【作者】 徐海清；

【导师】 唐辉明；

【作者基本信息】 中国地质大学 ， 环境工程， 2004， 硕士

【摘要】 水利工程是对流域资源、环境、生态等系统进行人工调控的工程。随着经济、社会和科技的发展及承受和抵御灾害能力的增强，建坝有效地改变了我们在自然条件下利用水能的状况。但是，当坝体建成运行后，坝体结构也变成了一种潜在的巨大的危险物，其安全性和使用性对社会和经济的影响相当大。岩体力学研究中，岩体参数的确定一直是研究热点课题之一。不同力学参数产生不同的计算结果，不当的力学参数还会对工程设计与施工起误导作用，许多工程地质问题都源于岩体参数估算的偏差。因此，坝区及坝基岩体力学参数取值直接关系到水利工程的安全与经济，是工程勘察、设计与施工所关心的核心问题之一。 对于重大工程而言，确定岩体力学参数最直接的方法是现场原位试验，但现场原位试验成本高、施工难度大，也不能完全避免尺度效应的影响。大量工程经验表明，小块岩石的强度与岩体强度之间存在某种内在关系。因此，在实际工程中总是试图利用小块岩石室内实验结果以及现场调查的资料来估算大范围的岩体力学参数，并通过多种估算方法对比研究得出与工程实际相吻合的参数计算值。 本文以贵州省鱼简河水库坝基出露的松子坎和狮子山组地层为例，分析岩体的结构特征和力学性质，对岩体变形及强度参数的选取分别进行了深入探讨。作者认为，岩体是岩块和结构面组成的复杂的结构体，其结构及结构面的力学性质是影响岩体力学性质的重要因素。本文深入分析了区内各岩体结构面的结构特征及力学性质，结合岩块的物理力学性质的试验结果，采用分形和神经网络在内的多种方法估算了坝基岩体的变形参数。通过对比分析各种计算方法所确定的力学参数，结合岩体所处的力学状态，为水库工程建设提供了合理的变形参数及强度参数建议值，并得出如下研究成果： 1．通过分析岩体结构面的结构特征及力学性质，结合岩块力学性质试验研究来评价岩体力学性质，获取力学参数是可行的。 2．对结构面野外实测数据进行统计分析，运用蒙特卡洛方法进行模拟，模拟岩体结构特征，为岩体力学参数估算提供重要依据。 3．通过岩体结构面网络模拟可实现岩体质量评价中的RQD值分析，以此为主要依据，应刚恰当的分类方法，可得到比较可信的岩体质量评价结果。 4．对于结构面较发育的裂隙化岩体，其单轴抗压强度、单轴抗拉强度用Hoek-Brown经验公式对比其他的许多经验公式法，估算结果与工程实践较接近。 5．在对比多种力学参数估算方法时，RMR法估算结果与Q方法估算结果普遍偏高，约为参数建议值的约为2～3倍，此结论在许多工程中都被得到证实。因此，在应用这两种方法进行岩体参数估算时，应对参数进行合理折减。 本论文的创新之处主要体现在以下3个方面： 1．统计表明，岩体结构展布特征具有极强的分形特性。将分形统计模型理论应用于岩体损伤力学研究中，并结合结构面网络模拟结果计算岩体分维数(D)可以比较合理估算岩体变形模量低二件一D)E)。说明分形方法是一种估算岩体参数的有效手段. 2.利用人工神经网络理论强大的大规模自适应、自学习、白组织并行处理能力，引入人工神经网络方法尝试一种实用、可靠的岩体参数估算新途径是十分必要的。本论文研究表明其预测精度很高，与工程设计的参数建议值较为吻合。 3.论文中提及的各种岩体参数估算方法都有其齐白优点和局限性，本论文结合实际工程—贵阳市息烽县鱼简河水利工程岩体所处的力学环境，本文提出将多种方法应用于坝基岩体力学参数估算中，对比各种计算结果，综合选取岩体力学参数，并指出综合选择岩体力学参数是今后工程建设中获取合理力学参数的一种趋势。关键词:坝基岩体，岩体力学参数，松子坎组，狮子山组，结构面，岩体质且，蒙特卡洛模拟，分维数，人工神经元网络更多还原

【Abstract】 Hydraulic engineering can artificially control the resources, environment and ecology system of the drainage area. With the development of economy, society and science and technology, the ability of enduring and resisting disasters is becoming much stronger, and the situation that people use water resources has effectively changed by dams.However, after dams have been constructed and operated, they also have become a kind of potential and huge threaten, and their security and service performance impact greatly the society and economy. The parameters estimation of rock mass in the study region is one of the hot topic in rock mass mechanics. Different mechanical parameters leads to different results, and improper mechanics parameters even incorrectly instructs to project design and construction. Many project design defaults derived from the deviation of the estimated parameters. The determination of mechanics parameters of rock mass is an important problem during the investigating, designing and constructing of projects because its rationality is the main determinant of the security and economy of the hydraulic engineering.For the important projects, the in-situ test is a direct method to determine the mechanics parameters of rock mass. Generally, it is very expensive and difficult to construct, and it is also affected by size effect. A large number of project experiences show that the internal relation exists between the small rock strength and the rock mass strength. So we always try to estimate the mechanical parameters of rock mass by analyzing the experimental data of mechanics parameters of rock samples and the geological data based on the field investigation, and acquire the parameters which are according with the practice by the contrast of multi-methods.Taken SongZiKan group and ShiZiShan group strata appeared in the dam foundation of YuJianHe reservoir in GuiZhou province for example, the structures characters and mechanics properties of rock mass were analyzed in this paper, and the methods of choosing the deformation and strength parameters of rock mass were discussed too. The rock is made up of the rock mass and the structural plane, so the rock mass structure and the mechanics properties of the structural plane are important factors of influencing the rock mechanics properties. So the structure characteristics and mechanics properties of structural plane are studied in detail in this paper. Based on the test results of the physical mechanics properties of the rock, the rock mechanical properties are analyzed, and the deformation parameters of rock masses were estimated by many methods including the fractal method and neural network method. Through comparison of the advantages and disadvantages of these methods, considering the engineering geology of the rock mass, the deformation parameters and strength parameters of rock masses can be selected comprehensively. Several study achievements have been obtained, which are described as follows:1. It is feasible to evaluate the rock mechanics properties and obtain the mechanical parameters of rock mass by analyzing the structure characteristics and mechanics properties of discontinuities, combining with experimental study of mechanical properties of rock.2. The characteristics of rock mass structures can be analyzed by statistical analysis of geometrical characteristics of discontinuities and Monte-Carlo simulation, which can also offer many important basis for estimating mechanical parameters of rock mass.3. The analysis of RQD can be realized through rock discontinuity network modeling and the credible result of rock mass quality evaluation can be obtained through proper classification methods.4 . For the fractured rock which has many discontinuities, Hoek-Brown experience formula is more proper to calculate the uniaxial compressive strength and uniaxial tensive strength comparing to other formulas and the result is more close to engineering facts.5. For methods used to evaluate mechanics parameters, we find that the values calculated by RMR method and Q m更多还原