开采沉陷动态预计模型构建与算法实现

发布时间：2018-06-15 16:27

  本文选题：开采沉陷 + 动态预计模型　； 参考：《中国矿业大学(北京)》2017年博士论文

【摘要】：现今我国的主要能源仍然是以煤炭为主,其消费量占能源消费总量的比重超过了60%。从长期来看,随着我国大气污染形势的日趋严峻,用清洁能源逐步取代煤炭能源的趋势已不可避免,但以煤炭为主的消费结构在短期内还将继续维持。巨量的煤炭消费导致了煤炭的大量开采,而由煤炭开采所引起的土地塌陷对农田和地表环境的危害已十分严重。同时由于我国的城镇化发展迅速,直接导致了我国建筑物下、水体下及铁路下(简称“三下”)的压煤量逐年增长,“三下”煤炭开采量也在逐年增大,由此带来的地表沉陷危害和造成的经济损失也越来越大,带来了一系列严重的社会问题。对采矿引起的地表及覆岩破坏规律的研究可追溯到19世纪末,通过文献分析可知,对开采沉陷稳定后的地表及覆岩移动变形研究所占比例较大,而对开采过程中地表及覆岩移动变形的动态研究相对较少。矿山开采地表移动变形动态预计是开采沉陷研究领域的重要内容之一,掌握开采沉陷随时间变化的动态过程,不但可以随时掌握地表及覆岩的移动与变形分布规律,还可实时得出地表最大变形值出现的时间和空间位置,对评价采动过程中地表建构筑物的变形过程和破坏程度,及时制定建筑物的维修与加固方案和确定实施时间等都有重要的指导意义。本文在其他学者研究的基础上,对开采沉陷动态预计存在的一些问题进行了深入的研究,主要包括以下几个方面:一,对可用于开采沉陷动态预计的常用时间函数模型进行研究,指出了它们的优缺点及适用性,对分段Knothe时间函数进行分析,指出其在实际应用中所存在问题,针对问题提出了相应的模型修正方法;二,对矩形工作面开采走向主断面、倾向主断面及地表任意点动态移动变形预计方法进行了研究,并设计了相应的算法;三,对不规则工作面积分区域的三角形划分方法进行了研究,并设计了相应的算法;四,研究了不规则工作面开采地表移动变形动态预计方法与算法;五,采用MATLAB语言实现了论文所提出的算法,完成了开采沉陷预计一体化系统的开发;六,研究了数值模拟岩层力学参数求取的正交实验方法,并利用数值模拟方法研究开采沉陷动态问题;最后,针对复杂地形的3DEC数值建模问题,提出了一种新的技术方法。本文所取得的主要研究成果如下:(1)对分段Knothe时间函数进行了研究与优化。通过研究,指出了分段Knothe时间函数模型存在的两个主要问题:一,在分段点t时刻,其时间函数值与理论值不相符,如采用该时间函数进行动态预计,在分段的t时刻会出现预测误差,并且这种偏差会随着两个参数乘积的减小而增大,只有当两个参数乘积大于一定的数值时才能有效地降低误差,这一问题大大的限制了使用该时间函数进行动态预计的适用性;二,分段Knothe时间函数的值在两参数乘积较小时并不收敛到1,而是收敛到小于1的值,并且其乘积越小,收敛值与1的偏差就越大,这会降低其对地表终态(最大)下沉值的预计精度。针对分段Knothe时间函数存在的问题,本文提出了修正该时间函数模型的方法,对原函数进行了优化,经过优化后的模型,无论其参数t和c如何取值,其函数值在t时刻均与理论值相符,并且时间函数值最终收敛到1,成功的解决了原时间函数在实际应用中存在的问题。(2)本文引入了一种新的时间函数—正态分布时间函数,对该时间函数值的变化范围和变化规律进行了对比分析,探讨了可用于开采沉陷动态预计的时间函数应具备基本条件,指出了正态分布时间函数的参数意义,并对该时间函数的动态预计适应性进行了研究,指出:一,正态分布时间函数的下沉速度和下沉加速度与地表动态下沉规律是相吻合的;二,当正态分布时间函数的参数d取值不同时,其时间函数的值并不都能收敛到1,且随着d取值由大到小,时间函数值收敛于1的误差就会由小到大,这说明了d取值越小,越不适合作为动态预计的时间函数,在进行精度要求不高的动态预计时,d可取大于等于2的值,当预测精度要求较高时,则要求d取大于或等于3的值,当d小于1时,用其进行动态预计时的误差会很大。(3)在概率积分法的基础上,结合经优化的分段Knothe时间函数模型,本文研究了矩形工作面动态开采单元的划分方法、时间函数参数的确定方法,及各动态开采单元所对应的时间函数值的求取方法。在倾向主断面动态预计中,由于每个动态单元所处的上山及下山方向采深不同,所对应的预计参数也将不同,因此本文还研究建立了倾向主断面各动态单元概率积分参数的计算公式。最后,分别建立了矩形工作面走向主断面、倾向主断面及地表任意点的动态预计模型,并设计了相应的编程算法。(4)针对优化分段Knothe时间函数的参数求取问题,本文提出了基于实测数据的“反求时间函数对比法”,该方法步骤简洁,结果可靠。另外,根据矿区已有概率积分参数,在波兰学者Knothe“图解法”求参方法的基础上,本文还研究了时间函数参数的“直接计算法”,根据该方法,只要知道某矿区的概率积分预计参数,便可较为准确的求取该矿区所对应的时间函数的c参数值,计算出的c值不再是一个估计区间,而是一个定值,这非常便于计算机编程实现。(5)对不规则工作面积分区域的三角形划分方法进行了研究,给出了不同形状三角形积分区域的划分原则,推导了分步积分过程中,积分区间上下限的确定方法,提出了不规则工作面积分区域的“三角剖分”概念,并设计了相应的算法,该算法可满足水平煤层和缓倾斜煤层不同采深、不同形状的多工作面开采时的静态沉陷预计。另外,为了使动态预计公式及算法简洁、高效,在进行动态预计之前,通常需要将矿区井上下对照图进行坐标变换。坐标变换主要包括将工作用图的坐标系由国家坐标系转换为工作面坐标系,还包括将预计结果的坐标从工作面坐标系向国家坐标系或向其他坐标系进行转换。因此,本文详细的研究了坐标系之间的相互转换方法,给出了实际操作时各坐标转换参数的确定方法。(6)基于不规则工作面积分区间的“三角剖分”算法,对不规则工作面开采时的动态预计方法进行了研究,给出了不规则工作面动态预计积分域的实时确定方法、任意给定时刻各动态单元时间函数区间的确定方法,研究了工作面顶点顺序排序及任意排序时的两种动态预计方法,并设计了相应的算法,同时,对特殊的凹多边形工作面开采时的动态预计方法进行了探讨,提出了相应的计算流程和相应的算法,解决了凹多边形工作面积分区间和时间函数区间的实时确定问题。(7)采用MATLAB软件实现了论文所提出的各种算法,完成了开采沉陷预计系统的开发,在水平或缓倾斜煤层开采条件下,该系统可对任意形状工作面开采进行动态预计和静态预计,可对不同水平、不同地质采矿条件下的多工作面开采进行地表移动变形的动/静态预计。另外,该系统界面友好,功能齐全,具有强大的数据处理及三维可视化表达功能,可不必采用第三方软件而直接对预计结果进行操作,提高了应用效率。最后通过实例验证可知,采用本系统进行预测,预测结果的相对误差多可控制在8%以内,证明了系统的可靠性。(8)针对复杂地形的3DEC数值建模问题,本文提出了一种基于地表等高线数据,采用Auto CAD和MATLAB软件编程来建立复杂地形数值模型的技术方法。利用该方法建立模型,前期数据处理工作量小、建模速度快,且能够建立任意复杂地形的三维数值模型,所建数值模型的地表曲面与实际地形表面高度吻合。(9)对王庄矿6206复杂形状工作面开采进行了数值模拟研究,详细介绍了该工作面数值模型的建立方法;针对数值模拟中岩体物理力学参数获取困难的问题,根据已有地表倾向和走向观测站的地表下沉监测数据,详细论述了基于正交试验与数值模拟相结合的岩层力学参数求取方法;(10)在数值模拟中,通过研究得出:当开采速度很大时,由于一次性开采面积较大,在较短时间内,由开采引起的岩层与地表的移动范围就已基本形成,但其数值都还相对较小;当模型中存在断层时,岩层的竖向位移在断层的两侧明显增大,以断层为分界线,在其两侧一定的范围内,岩层的竖向位移较大,呈塔形发育,随着开采时间的增加塔形范围逐渐增大,竖向位移也逐渐增大,然后,随着时间的再增加,塔形范围逐渐消失,其与周边岩层的位移之差也逐渐缩小,但断层两侧的位移量和周边岩层相比明显较小,岩层的位移量仍然会沿着断层呈现出较为对称的形态。(11)在不规则工作面的数值模拟中,无论倾向开采长度是多少,在开始阶段,当开采面积较小时,随着走向开采长度的增大,模型稳定所需的计算时步数迅速增加,但当走向开采长度增大到一定的数值后,尽管走向开采增加的长度相同,但模型稳定所需的计算时步增加量明显减小;通过分析,本文拟合得到了模型移动稳定时的计算时步数和开采面积之间的关系式:n=8.83A~(0.7032)-0.2481A。
