被保护煤层残余瓦斯含量预测模型研究
发布时间：2018-08-26 14:27
【摘要】：开采保护层是现在煤矿生产过程中防止瓦斯事故发生应用较普遍的措施,开采保护层会使得被保护煤层裂隙增多,瓦斯发生大量解吸,从而方便进行瓦斯的抽采。煤层残余瓦斯含量是检验保护层开采效果的重要指标之一,检测煤层瓦斯含量的主要方法是井下直接测量或采集煤样在实验室间接测量,但这不能实时得到煤层的瓦斯含量的数据及测量存在一定误差。基于此,本文通过SPH(光滑粒子)方法对穿层钻孔抽采被保护层瓦斯进行模拟,主要研究在穿层钻孔抽采下,被保护层的内的瓦斯运移情况以及对有可能影响被保护层残余瓦斯含量的因素进行分析,然后利用灰色关联度分析方法得到影响煤层残余瓦斯含量的最主要的因素,通过查阅资料及数值模拟的数据作为训练BP神经网络的样本数据,得到基于灰色关联度分析的BP神经网络来对被保护层残余瓦斯含量预测模型,可以实时得到被保护层的残余瓦斯含量,有利于煤层开采的安全进行。本文的主要工作如下:(1)依据瓦斯的渗流理论,建立了穿层钻孔抽采被保护煤层瓦斯SPH模型;通过计算均质煤层中瓦斯压力及涌出量的解析解和SPH解,结果表明采用SPH方法对煤层瓦斯渗流进行数值模拟具有准确性和可行性;并且得到了被保护煤层瓦斯在抽采过程中运移的一般规律。(2)通过所建立的穿层钻孔抽采被保护煤层瓦斯SPH模型,模拟研究了不同的原始瓦斯压力、透气性系数、抽采负压等因素下被保护煤层中的瓦斯流动规律,研究了钻孔瓦斯流量与被保护层残余瓦斯含量的关系,得到流量拟合参数m和n随着煤层瓦斯压力的变大,其值也会随之增大,反之亦然;当煤层的透气性系数增大时,m的值会随之增大,而n的绝对值会随之减小。研究了通过瓦斯涌出规律来预测煤层残余瓦斯含量这种方法的可行性。(3)分析BP神经网络的运算原理及算法的实现,对所构建神经网络所需要的样本数据进行收集,并且通过灰色关联度分析对神经网络的输入层进行选择。确定该预测被保护层残余瓦斯含量的BP神经网络的输入层、隐含层和输出层的节点数分别为5、7、3,传递函数、训练函数及学习函数分别为logsig函数、trainlm函数及Learngdm函数。(4)被保护层残余瓦斯含量快速预测模型的实现及结果分析,通过前面一系列的准备工作,然后利用MATLAB软件对所构建的神经网络进行训练,将现场所得到的数据代入所得到的模型中进行检测,通过误差率的分析,使得模型的误差率在可以接受的范围之内,这就得到了被保护层残余瓦斯含量快速预测模型。
[Abstract]:Mining protective layer is a common measure to prevent gas accident in the process of coal mine production. Mining protective layer will cause more cracks in protected coal seam and a large amount of gas desorption, thus facilitating the extraction of gas. The residual gas content in coal seam is one of the important indexes to test the mining effect of protective layer. The main method to detect the gas content in coal seam is to measure directly underground coal or to measure coal samples indirectly in laboratory. However, there are some errors in the data and measurement of gas content in coal seam. Based on this, this paper uses SPH (smooth particle) method to simulate the gas extraction of the protected layer by drilling through the layer, and mainly studies the extraction of gas in the hole through the layer. The gas migration in the protected layer and the factors that may affect the residual gas content in the protected layer are analyzed, and the main factors affecting the residual gas content in the coal seam are obtained by using the grey correlation analysis method. Through consulting the data and numerical simulation data as the sample data of training BP neural network, the BP neural network based on grey correlation degree analysis is obtained to predict the residual gas content in protected layer. The residual gas content of the protected layer can be obtained in real time, which is beneficial to the safety of coal seam mining. The main work of this paper is as follows: (1) based on the seepage theory of gas, the SPH model of gas extraction in protected coal seam by borehole drilling is established, and the analytical and SPH solutions of gas pressure and emission in homogeneous coal seam are calculated. The results show that the numerical simulation of coal seam gas seepage with SPH method is accurate and feasible. The general law of gas migration in the course of coal seam gas extraction is obtained. (2) through the established SPH model of gas extraction by drilling holes, the different original gas pressure and permeability coefficient are simulated and studied. In this paper, the relationship between gas flow rate of borehole and residual gas content in protected seam is studied, and the flow fitting parameters m an dna re obtained with the increase of gas pressure in coal seam. When the permeability coefficient of coal seam increases, the value of m will increase and the absolute value of n will decrease. The feasibility of the method of predicting the residual gas content of coal seam through the law of gas emission is studied. (3) the operation principle and algorithm of BP neural network are analyzed, and the sample data needed to build the neural network are