正态化,然后,利用独立正态化的参数来拟合响应面函数模型,据此构造的响应面模型能适用于实际边坡工程中参数分布形式的多样性及参数间的相关性特征,最后,将前述拟合出的响应面模型和一次逆可靠度理论结合,构建出更具合理性和有效性的边坡失稳风险主动控制评估方法。再次,为进一步完善并丰富边坡失稳风险主动控制评估方法的研究内容,提高响应面建模手段在处理边坡工程隐式功能函数中的适用性,结合边坡工程极限状态功能函数复杂、高度非线性的特征,综合考虑了响应面模型中函数表达形式及抽样方式。在响应面模型的选择上,针对多项式模型适用范围的局限性,采用理论上更为严谨、操作上更具灵活性、对复杂非线性问题处理更为有效的Kriging代理模型；对于试验样本点的采样方式,则将对复杂非线性模型分析效率更高、适应性更强的拉丁超立方试验设计方法与具有主动搜索最佳训练样本的主动学习法相结合。上述响应面模型与抽样方式的融合将建立起一种更为完善的边坡稳定逆可靠度响应面技术,拓宽了解决边坡复杂隐式功能函数失效风险评估方法的适用范围。最后,利用本文所构建出的一套较为完善、适合不同状况下的边坡稳定影响因素显著性辨识方法及失稳风险主动控制评估方法指导了连接湘北与鄂东的山区高速公路一通平高速公路沿线大量边坡的稳定性分析及设计。在第六章中展示了基于上述方法指导的其中两个路段边坡的稳定性分析过程及主要结论。工作成果为通平高速公路沿线边坡提供了实用有效的处治方案,同时验证了本文所建立的边坡失稳风险主动控制评估方法的合理性与适用性。
[Abstract]:Slope engineering is one of the classical research fields in geotechnical engineering. The object of slope engineering is the rock and soil mass which has undergone a long geological history for hundreds of millions of years. The factors such as its occurrence environment, composition, physical and mechanical characteristics are uncertain in varying degrees. The safety factor or reliability index is often used to characterize the failure. But the safety factor focuses on revealing the relative quantitative relationship between the average resistance and the average load, and the reliability index indicates that there is a probability of passive failure risk under given conditions. Risk is impossible, the key point is to make clear the influence degree of the relevant factors on the risk, according to the pertinent method, through the corresponding design analysis, to make reasonable engineering treatment measures to indicate the direction, and actively control the risk within the acceptable range. In this paper, a series of research work will be carried out on the basis of the existing slope stability analysis methods and the characteristics of slope engineering, and a new method of slope stability assessment based on the concept of risk active control will be established.
Firstly, according to the two failure modes of sliding along regular sliding surface and arbitrary sliding surface in slope failure, Bishop model and simple distribution model are selected as representative from the limit equilibrium theory system of slope. Based on the risk control concept, the inverse reliability algorithm of slope stability is deduced by using the principle of primary inverse reliability, and two kinds of inverse reliability algorithms are derived. A framework of inverse reliability analysis method for slope stability with double control indexes of reliability and safety factor is developed. The method is based on the pre-acceptable failure risk (target reliability index) to design various parameters affecting slope stability so as to achieve reliability. The standard and stability coefficient correspond to each other, so as to achieve the purpose of active control of slope failure risk.
Secondly, the non-deterministic description of geotechnical parameters and the functional expression of slope stability state determine the practicability of inverse reliability design and analysis method and the validity of calculation results. Many parameters affecting slope stability are often manifested as random, interval, grey and other diversified phenomena in uncertain characteristics. For the non-uniformity of the influence degree of each parameter on the stability state of the slope, the grey relational theory is embedded into the interval analysis model to solve the interval correlation degree between each factor and the response value, and the relative degree is combined by establishing the method of extracting test samples based on the uniform design which is full of space, multi-level and robustness. The dominance degree theory ranks the interval correlation degree and constructs a comprehensive evaluation method for the significance of the influencing factors of slope stability based on the uncertainty theory. The practicability of the control evaluation method is prepared accordingly.
Combining with the actual engineering characteristics of the different distribution patterns of slope parameters and the correlation between parameters, this paper proceeds with the construction conditions of the response surface function model, and then proceeds to carry on the thorough research and discussion. The response surface function model is fitted by independent normalized parameters, and the response surface model constructed by this method can be applied to the diversity of parameter distribution forms and the correlation characteristics between parameters in practical slope engineering. Finally, the response surface model and the first-order inverse reliability theory are combined to construct a more reasonable and effective response surface model. Active risk assessment method for slope instability risk.
Thirdly, in order to further improve and enrich the research content of active control and evaluation method of slope instability risk, and to improve the applicability of response surface modeling method in dealing with implicit function of slope engineering, the function expression form of response surface model is considered comprehensively considering the characteristics of complex and highly nonlinear function of slope engineering limit state. In response surface model selection, the Kriging proxy model which is more rigorous in theory, more flexible in operation and more effective in dealing with complex nonlinear problems is adopted in view of the limitation of the application scope of the polynomial model. A more adaptable Latin hypercube test design method is combined with an active learning method with active search for the best training samples. The integration of the above response surface model and sampling method will establish a more perfect inverse reliability response surface technique for slope stability, which will widen the risk assessment of failure of complex implicit function of slope. The applicable scope of the method.
Finally, a set of comparatively perfect identification method of slope stability influencing factors under different conditions and assessment method of Instability Risk active control are used to guide the stability analysis and design of a large number of slopes along the mountain expressway-Tongping expressway connecting northern Hunan and Eastern Hubei. The results provide a practical and effective treatment scheme for the slope along the Tongping Expressway, and verify the rationality and applicability of the proposed method.
【学位授予单位】：湖南大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TU43

【参考文献】