[Abstract]:At present, the main energy of our country is coal mainly. The proportion of consumption in the total amount of energy consumption exceeds 60%. in the long run. With the increasingly severe air pollution situation in our country, the trend of replacing coal energy with clean energy is inevitable, but the consumption structure based on coal will continue to maintain in the short term. A huge amount of coal consumption has led to a large amount of coal mining, and the land collapse caused by coal mining has been seriously harmful to the farmland and the surface environment. At the same time, the rapid development of the urbanization in China has directly led to the construction of China, under the water body and under the railway (simply called "three") the volume of coal pressure increased year by year, "three under". The amount of coal mining is increasing year by year, and the damage and economic loss caused by the surface subsidence are becoming more and more serious, which brings a series of serious social problems. The study of the destruction law of the surface and overlying rock caused by mining can be traced back to the end of the nineteenth Century. The movement of the surface and overlying rock after the mining subsidence is stable The deformation research occupies a large proportion, and the dynamic deformation of the surface and overlying rock in the mining process is relatively less. The dynamic prediction of the surface movement and deformation of the mine mining is one of the important contents in the field of mining subsidence research. The time and space position of the maximum deformation value of the surface can be obtained in real time. It has important guiding significance for the evaluation of the deformation process and damage degree of the surface construction in the mining process, the timely formulation of the maintenance and reinforcement scheme of the building and the determination of the implementation time, etc. This paper is based on the research of other scholars, Some problems in the dynamic prediction of mining subsidence are studied in depth, including the following aspects: first, the common time function model, which can be used for mining subsidence dynamic prediction, is studied, their advantages and disadvantages and applicability are pointed out, and the subsection Knothe time function is analyzed, and it is pointed out that it is stored in the practical application. In the problem, the corresponding model correction method is put forward for the problem. Two, the method of moving the rectangle working face to the main section, the tendency of the main section and the dynamic deformation of the surface at any point is studied, and the corresponding algorithm is designed. Three, the triangle division method of the irregular working area subregion is studied and the design of the method is also designed. Corresponding algorithms; four, the method and algorithm for dynamic prediction of surface movement and deformation in irregular working face are studied. Five, the algorithm proposed in the paper is realized by MATLAB language, and the development of the integrated system for mining subsidence prediction is completed. Six, the orthogonal experimental method for calculating the parameters of the numerical simulation of rock strata is studied and the numerical simulation is used. This paper studies the dynamic problem of mining subsidence. Finally, a new technical method is proposed for the 3DEC numerical modeling of complex terrain. The main results