基于粗糙集和人工神经网络的洞室岩体质量评价

针对洞室岩体质量问题,从洞室工程的角度选取能够反映岩体综合工程特性的6个参数,用可拓评判和专家审定的方法构建了决策样本集;再利用粗糙集理论对原始决策样本集进行约简操作,并分析各指标对决策的相对重要性;最后将约简结果生成的规则作为人工神经网络的输入,建立了洞室岩体质量评价模型。通过工程实例分析对比,该模型有效地简化神经网络的网络结构,减少网络的训练步数,提高网络的学习效率,能够较准确地反映洞室岩体的工程特性。     

An Integrated Numerical Modelling–Discrete Fracture Network Approach Applied to the Characterisation of Rock Mass Strength of Naturally Fractured Pillars

Naturally fractured mine pillars provide an excellent example of the importance of accurately determining rock mass strength. Failure in slender pillars is predominantly controlled by naturally occurring discontinuities, their influence diminishing with increasing pillar width, with wider pillars failing through a combination of brittle and shearing processes. To accurately simulate this behaviour by numerical modelling, the current analysis incorporates a more realistic representation of the mechanical behaviour of discrete fracture systems. This involves realistic simulation and representation of fracture networks, either as individual entities or as a collective system of fracture sets, or a combination of both. By using an integrated finite element/discrete element–discrete fracture network approach it is possible to study the failure of rock masses in tension and compression, along both existing pre-existing fractures and through intact rock bridges, and incorporating complex kinematic mechanisms. The proposed modelling approach fully captures the anisotropic and inhomogeneous effects of natural jointing and is considered to be more realistic than methods relying solely on continuum or discontinuum representation. The paper concludes with a discussion on the development of synthetic rock mass properties, with the intention of providing a more robust link between rock mass strength and rock mass classification systems.     

The Contribution of Landslide Susceptibility Factors Through the Use of Rock Engineering System (RES) to the Prognosis of Slope Failures: An Application in Panagopoula and Malakasa Landslide Areas in Greece

This paper presents an application of the rock engineering system (RES) in an attempt to assess the proper landslide parameters and estimate the instability index, using two disastrous landslides in Greece which took place in Panagopoula (1971) and Malakasa (1995). RES has been developed by Hudson (Rock engineering systems: theory and practice. Ellis Horwood Limited, 1992) to determine interaction of a number of parameters in rock engineering design and calculate instability index for rock slopes. In this paper, an attempt is made to prove, how RES can be implemented in large-scale instability areas where natural slopes are associated with a variety of geomaterials (soils, rocks, weathering mantle, etc.), by selecting each time the most appropriate parameters that are relevant to the ad hoc potential slope failure and which can be quantified easiest than those of time and money consuming ones. RES approach allows the utilization of those parameters which are particularly active at the site, evaluates the importance of their interactions, taking into account the particular problems at any investigated site. The instability index for both study areas were calculated and found 89.47 for Panagopoula site and 81.59 for Malakasa (out of 100). According to the classification for landslide susceptibility by Brabb et al. (Landslide susceptibility in San Mateo County, California, 1972), both the examined case studies are classified as landslides, approving their existence as two serious slope failures. Thus, RES could be a simple and efficient tool in calculating the instability index and consequently in getting the prognosis of a potential slope failure in landslide susceptible areas, for land use and development planning processes.     

基于RES的地下巷道工程稳定性全耦合分析方法

针对地下巷道工程稳定性问题,基于岩石工程系统(RES)理论,提出了全耦合分析法及优化设计思想.首先简要地叙述了全耦合分析过程的思想及流程,并在收集的27个样本的基础上,应用神经网络编码法建立了地下巷道工程的综合交互作用矩阵(GIM),分析了主要因素影响和作用机制.根据本巷道工程系统的稳定性能分析结果,提出了稳定性评价参数,给出了稳定状态范围.针对沙曲煤矿顺槽进行了优化支护设计举例,并用Q系统岩体质量评价法对其进行对比分析及验证.现场监测及对比分析结果表明,提出的全耦合分析法对稳定性分析及支护设计方法有实用价值.     

Development of a New Index to Assess the Rock Mass Drillability

Knowledge of drillability of rock masses in engineering projects is very important in determining drilling costs. In drilling operations, so many parameters such as the properties of rock and the drilling equipment affect the drilling performance. In this study, after discussing the rock mass drillability process and identifying all the effective parameters, interaction matrixes based on the rock engineering systems, that analyze the interrelationship between the parameters affecting rock engineering activities, is introduced to study the rock mass drillability tribosystem. Given that interaction matrix codes are not unique numbers, and then possible interactive intensities are calculated for each matrix and a group decision-making method, Fuzzy–Delphi–AHP technique has been used to obtain appropriate weights. As a result, rock mass drillability index (RMDI) is presented to classify the rock mass drillability. The results indicate the ability of this method to analyze rock mass drillability procedure. Drilling data along with laboratory rock properties from Sungun copper mine were collected and were ranked according to the new classification system. Fifteen zones at the mine site were ranked based upon the new index RMDI and a reasonable correlation was obtained between measured drilling rate at the zones and RMDI data.     

