基于DILSSVM的岩石强度预测研究

为解决经典强度准则预测岩石破坏强度不准确的问题,本文使用交叉最小二乘支持向量机(DILSSVM)在单轴及三轴加载情况下对岩石强度进行预测。将现场和实验室试验得到的各种岩石相关数据划分成训练和验证子集。将抗压强度c和最小主应力3作为输入值,最大主应力值1f作为输出值训练DILSSVM。使用训练后的DILSSVM预测岩石破坏时的1f,同时使用训练子集反演Hoek-Brown经验公式的常数m。使用测试数据对训练后DILSSVM预测的目标岩石强度的准确性进行验证,同时将测试数据带入2类Hoek-Brown(m值的取值不同)经验公式预测目标岩石强度1f。比较DILSSVM和Hoek-Brown预测结果, 表明:DILSSVM预测结果的均方差减小了45% ～55%,决定系数增加了0.055～0.085,更接近于1。说明DILSSVM对岩石强度的预测范围较宽,且适合岩石种类多变的复杂非线性情况。     

砂岩试验强度与强度准则预测结果对比分析

对山西大同煤峪口矿巷道砂岩进行了巴西劈裂、单轴和常规三轴试验,分析了其变形和强度特性,以整体平方根误差最小为依据拟合确定了5种强度准则参数,比较各准则的整体拟合效果以及抗拉强度和抗压强度预测值与试验值之间的差异。结果表明:砂岩单轴和低围压下的脆性特征显著,随着围压增大屈服段明显增加,逐渐由脆性向延性转变;对试验数据的整体拟合效果,Mohr-Coulomb准则最不理想,指数强度准则最好,Hoek-Brown准则和Bieniawski准则相接近;对抗拉强度的预测,Mohr-Coulomb准则明显高估了砂岩的抗拉强度,不适合预测抗拉强度,指数强度准则最为理想,Bieniawski准则不能反映岩石抗拉特性;单轴和不同围压下砂岩强度,直线型Mohr-Coulomb准则偏离了大多数试验数据点,特别是高围压下偏离显著结果失真,Hoek-Brown准则和Bieniawski准则预测结果几乎一致,Sheorey准则高围压下预测结果稍有偏高,指数强度准则接近大多数试验数据,预测效果最佳。      

三轴试验抗剪强度参数值回归分析法的区别与修正

对岩土三轴试验数据采用两种线性回归法计算所得的抗剪强度参数c、φ值会有差别。为了揭示这两种回归方法差别的根本原因,深入最小二乘法线性回归的基本原理指出,按大小主应力回归的结果是最佳估值,而按p-q曲线回归的计算结果不可靠,为了得到统一的结果,对后者所基于的最小二乘法进行了修正;推导了修正后的回归系数计算公式,证明了修正后两种方法得到的结果相同,并用一组千枚岩三轴试验数据进行了验证。结果表明,修正后用这两种方法确定三轴试验抗剪强度参数值都将得到同样的最佳无偏估值。     

岩体抗剪强度参数的线性优化方法

针对岩体结构面力学参数的重要性,以大量工程实例为依据,以统计误差分析相关理论为基础,研究最小二乘法与M-估计法在分析结构面抗剪强度参数时的适用性。研究结果表明,在难以判断粗差的数量和比例时,很难对两种线性回归方法的优劣进行规律性的比较。考虑到数据是具有工程特性的,为更准确地处理实验数据,在以上两种方法的基础上提出了一种新的实验数据处理方法,并应用于计算李家河水库的抗剪强度参数。     

高陡山区房柱法开采岩体移动预测及山体失稳判据

在实测资料统计分析基础上,以概率理论为基础,建立了高陡山区房柱法地下开采岩体移动变形预测分析模型,并给出了岩体变形破坏的极限值。利用本模型对山区地下开采工程实例进行了具体的计算分析,通过工程实例分析说明,理论预测结果与现场实测结果吻合。     

