| "Digital Mine, DM" is the common digital representation of the entity and its related phenomena on real mines. It is a "Virtual Mine" that reflects mining development and management process via a digital way. Based on the computer and network technology, through mine data collection, transmission, storage, interpretation, processing and reformulation, DM achieves the goals of optimizing production planning, improving administration effectiveness and supporting scientific decision making by applying mine data in mine production and construction process. Studying 3D geology modeling and visualization of mines is the core technology of "Digital Mine", which is a mathematical model representing the characteristic of the geological structure based on specific data structure; meanwhile, which is also a process of using 3D visual technology to display and analyze the attribute information of the mine model. Due to the underlying process of Digital Mine,3D mine model provides an abundant geological evidences and effective reference for mining enterprises in exploration design, construction plan and object decision.During the mining process, mining enterprises have multi-level requirements in the prediction, evaluation, design and other aspects. In the stage of mineral resources evaluation, the changing characteristic of recoverable reserves in the mining and transportation process should be predictable. In the stage of mining design, the mining plans and material proportioning schemes should be worked out according to the ore grade and thickness of ore body, In the stage of extraction of mining exploration, accurate estimation of extraction reserves in the three grade ore reserves is needed. According to these different purposes mentioned above, the 3D model of mines should be the true reflection of overall importance, volatility and accuracy. Comparing to the multi-degree characteristics of the mine 3D entity, the traditional modeling approaches and their single-degree characteristics have become a bottleneck that restricts the application of 3D model of mines. Therefore, based on the multi-scale features of cognition during the mining process and different spatial interpolation, building multiple sequences of three-dimensional models of mines with multiple levels and characteristics can improve the design capacity and production efficiency of mining enterprises. At the same time, during the process of mine modeling, the accuracy of mining will be influenced by many aspects such as the instability of 3D geologic bodies, the complexity of geological structures, the sparsity of sample data and the cognition subjectivity of geologists.3D mine model is an approximate description of natural geological bodies and usually contains uncertainties,these uncertainties always trigger deviations of spatial resolutions, shapes and topologies between 3D models of mines and geologic bodies. Therefore, based on the investigation of main sources of uncertainties in mine modeling, analyzing the propagation of uncertainties in modeling process, building quantitative analysis and standard expressions of uncertainties in 3D models of mines, establishing a framework of uncertainty estimation of mine models will greatly enhance the practical values of 3D models of mines.The research domains of this dissertation are the basic theories and essential techniques of geology spatial data modeling, approaches for 3D geology modeling and visualization methods for 3D geological entities, geological statistics methods and spatial information prediction. It discusses the fundamentals and the process of 3D geology modeling, improves the profound comprehension of geologic entity and modeling approaches. From the perspective of semantic scale, it integrates geological cognition and the mining process, proposes principles and methods for multiple sequences of 3D models of mines. In order to analyzing the origin of uncertainty and propagation of uncertainty in modeling process, this dissertation extendsthe Geological Metadata Standard in uncertainty fields, applies information entropy as an objective measure in researching uncertainties of 3D models of mines and implements a process of quantitative analysis and model updating.This dissertation assimilates and borrows ideas and research findings from other researchers in 3D geology modeling technology. Centering the multiple level requirements analysis and uncertainty analysis of the 3D model of mines as two basic points, the research developed new methods on the application of mine models in production design, decision support and reserve evaluation. The results of the modeling and uncertainty analysis of the experimental copper deposit data show that the proposed theories and methods by this thesis is feasible, practical and rational.This dissertation is organized as follows:1. In Chapter One, the paper first briefly evaluates problems such as design, production and decision related to the application of 3D model of mines. Then the direction of multi-degree expression and uncertain analysis of the mine model are described. It systematically reviews the state-of-art technologies of 3D geology modeling and uncertainty, and defines the primary research subject and the research context.2. In Chapter Two, the thesis reviews the concepts and theories and gives a detailed description of spatial data modeling in 3D geologic entities. It systematically analyzes 3D vector models,3D volume models and hybrid models, and elaborate the function of geological statistics methods and spatial information prediction in 3D mine modeling. In this dissertation, the basic concepts and related terms of uncertainty are summarized and analyzed, theories and methods of uncertainties in Geographical Information Systems (GIS) are described as well. By expatiating the theories of architecture of the 3D model of mines and uncertainty analysis, this chapter offers reference for construction of multiple sequences of mine models and uncertainty analysis3. In Chapter Three, the geological cognition and the mining process are discussed,the multi-scale