obtained in this paper are as follows: (1) the piecewise Knothe time function is studied and optimized. Through the study, the two main problems of the piecewise Knothe time function model are pointed out. First, the time function value is not consistent with the theoretical value at the piecewise point t time. If the time function is used for dynamic prediction, the prediction error will appear in the t moment of the segment, and the deviation will increase with the decrease of the product of the two parameters. Only when the product of the two parameters is larger than a certain value can the error be reduced effectively. The problem greatly limits the applicability of the dynamic prediction using this time function; two, the value of the piecewise Knothe time function does not converge to 1 when the product of the two parameter is small, but converges to a value less than 1, and the smaller the product's product, the greater the deviation of the convergence value from 1, which will reduce the expected precision of the final (maximum) subsidence value of the surface. In view of the problems existing in the piecewise Knothe time function, this paper proposes a method to modify the time function model and optimizes the original function. After the optimized model, the value of its parameter T and C is consistent with the theoretical value at the time of T, and the value of the time function converges to 1, and the original time function is successfully solved. There are problems in practical application. (2) a new time function, normal distribution time function, is introduced in this paper. The variation range and change law of the time function value are compared and analyzed. The basic conditions for the time function which can be used for mining subsidence dynamic prediction should be discussed, and the reference of the normal distribution time function is pointed out. The dynamic prediction adaptability of the time function is studied. First, the sinking velocity and subsidence acceleration of the normal distribution time function are consistent with the law of the surface dynamic subsidence. Two, when the parameter D values of the normal distribution time function are not at the same time, the value of the time function can not all converge to 1, and the value of the time function can be obtained with the value of the time function. From large to small, the error of the time function value converging to 1 will be from small to large, which shows that the smaller the D value is, the more unsuitable for the dynamic prediction time function, when the dynamic prediction of the precision is not high, the D should be more than equal to the value of 2. When the precision of the prediction is higher, the D is required to be greater than or equal to the value of 3, when D is less than 1, use it The error of dynamic prediction is very large. (3) on the basis of the probability integral method, combined with the optimized piecewise Knothe time function model, this paper studies the method of dividing the dynamic mining unit of the rectangular working face, the method of determining the parameters of the time function, and the method of obtaining the time function values corresponding to each dynamic mining single element. In the dynamic prediction of main section, the corresponding prediction parameters will be different because of the different depth of the mountain and the downhill direction of each dynamic unit. Therefore, the calculation formula of the probability integral parameters of the dynamic elements of the main section