煤矿井下煤质预测及工作面三维可视化研究与实现

发布时间：2018-03-24 21:19

本文选题：煤质预测　切入点：Kriging插值　出处：《西安科技大学》2017年硕士论文

【摘要】：煤炭是我国的主要能源资源,科学、合理的开采一直被煤炭企业所重视。了解井下煤炭质量(以下简称煤质)变化趋势及煤层结构分布情况对于煤炭企业进行生产计划、开采方案等的制定具有重要指导意义。井下煤质预测采用的方法是根据样本点煤质信息对待预测区域做插值处理,从而达到对范围或任意点的煤质进行预测的目的。传统的插值方法没有考虑样本点之间的空间自相关性,且容易受样本点变化的影响。本文在对样本数据进行离群值分析和正态校验的基础上,给出了 Kriging进行煤质预测的具体方法,并针对变差函数拟合前的步长选择问题,提出平均最邻近距离进行优化选择。实验证明,优化后的步长相比于传统的经验法确定的步长能拟合出更优的变差函数模型。同时在该步长下,分别以球状模型、指数模型、高斯模型为基础,进行了煤质预测实验,得出最适合本文数据的为指数模型。在井下工作面三维可视化研究中,以面片表示工作面空间结构时,一般多项式曲面的数学解析式对于复杂曲面无法准确表示,甚至无法表示。因此本文将B样条引入工作面的分层模型中。通过分析煤层煤样的数据特征和存储结构,基于B样条曲面给出了工作面单层模型和多层模型的统一表达,并通过反求控制点,使得曲面可以精确的通过已知样本点。对多层模型建立过程中的缺层现象给出了虚拟煤样的解决方案,实验结果表明,B样条曲面计算简单稳定,可以取得较好的建模效果。在深入分析浏览器端三维建模优势和Three.js三维建模原理的基础上,本文实现了浏览器端工作面的三维可视化。为了降低编码耦合,设计了面向对象的三维数据组织模型,给出了基于数据驱动的三维建模方案。并优化了数据请求时延和建模脚本文件安全性的问题,同时基于模型拾取实现了煤质指标与三维工作面的动态关联。在进行旋转、缩放、平移等交互操作时,页面均可以做出流畅的响应。最终,通过采用J2EE标准,开发了具有基础信息管理、报表自动生成、煤质信息预测以及工作面三维可视化功能的Web系统,经过实验测试效果良好,从而证明了本文研究成果的可行性与实际应用价值。
[Abstract]:Coal is the main energy resource of our country, scientific and reasonable mining has been paid attention to by coal enterprises all the time. Knowing the change trend of underground coal quality and the distribution of coal seam structure, it is necessary for coal enterprises to plan their production. The method of underground coal quality prediction is to interpolate the predicted area according to the sample point coal quality information. The traditional interpolation method does not consider the spatial autocorrelation between the sample points. On the basis of outlier analysis and normal check of sample data, the concrete method of coal quality prediction by Kriging is given, and the selection of step size before variation function fitting is discussed. It is proved by experiments that the optimized step size can fit a better variation function model than that determined by the traditional empirical method. At the same time, the spherical model and the exponential model are used respectively under the step size. On the basis of Gao Si model, coal quality prediction experiment is carried out, and the index model is obtained, which is the most suitable for the data of this paper. In the 3D visualization study of underground working face, when the face slice is used to represent the spatial structure of the working face, The mathematical solution of the general polynomial surface can not be accurately represented or even represented for the complex surface. Therefore, this paper introduces B-spline into the stratified model of the coal face. By analyzing the data characteristics and storage structure of the coal sample in coal seam, the paper introduces the B-spline into the stratified model of the coal face. Based on B-spline surface, the unified expression of single-layer model and multi-layer model of working face is given, and the control points are obtained by reverse calculation. The surface can pass through the known sample points accurately. The solution to the phenomenon of missing layer in the process of multi-layer model establishment is given. The experimental results show that the B-spline surface is simple and stable. On the basis of deeply analyzing the advantages of browser side 3D modeling and the principle of Three.js 3D modeling, the 3D visualization of browser end face is realized in this paper. In order to reduce coding coupling, The object oriented 3D data organization model is designed, and the data driven 3D modeling scheme is given, and the delay of data request and the security of modeling script file are optimized. At the same time, the dynamic correlation between coal quality index and 3D working face is realized based on model pick-up. When rotating, zooming, translation and other interactive operations, the page can all make smooth response. Finally, by adopting J2EE standard, The Web system, which has the functions of basic information management, automatic generation of report forms, coal quality information prediction and 3D visualization of coal face, has been developed. The experimental test results show that the feasibility and practical application value of the research results in this paper are good.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52;TD82

【参考文献】