characteristics and semantic content of geological data of mines are explained in detail. Based on theories of 3D mine modeling and geological statistics methods,this dissertation presents a conceptual framework of 3D semantic scale modeling in terms of the multi-degree characteristics of the mine 3D entity. According to different semantic levels of mining processes,various application scopes of mine models and the framework of 3D semantic-scale, by means of kriging interpolation and stochastic simulation algorithm, this dissertation proposes the principles and an approach for multiple sequences of 3D models of mines based on semantic scale, which makes the 3D models meet the demand of mining enterprises at multiple levels including overall, correlation and accuracy aspects. Meanwhile, in the building process of multiple sequence model, using confidence parameters to constrain the simulation computation, the modeling approach implements quantitative expression for semantic granularity in the stochastic simulation algorithm.4. In Chapter Four, the origin of uncertainty and the uncertainty propagation model and uncertainty algorithm implementation during the process of 3D modeling of mine are described in detail. Based on the analysis of the characteristics and workflow of modeling processes of mines, this research divided the origin of uncertainties of 3D mine model into five categories:data acquisition and measurement error, inconsistency and incompleteness of data, stochastic uncertainty, uncertainty caused by incomplete cognition, uncertainty caused by modeling software.Using reasoning theory in artificial intelligence, this research explored a new uncertainty propagation algorithm in mine modeling and built an uncertainty propagation model of mine modeling, analyzed the uncertainty propagation mechanism of mine modeling in detail.5. In Chapter Five, as soon as the definition and properties of the information entropy are discussed and analyzed, this dissertation extends the information entropy as an objective measure and a type of quantitative analysis tools to evaluate the uncertainty of 3D mine model as follows:Based on cell-based discretization of model, the 3D mine model should be discretized as the same size voxel. A number of simulation ware carried out to determine the probability of each voxel and its value of information entropy can also be worked out by the probabilistic field of each voxel. Utilizing the information entropy as a measurement to quantify uncertainties for each voxel of the mine model, the purposes of quantitative analysis, evaluation and visualization of uncertainties for 3D mine model are achieved. This dissertation proposes the framework and the implementation steps of quantitative analysis, error measurement and model updating for 3D mine model and related key algorithms are implemented for a 3D grade model of mines.6. In Chapter Six, using geological information and data in different stages of an experimental copper deposit and the 3D grade model of copper ore as a case study, this research built the multiple sequences of 3D models of mine based on multiple semantic scale. From the view of frequency histogram, variation function and the accuracy of sampling point of mine models, the overall importance, volatility and accuracy are investigated to verify that multiple sequence of 3D models have multi-degree characteristics. By using entropy-based method, the thesis implemented quantitative analysis of uncertainties of 3D grade model of copper ore. According to the experimental results, the range and position of supplementary geologic borehole can be determined and the updating process of the 3D mine mode can be achieved. According to changes in the state of 3D mine model before and after the model updating experiment, the expected results were achieved and it shows that the theory of quantitative analysis of uncertainties for mine model proposed by this dissertation is feasible, practical and rational.7. In Chapter Seven, summary of the research and future work are given.The contributions of this dissertation are summarized as follows:1.On the basis of studying scale of geographical information, combining theories and methods of 3D modeling of mines, this dissertation presents a conceptual framework of 3D semantic-scale modeling of mines. It discusses the principle of geological statistics methods and spatial information prediction, develops mechanism of building multiple sequence of 3D models of mines based on semantic scale, implements the multi-degree expression of the 3D model of mines.2. It expounds the mechanism of using confidence parameters to control the simulation computation and implements the quantitative expression for semantic granularity in stochastic simulation algorithm. Aiming at improvement of traditional methods of 3D mine modeling and the framework of multiple sequence of 3D models of mines based on semantic scale, the workflow of constructing structural model and attribute model of mine are enhanced. Through building multiple sequences of 3D grade models of mines of an experimental copper deposit, the related theories and feasibility of the method are verified.3. It summarizes the factors of uncertainties in the process of mine modeling, and it takes reasoning theory in artificial intelligence as foundation to develop a new uncertainty propagation algorithm in mine modeling and an uncertainty propagation model.4. It applies the information entropy theories to the research process of uncertainties in 3D modeling of mines. Using the information entropy as measurement tools, it develops the framework and the implementation resutls of quantitative analysis of the uncertainty of mining models and related quantitative analysis are applied to the 3D grade model of mines. According to the process of quantitative analysis and model updating for grade model of mines in the studying of an experimental copper deposit, it shows that the theory of quantitative analysis of uncertainties for mine model proposed by this dissertation is feasible, practical and rational. |
PDF Full Text Request