is also established. Finally, the main section of the rectangular working face is established, the main section and the ground are inclined to the main section. The dynamic prediction model of an arbitrary point is presented and the corresponding programming algorithm is designed. (4) for the optimization of the parameters of the optimized piecewise Knothe time function, a "inverse time function comparison method" based on the measured data is proposed in this paper. The method is simple and the result is reliable. In addition, according to the existing probability integral parameters of the mining area, the Poland scholar Knothe On the basis of the "graphical method" method, this paper also studies the direct calculation method of time function parameters. According to this method, the C parameter of the time function corresponding to the mining area can be calculated accurately as long as the parameters of the probability integral of a mining area are known, and the calculated C value is no longer an estimate interval, but a definite interval. Value, this is very convenient for computer programming. (5) the triangle division method of irregular area of area is studied, the principle of dividing the area of different shape triangle integral is given. The method of determining the lower limit on the integral interval in the step integral process is derived, and the triangulation of the irregular working area subregion is put forward. According to the concept, the corresponding algorithm is designed, which can meet the static subsidence prediction of different mining depth of horizontal and gently inclined coal seam with different shape and multi face mining. In addition, in order to make the dynamic prediction formula and algorithm simple and efficient, it usually needs to coordinate the coordinate change of the upper and lower control charts in the mining area before the dynamic prediction is carried out. The coordinate transformation mainly includes converting the coordinate system of working diagram from the national coordinate system to the working plane coordinate system. It also includes the conversion of the coordinates of the predicted results from the working face coordinate to the national coordinate system or to the other coordinate systems. Therefore, this paper studies the mutual conversion method in the coordinate system in detail, and gives the actual operation. The method of determining the parameters of each coordinate conversion. (6) based on the "triangulation" algorithm between the irregular working area partition, the dynamic prediction method of the irregular working face mining is studied. The real-time determination method of the dynamic prediction integral domain of the irregular working face is given, and the time function interval of the dynamic units is engraved at the time. According to the method, two dynamic prediction methods for the order of vertex order and arbitrary ordering of the working face are studied, and the corresponding algorithms are designed. At the same time, the dynamic prediction method of the special concave polygon working face is discussed, the corresponding calculation process and the corresponding algorithm are put forward to solve the partition between the working area of the concave polygon. The real time determination of the interval of the time function. (7) the various algorithms proposed in the paper are realized by MATLAB software, and the development of the mining subsidence prediction system is completed. Under the conditions of horizontal or gently inclined coal seam