岩体结构面强度的 Mohr—Coulomb破坏判据与 Barton破坏判据的试验对比

岩体结构面强度是工程岩体稳定性设计中的重要参数,目前工程上常用的两个强度判据,即Mohr-Coulomb准则和Barton准则有各自适用条件。本文在对七组具不同岩性、不同粗糙度及起伏度红层样品进行试验的基础上,探讨了在试验过程中所揭示的强度周期性变化现象,同时基于两准则试验数据,分析了它们之间的区别,获得一些重要结论.即JRC、JCS越大,两准则差别越大;在低法向应力下两准则较接近,在高法向应力时,MC准则偏于危险。     

岩石点荷载强度与单轴抗压强度线性关系试验

针对目前岩石点荷载试验大部分采用规则试件进行,试件加工制作及试验的过程繁琐、成本高的问题,通过自制点荷载试验仪以及RMT-150B岩石力学伺服试验机,对煤矿顶底板常见岩性进行不规则岩石试件点荷载试验和规则岩石试件单轴抗压试验。试验结果表明:不规则岩石点荷载强度基本上满足正态分布规律,同时,点荷载破坏载荷与破坏面积之间呈线性关系,不同载荷破坏面积对点荷载试验中的尺寸、形状效应基本上没有太大的影响。通过分析点荷载强度与单轴抗压强度两者间的关系,得出通过点荷载强度确定单轴抗压强度的经验公式。      

含两组交叉贯通节理岩体的强度及破坏特征分析

天然岩体中常含有多组相互交叉的贯通节理,它们的存在极大地削弱了岩体的力学性质。为了研究含两组交叉贯通节理岩体的强度及破坏特征,基于弹塑性数值流形方法,对不同应力状态下,节理倾角和节理间距不同的岩体压缩试验进行数值模拟。结果表明：岩体强度随节理倾角的变化曲线呈现出多波峰、多波谷特点,岩体强度随节理间距变化曲线符合负指数函数形式。根据节理状态,两组节理岩体的破坏模式可以分为3种：岩块破坏、沿一组节理滑移和沿两组节理滑移。两组节理均会影响岩体强度,一组起主要作用,另一组起次要作用,并且节理组之间存在相互影响。通过对数值计算结果进行回归分析,基于一组节理岩体强度预测模型,量化节理组之间的相互影响,提出适用于含两组交叉贯通节理岩体的强度预测模型。模型形式简单,使用方便,可为实际工程中正确评估岩体强度提供指导性意见。     

Influence of Rock Mass Parameters on the Performance of a TBM in a Gneissic Formation (Varzo Tunnel)

Summary The paper analyses the influence of rock mass quality on the performance of a double shield TBM in the excavation of a tunnel in a gneiss formation which is characterized by high strength and low fracture intensity.As full observation of the rock conditions was prevented by the use of segmental lining, a geomechanical survey of the face was performed during maintenance downtime and the observed conditions were correlated with the machine performance parameters for that same day. A statistical analysis of the data shows that penetration rate correlates well with a slightly modified RMR index (in which the influence of the water conditions and joint orientation were discounted), but the most important factor is by far the partial rating of the RMR classification related to joint spacing only. However in tunnels characterized by greater variability in rock strength and joint conditions, it could be worthwhile using the complete RMR index.Given the toughness of the rock, failure of the cutter bearings and supports were a frequent occurrence during excavation. Owing to this factor the influence of rock quality on the rate of advance was found to be weak and the correlation more scattered.The results obtained for the Varzo tunnel were compared with those relative to other tunnels in granitic rocks and found to be in good agreement. However the relationships obtained should be considered valid only for this type of rock; machine behaviour could be found to be markedly different in other rock types, even where rock material strength and joint frequency are the same.     

A particle mechanics approach for the dynamic strength model of the jointed rock mass considering the joint orientation

In nature, there exist several forms of anisotropy in rock masses due to the presence of bedding planes, joints, and weak layers. It is well understood that the anisotropic properties of jointed rock masses significantly affect the stability of surface and underground excavations. However, these critical anisotropic characteristics are often ignored in existing uniaxial dynamic failure criteria. This study investigates the effect of a pre-existing persistent joint on the rate-dependent mechanical behaviours of a rock mass using a particle mechanics approach, namely, bonded particle model (BPM), to realistically replicate the mechanical response of the rock mass. Firstly, in order to capture the rate-dependent response of the jointed rock mass, the BPM model is validated using published experimental data. Then, a dynamic strength model is proposed based on the Jaeger criterion and simulation results. To further investigate the dynamic behaviours, the dynamic uniaxial compressive strength (UCS) for anisotropic rock masses with various joint orientations is investigated by subjecting the BPM models to uniaxial compression numerical tests with various strain rate. The proposed dynamic strength model is validated based on numerical simulation results. Finally, the fragmentation characteristics of the jointed rock masses are analysed, which demonstrate that the failure mode affects the dynamic UCS. This is further confirmed by the analysis of the orientations of microscopic cracks generated by the compression loading.     