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.     

GIS数据融入遥感图像理解的模型初探

作者在本文中讨论了利用电磁波散射的数值计算、高光谱、人人智能进行遥感图像理解的现状，说明了地理信息系统（GIS）数据融入遥感图像理解的重要性，概述了GIS数据融入所面临的问题，指出了GIS融入遥感图像理解对处理系统的一般要求，分析了基于数据挖掘的专家系统、神经网络、进行计算的特点和性能。结合神经网络和进化计算能融合多源数据、高度并行、自适应、自组织能力和知识处理的能力，以及进化计算通过重组、变异和复制算法具有优化选择的功能，构建了基于进化计算的神经网络的GIS数据，并融入模型。模型中GIS属性数据惯穿遥感图像理解的全过程，是一种高层次的数据融合，且GIS数据特征的提取、处理和遥感图像理解是高度并行的。它是一种实现GIS数据融入遥感图像理解的有效途径。基于数据挖掘的专家系统也具有上述功能，但存在数据挖掘的困难。     

基岩裂隙水神经网络专家系统实例研究

详细介绍了基岩裂隙水寻找与开发的神经网络专家系统的建立,包括神经网络专家系统的组成、结构和神经网络模型的训练,并将训练后的神经网络模型应用于某裂隙水水源地的勘探工作中,取得了较好的效果.     

Updating the models and uncertainty of mechanical parameters for rock tunnels using Bayesian inference

Rock mechanical parameters and their uncertainties are critical to rock stability analysis, engineering design, and safe construction in rock mechanics and engineering. The back analysis is widely adopted in rock engineering to determine the mechanical parameters of the surrounding rock mass, but this does not consider the uncertainty. This problem is addressed here by the proposed approach by developing a system of Bayesian inferences for updating mechanical parameters and their statistical properties using monitored field data, then integrating the monitored data, prior knowledge of geotechnical parameters,and a mechanical model of a rock tunnel using Markov chain Monte Carlo(MCMC) simulation. The proposed approach is illustrated by a circular tunnel with an analytical solution, which was then applied to an experimental tunnel in Goupitan Hydropower Station, China. The mechanical properties and strength parameters of the surrounding rock mass were modeled as random variables. The displacement was predicted with the aid of the parameters updated by Bayesian inferences and agreed closely with monitored displacements. It indicates that Bayesian inferences combined the monitored data into the tunnel model to update its parameters dynamically. Further study indicated that the performance of Bayesian inferences is improved greatly by regularly supplementing field monitoring data. Bayesian inference is a significant and new approach for determining the mechanical parameters of the surrounding rock mass in a tunnel model and contributes to safe construction in rock engineering.     

A novel type of neural networks for feature engineering of geological data:Case studies of coal and gas hydrate-bearing sediments

The nature of the measured data varies among different disciplines of geosciences.In rock engineering,features of data play a leading role in determining the feasible methods of its proper manipulation.The present study focuses on resolving one of the major deficiencies of conventional neural networks(NNs) in dealing with rock engineering data.Herein,since the samples are obtained from hundreds of meters below the surface with the utmost difficulty,the number of samples is always limited.Meanwhile,the experimental analysis of these samples may result in many repetitive values and 0 s.However,conventional neural networks are incapable of making robust models in the presence of such data.On the other hand,these networks strongly depend on the initial weights and bias values for making reliable predictions.With this in mind,the current research introduces a novel kind of neural network processing framework for the geological that does not suffer from the limitations of the conventional NNs.The introduced single-data-based feature engineering network extracts all the information wrapped in every single data point without being affected by the other points.This method,being completely different from the conventional NNs,re-arranges all the basic elements of the neuron model into a new structure.Therefore,its mathematical calculations were performed from the very beginning.Moreover,the corresponding programming codes were developed in MATLAB and Python since they could not be found in any common programming software at the time being.This new kind of network was first evaluated through computer-based simulations of rock cracks in the 3 DEC environment.After the model's reliability was confirmed,it was adopted in two case studies for estimating respectively tensile strength and shear strength of real rock samples.These samples were coal core samples from the Southern Qinshui Basin of China,and gas hydrate-bearing sediment(GHBS) samples from the Nankai Trough of Japan.The coal samples used in the experiments underwent nuclear magnetic resonance(NMR) measurements,and Scanning Electron Microscopy(SEM) imaging to investigate their original micro and macro fractures.Once done with these experiments,measurement of the rock mechanical properties,including tensile strength,was performed using a rock mechanical test system.However,the shear strength of GHBS samples was acquired through triaxial and direct shear tests.According to the obtained result,the new network structure outperformed the conventional neural networks in both cases of simulation-based and case study estimations of the tensile and shear strength.Even though the proposed approach of the current study originally aimed at resolving the issue of having a limited dataset,its unique properties would also be applied to larger datasets from other subsurface measurements.     