mining, the system can dynamically predict and predict the mining of arbitrary shape working face, and can be used for different levels and different geological conditions. In addition, the system has friendly interface and full function, and has powerful data processing and three-dimensional visualization expression. It can not use third party software to operate the predicted results directly and improve the application efficiency. Finally, the application efficiency is improved. Finally, the example is proved to be known. The relative error of the prediction results can be controlled within 8%, which proves the reliability of the system. (8) in view of the 3DEC numerical modeling of complex terrain, this paper presents a technical method based on the surface contour data, using the Auto CAD and MATLAB software programming to establish a complex terrain numerical model. In this method, the model is built, the workload of the early data processing is small, the modeling speed is fast, and the three-dimensional numerical model of any complex terrain can be established. The surface surface of the numerical model is in good agreement with the surface of the actual terrain. (9) the numerical simulation of the mining of the 6206 complex shape working face of Wang Zhuang mine is studied, and the numerical model of the working face is introduced in detail. In view of the difficulties in obtaining the physical and mechanical parameters of rock mass in numerical simulation, according to the surface subsidence monitoring data of the existing surface tendencies and the direction of the observation station, the method of obtaining the mechanical parameters of rock strata based on the combination of orthogonal test and numerical simulation is discussed in detail. (10) in numerical simulation, it is obtained that when mining is exploited. When the velocity is very high, the moving range of rock layer and surface caused by mining is basically formed in a relatively short time, but its numerical value is relatively small. When there is a fault in the model, the vertical displacement of the rock layer increases obviously on both sides of the fault, and the fault is divided into a certain range on both sides of the fault. The vertical displacement of the rock stratum is larger and the tower shape develops. With the increase of mining time, the tower shape gradually increases and the vertical displacement increases gradually. Then, as the time increases, the range of the tower shape gradually disappears, and the difference between the displacement and the surrounding rock is gradually narrowed, but the displacement on both sides of the fault layer is smaller than that of the surrounding rock layer. The displacement will still show a more symmetrical shape along the fault. (11) in the numerical simulation of the irregular working face, no matter how much the mining length is inclined, at the beginning, when the mining area is small, the number of steps required for the model stability increases rapidly with the increase of the mining length, but when the mining length is increased to the length of mining, After a certain number of values, although the length of the mining increase is the same, the step increment required for the model stability decreases obviously. Through the analysis, the relationship between the time of calculation and the area of the mining area when the model is stable is obtained by fitting the model: n=8.83A~ (0.7032) -0.2481A.
【学位授予单位】：中国矿业大学(北京)
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TD327
，

本文编号：2022674