基于人工神经网络模型的岩石特性预测

近年来,软计算技术被用作替代的统计工具。如人工神经网络（ANN）被用于开发预测模型来估计所需的参数。在本研究中,通过利用冲击钻进过程中的一些钻进参数（气压、推力、钻头直径、穿透率）和所产生的声级,建立了预测岩石性质的神经网络模型。在实验室中所产生的数据,用于开发预测岩石特性（如单轴抗压强度、耐磨性、抗拉强度和施密特回弹数）的神经网络模型,并使用各种预测性能指标对所建模型进行检验,结果表明人工神经网络模型适用于岩石性质的预测。     

Characterization of Strength of Intact Brittle Rock Considering Confinement-Dependent Failure Processes

As technologies for deep underground development such as tunneling underneath mountains or mass mining at great depths (>1,000 m) are implemented, more difficult ground conditions in highly stressed environments are encountered. Moreover, the anticipated stress level at these depths easily exceeds the loading capacity of laboratory testing, so it is difficult to properly characterize what the rock behavior would be under high confinement stress conditions. If rock is expected to fail in a brittle manner, behavior changes associated with the relatively low tensile strength, such as transition from splitting to the shear failure, have to be considered and reflected in the adopted failure criteria. Rock failure in tension takes place at low confinement around excavations due to tensile or extensional failure in heterogeneous rocks. The prospect of tensile-dominant brittle failure diminishes as the confinement increases away from the excavation boundary. Therefore, it must be expected that the transition in the failure mechanism, from tensile to shear, occurs as the confinement level increases and conditions for extensional failure are prevented or strongly diminished. However, conventional failure criteria implicitly consider only the shear failure mechanism (i.e., failure envelopes touching Mohr stress circles), and thus, do not explicitly capture the transition of failure modes from tensile to shear associated with confinement change. This paper examines the methodologies for intact rock strength determination as the basic input data for engineering design of deep excavations. It is demonstrated that published laboratory test data can be reinterpreted and better characterized using an s-shaped failure criterion highlighting the transition of failure modes in brittle failing rock. As a consequence of the bi-modal nature of the failure envelope, intact rock strength data are often misinterpreted. If the intact rock strength is estimated by standard procedures from unconfined compression tests (UCS) alone, the confined strength may be underestimated by as much as 50 % (on average). If triaxial data with a limited confinement range (e.g., σ₃ ? 0.5 UCS due to cell pressure limitations) are used, the confined strength may be overestimated. Therefore, the application of standard data fitting procedures, without consideration of confinement-dependent failure mechanisms, may lead to erroneous intact rock strength parameters when applied to brittle rocks, and consequently, by extrapolation, to correspondingly erroneous rock mass strength parameters. It follows that the strength characteristics of massive rock differ significantly in the direct vicinity of excavation from that which is remote with higher confinement. Therefore, it is recommended to adopt a differentiated approach to obtain intact rock strength parameters for engineering problems at lower confinement (near excavation; e.g., excavation stability assessment or support design), and at elevated confinement (typically, when the confinement exceeds about 10 % of the UCS) as might be encountered in wide pillar cores.     