The introduction of neural network system and its applications in rock engineering

The neural network system has been developing very fast recently. It has been widely used in many industries such as automation, nuclear power plant, chemical industry, etc. Neural network systems have a great advantage in dealing with problems in which many factors influence the process and result, and the understanding of this process is poor, and there are experimental data or field data. In rock engineering, many problems are of this nature. In this paper, a brief introduction to neural network systems is given. Problems such as what is a neural network, how it works and what kind of advantages it has are discussed. After this, several applications in rock engineering, made by us, are presented. Case 1 is ore boundary delineation. In this case, the rock are divided into three classes, i.e.: (1) waste rock; (2) semi-ore; and (3) ore for mining purposes. The neural network system built can judge whether it is ore, semi-ore or waste rock along the borehole according its corresponding geophysical logging data, such as gamma-ray, gamma-gamma, neutron and resistivity. Case 2 is aggregate quality prediction. In this case, the quality parameters: (1) impact value; (2) abrasion value I; and (3) abrasion value II are predicted by using a neural network system based on density, point load, content of quarts and content of brittle minerals. Case 3 is rock indentation depth prediction. In this case, the rock indentation depth under indentation load is predicted by the established neural network system based on the indentation load on rock, indenter type and rock mechanical properties, such as uniaxial compressive strength, Young's modulus. Poisson's ratio, critical energy release rate and density. In all these cases, the neural network systems have been applied successfully. The testing results are satisfactory and better than the existing techniques.     

Rock mass characterization for the underground cavern design of Khiritharn pumped storage scheme

A general approach to rock engineering designing aspects adopted at the Khiritharn Pumped Storage Scheme is described. The scheme involves excavation of three large caverns and tunnels in jointed sandstone within a suture zone in Southeast Thailand. Geological condition and engineering properties of the sandstone were investigated. Strength and modulus properties of the intact rock were determined from laboratory tests and properties of rock mass were empirically estimated for the design analysis in the de.nite study stage on the basis of three rock mass classi.cation systems namely the Rock Mass Rating (RMR), Geological Strength Index (GSI) and a Japanese system (EPDC). While the GSI gives strength and modulus of deformation values slightly higher than the RMR classi.cation, the EPDC gives a lower value of modulus of deformation but comparable rock mass strength value for the level of con.ning pressures at the depth of the cavern excavation. The results of stress analysis and loosening wedge analysis for the cavern excavations suggest favorable excavation condition.     

Prediction of the Rock Mass Diggability Index by Using Fuzzy Clustering-Based,ANN and Multiple Regression Methods

Rock mass classification systems are one of the most common ways of determining rock mass excavatability and related equipment assessment. However, the strength and weak points of such rating-based classifications have always been questionable. Such classification systems assign quantifiable values to predefined classified geotechnical parameters of rock mass. This causes particular ambiguities, leading to the misuse of such classifications in practical applications. Recently, intelligence system approaches such as artificial neural networks (ANNs) and neuro-fuzzy methods, along with multiple regression models, have been used successfully to overcome such uncertainties. The purpose of the present study is the construction of several models by using an adaptive neuro-fuzzy inference system (ANFIS) method with two data clustering approaches, including fuzzy c-means (FCM) clustering and subtractive clustering, an ANN and non-linear multiple regression to estimate the basic rock mass diggability index. A set of data from several case studies was used to obtain the real rock mass diggability index and compared to the predicted values by the constructed models. In conclusion, it was observed that ANFIS based on the FCM model shows higher accuracy and correlation with actual data compared to that of the ANN and multiple regression. As a result, one can use the assimilation of ANNs with fuzzy clustering-based models to construct such rigorous predictor tools.     

Risk assessment and prediction of safety factor for circular failure slope using rock engineering systems

Circular failure is generally observed in the slope of soil, highly jointed rock mass, mine dump and weak rock. Accurate estimation of the safety factor (SF) of slopes and their performance is not an easy task. In this research, based on rock engineering systems (RES), a new approach for the estimation of the SF is presented. The introduced model involves six effective parameters on SF [unit weight (γ), pore pressure ratio (r _u), height (H), angle of internal friction (φ), cohesion (C) and slope angle (\(\beta\))], while retaining simplicity as well. In the case of SF prediction, all the datasets were divided randomly to training and testing datasets for proposing the RES model. For comparison purposes, nonlinear multiple regression models were also employed for estimating SF. The performances of the proposed predictive models were examined according to two performance indices, i.e., coefficient of determination (R ²) and mean square error. The obtained results of this study indicated that the RES is a reliable method to predict SF with a higher degree of accuracy in comparison with nonlinear multiple regression models.     