The influence of the intermediate principal stress on rock failure behaviour: A numerical study

The influence of the intermediate principal stress on rock strength has been studied comprehensively by previous researchers. However, the reason why rock strength firstly increases and subsequently decreases with the increase of intermediate principal stress is still unclear. In this paper, the mechanism of the intermediate principal stress effect on rock failure behaviour is revealed through a numerical method using the EPCA3D system (Elasto-Plastic Cellular Automaton). In this study, both homogeneous and heterogeneous rocks are considered. The heterogeneity of a rock specimen is modelled by introducing Weibull's statistical distribution. Two criteria, i.e. the Drucker–Prager and Mohr–Coulomb models, are used to determine whether a meso-scopic element in the rock specimen is in a failure state or not during the polyaxial stress loading process. The EPCA3D simulation reproduces the typical phenomenon of the intermediate principal stress effect that occurs in some rock experiments. By studying the EPCA3D simulated acoustic emission and complete stress–strain curves illustrating failure initiation, propagation and coalescence in the failure process of rocks, the essence of the intermediate principal stress effect is tracked. It is concluded that the heterogeneous stress distribution induced by the natural heterogeneity of rocks and the effect of the loading platen are two of the reasons producing the intermediate stress effect. Studies indicate that a moderate intermediate principal stress delays the onset of local failure, which in turn leads to an increase in the rock strength. However, once the intermediate principal stress reaches a certain value, local failure will be formed through the application of the intermediate principal stress. It is the number of failed elements in the pre-peak region that determines whether the rock strength decreases or not. The extent of rock strength reduction when the intermediate principal stress reaches a certain value is lessened with the increase in the minimum principal stress.     

Artificial neural network model for prediction of rock properties from sound level produced during drilling

In many rock engineering applications such as foundations, slopes and tunnels, the intact rock properties are not actually determined by laboratory tests, due to the requirements of high quality core samples and sophisticated test equipments. Thus, predicting the rock properties by using empirical equations has been an attractive research topic relating to rock engineering practice for many years. Soft computing techniques are now being used as alternative statistical tools. In this study, artificial neural network models were developed to predict the rock properties of the intact rock, by using sound level produced during rock drilling. A database of 832 datasets, including drill bit diameter, drill bit speed, penetration rate of the drill bit and equivalent sound level (L_eq) produced during drilling for input parameters, and uniaxial compressive strength (UCS), Schmidt rebound number (SRN), dry density (ρ), P-wave velocity (V_p), tensile strength (TS), modulus of elasticity (E) and percentage porosity (n) of intact rock for output, was established. The constructed models were checked using various prediction performance indices. Goodness of the fit measures revealed that recommended ANN model fitted the data as accurately as experimental results, indicating the usefulness of artificial neural networks in predicting rock properties.     

The Influence of Testing Procedures on Uniaxial Compressive Strength Prediction of Carbonate Rocks from Equotip Hardness Tester (EHT) and Proposal of a New Testing Methodology: Hybrid Dynamic Hardness (HDH)

The Equotip hardness tester (EHT) is a portable and non-destructive instrument used mainly for the dynamic rebound hardness testing of metals. Although various versions of the ‘single impacts’ and ‘repeated impacts’ testing procedures have been employed by different authors for different applications, it is not yet known whether a particular testing procedure is more relevant for a specific application in rock engineering. To be able to contribute to the subject, the present study was carried out to determine the suitability of different rebound testing procedures with this instrument for uniaxial compressive strength (UCS) estimations of some selected carbonate rocks. To achieve this goal, as well as four different existing rebound testing procedures, a newly proposed testing methodology involving the parameter hybrid dynamic hardness (HDH) was also employed. The statistical analyses performed on the experimental data, on the whole, showed that the test procedures which are based on single impacts test procedures outperformed the repeated impacts test procedures in terms of UCS prediction accuracy. The prediction capability of the newly introduced testing methodology was found to be superior to those of other procedures considered in this work, suggesting that it could be an efficient tool in practice for preliminary estimates of rock strength. The statistical analyses also indicated that, in practical applications of the EHT using different test procedures, it may be possible to predict the UCS more accurately when apparent density data is available. For the range of specimen sizes considered, no clear evidence of size effect was observed in the mean rebound values. The argument raised by some other authors that the EHT might not be a convenient instrument for the dynamic rebound hardness determination of relatively high-porosity rocks was not confirmed in this study.     

岩爆预测的灰色关联系统分析与评价

岩爆由于其发生的突然性和猛烈性,常给地下工程建设带来灾害性破坏,因此,岩爆预测研究具有重大的工程意义.岩爆由于影响因素众多且关系复杂,采用传统的指标和评价办法很难系统地考虑诸多因素,而采用灰色关联分析方法可建立起各种影响因素的相关性.灰色关联系统通过分析洞室围岩应力、岩性条件和能量条件等影响岩爆的主要因素,选取岩石抗压...     

基于PFC2D数值试验的异性结构面剪切强度特性研究

天然岩体中广泛发育两侧岩性不同的异性结构面,开展异性结构面变形和强度特性研究旨在为岩体稳定性评价和利用提供依据。选取三峡库区侏罗系典型的砂岩-泥岩异性岩层,首先运用分形几何理论,定量计算了平直和4种不同不规则起伏形态结构面的粗糙度系数JRC值,然后基于PFC2D颗粒流程序,分别开展了以上5种形态异性结构面的数值剪切试验,获得了各形态结构面在不同正应力下的剪切应力-位移曲线。根据数值试验结果,采用巴顿的JRC-JCS模型分析了异性结构面强度特性,并与同性结构面强度性质进行对比研究。最后,在考虑异性结构面剪切破坏机制的基础上,引入强度因子的概念,提出了新的适用于异性结构面强度评价的两类改进巴顿准则。研究结果表明：异性岩体结构面抗剪强度介于相同粗糙度的两种同性结构面强度之间,在较低正应力下接近软岩同性结构面强度,符合Ⅰ型改进巴顿准则;在较高正应力下偏向硬岩同性结构面强度,符合Ⅱ型改进巴顿准则。实际工程中可利用改进准则并根据异性结构面应力状态对岩体稳定性进行评价,弥补了以往研究的不足。     

Prediction of underground cavity roof collapse using the Hoek–Brown failure criterion

Preventing roof collapse in underground cavities is a challenge to geotechnical engineering. In this study, three independent methods have been used to evaluate the roof collapse of underground rectangular cavities for a range of geometries and rock properties. The rock mass strength has been described by the Hoek–Brown failure criterion. The results of the analysis allow for prediction of roof collapse and to determine whether the failure surface that develops in the rock mass remains localised or extends through the full depth of cover. This is of significance if there are overlying cavities and when estimating surface subsidence.     

基于遍布节理模型的边坡稳定性强度折减法分析

采用非线性数值分析方法分析边坡稳定性问题时,强度折减法因其具有较多的优点而得到广泛应用。岩土体一般采用理想弹塑性模型,屈服准则为广义米赛斯准则。对于密集节理岩质边坡稳定性问题,采用遍布节理模型可同时考虑岩块和节理属性,更符合岩体状态及工程实际,认为岩体经强度折减后潜在破坏可能首先出现在岩体中或沿节理面或二者同时破坏。结合工程实例,基于遍布节理模型的强度折减法计算结果表明,潜在滑移面为折线型滑面,下部潜在滑移面倾角与节理面等效内摩擦角基本一致,上部潜在滑移面与岩体拉破坏相关;节理倾角与边坡安全系数、潜在滑动范围密切相关,陡倾角节理对边坡稳定性影响较小。通过对边坡失稳判据和边坡滑移面确定的探讨,认为以力或位移不收敛作为边坡失稳判据是适当的,而边坡的剪应变速率物理意义十分明确,适于作为边坡潜在滑移面的确定依据。     

Analysis of strength and moduli of jointed rocks

This paper deals with two aspects of jointed rock mass behavior, first the finite element modeling of a jointed rock mass as an equivalent continuum, second the comparison of empirical strength criteria of a jointed rock mass. In finite element modeling the jointed rock properties are represented by a set of empirical relationships, which express the properties of the jointed medium as a function of joint factor and the properties of the intact rock. These relationships have been derived from a large set of experimental data of tangent elastic modulus. It is concluded that equivalent continuum analysis gives the best results for both single and multiple jointed rock. The reliability of the analysis depends on the estimation of joint factor, which is a function of the joint orientation, joint frequency and joint strength.Empirical strength criteria for jointed rocks, namely Hoek and Brown, Yudhbir et al., Ramamurthy and Arora, Mohr–Coulomb have been incorporated in a nonlinear finite element analysis of jointed rock using the equivalent continuum approach, to determine the failure stress. The major principal stress at failure, obtained using Ramamurthy's criteria, compares very well with experimental results.