Optimization of the generalization capability for rainfall–runoff modeling by neural networks: the case of the <Emphasis Type="Italic">Lez</Emphasis> aquifer (southern France)

Neural networks are increasingly used in the field of hydrology due to their properties of parsimony and universal approximation with regard to nonlinear systems. Nevertheless, as a result of the existence of noise and approximations in hydrological data, which are very significant in some cases, such systems are particularly sensitive to increased model complexity. This dilemma is known in machine learning as bias–variance and can be avoided by suitable regularization methods. Following a presentation of the bias–variance dilemma along with regularization methods such as cross-validation, early stopping and weight decay, an application is provided for simulating and forecasting karst aquifer outflows at the Lez site. The efficiency of this regularization process is thus demonstrated on a nonlinear, partially unknown basin. As a last step, results are presented over the most intense rainfall event found in the database, which allows assessing the capability of neural networks to generalize with rare or extreme events.     

Application of Fuzzy Set Theory to Rock Engineering Classification Systems: An Illustration of the Rock Mass Excavability Index

The characterization of rock masses is one of the integral aspects of rock engineering. Over the years, many classification systems have been developed for characterization and design purposes in mining and civil engineering practices. However, the strength and weak points of such rating-based classifications have always been questionable. Such classification systems assign quantifiable values to predefined classified geotechnical parameters of rock mass. This results in subjective uncertainties, leading to the misuse of such classifications in practical applications. Fuzzy set theory is an effective tool to overcome such uncertainties by using membership functions and an inference system. This study illustrates the potential application of fuzzy set theory in assisting engineers in the rock engineering decision processes for which subjectivity plays an important role. So, the basic principles of fuzzy set theory are described and then it was applied to rock mass excavability (RME) classification to verify the applicability of fuzzy rock engineering classifications. It was concluded that fuzzy set theory has an acceptable reliability to be employed for all rock engineering classification systems.     

Empirical and numerical analyses of support requirements for a diversion tunnel at the Boztepe dam site, eastern Turkey

This paper presents the engineering geological properties and support design of a planned diversion tunnel at the Boztepe dam site that contains units of basalt and tuffites. Empirical, theoretical and numerical approaches were used and compared in this study focusing on tunnel design safety. Rock masses at the site were characterized using three empirical methods, namely rock mass rating (RMR), rock mass quality (Q) and geological strength index (GSI). The RMR, Q and GSI ratings were determined by using field data and the mechanical properties of intact rock samples were evaluated in the laboratory. Support requirements were proposed accordingly in terms of different rock mass classification systems. The convergence–confinement method was used as the theoretical approach. Support systems were also analyzed using a commercial software based on the finite element method (FEM). The parameters calculated by empirical methods were used as input parameters for the FEM analysis. The results from the two methods were compared with each other. This comparison suggests that a more reliable and safe design could be achieved by using a combination of empirical, analytical and numerical approaches.     

Rock–support interaction analysis based on numerical modelling

The use of yield in supports to control the final loading that develops upon a support system has been one of the most important deformation control techniques used by tunnelling engineers, both historically and currently. Successful use of this approach requires a thorough understanding of the process of rock–support interaction as it is an approach that can fail dramatically if incorrectly applied. There is a fine line between the yield support technique improving the conditions, and the approach resulting in the development of a large area of failed rock, which could ultimately be detrimental. The relationship between the support action and the rock has historically been studied using analytical approaches with the application of significant simplifying assumptions.This paper presents a new approach, where a state-of-the-art numerical model is run repeatedly to develop rock–support interaction curves. This has the advantage of allowing more realistic tunnel geometry, stress states and ground conditions to be simulated. It does, however, use the familiar output form of the relatively simple rock–support interaction curve as opposed to complex and voluminous graphics. Its disadvantage lies in the considerable number of computer runs required to develop the full solutions. Computer software has, however, been written to automate much of this process using a programming language within the modelling package.The analysis approach has been further improved by plotting not one rock–support interaction curve but a whole family of curves representing variations in the rock mass quality of the assumed ground, since this is the most variable of the input parameters for most tunnelling situations. This form of output allows engineers to study the practical range of yield they may require for their rock conditions and also to define at what rock mass quality they can expect the yielding approach to cease to be an effective strategy. This new approach has been presented on a test case history with idealized rock mass properties to illustrate the approach. However, it is an approach that can be specially tailored to any set of rock conditions, tunnel geometry or stress.