collected. The input layer of neural network is selected by grey correlation analysis. The number of nodes in the hidden layer and the output layer of the BP neural network for predicting the residual gas content in the protected layer is 5 / 7 / 3, respectively, and the transfer function is used to determine the input layer of the BP neural network for predicting the residual gas content in the protected layer. The training function and the learning function are logsig function trainlm function and Learngdm function respectively. (4) the realization and result analysis of the fast prediction model of residual gas content in protected bed. Then the neural network is trained by MATLAB software, and the data obtained in the field is detected in the model. The error rate of the model is within the acceptable range through the analysis of the error rate. Thus, a fast prediction model of residual gas content in protected layer is obtained.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD712.3
本文编号：2205153
[Abstract]:Mining protective layer is a common measure to prevent gas accident in the process of coal mine production. Mining protective layer will cause more cracks in protected coal seam and a large amount of gas desorption, thus facilitating the extraction of gas. The residual gas content in coal seam is one of the important indexes to test the mining effect of protective layer. The main method to detect the gas content in coal seam is to measure directly underground coal or to measure coal samples indirectly in laboratory. However, there are some errors in the data and measurement of gas content in coal seam. Based on this, this paper uses SPH (smooth particle) method to simulate the gas extraction of the protected layer by drilling through the layer, and mainly studies the extraction of gas in the hole through the layer. The gas migration in the protected layer and the factors that may affect the residual gas content in the protected layer are analyzed, and the main factors affecting the residual gas content in the coal seam are obtained by using the grey correlation analysis method. Through consulting the data and numerical simulation data as the sample data of training BP neural network, the BP neural network based on grey correlation degree analysis is obtained to predict the residual gas content in protected layer. The residual gas content of the protected layer can be obtained in real time, which is beneficial to the safety of coal seam mining. The main work of this paper is as follows: (1) based on the seepage theory of gas, the SPH model of gas extraction in protected coal seam by borehole drilling is established, and the analytical and SPH solutions of gas pressure and emission in homogeneous coal seam are calculated. The results show that the numerical simulation of coal seam gas seepage with SPH method is accurate and feasible. The general law of gas migration in the course of coal seam gas extraction is obtained. (2) through the established SPH model of gas extraction by drilling holes, the different original gas pressure and permeability coefficient are simulated and studied. In this paper, the relationship between gas flow rate of borehole and residual gas content in protected seam is studied, and the flow fitting parameters m an dna re obtained with the increase of gas pressure in coal seam. When the permeability coefficient of coal seam increases, the value of m will increase and the absolute value of n will decrease. The feasibility of the method of predicting the residual gas content of coal seam through the law of gas emission is studied. (3) the operation principle and algorithm of BP neural network are analyzed, and the sample data needed to build the neural network are collected. The input layer of neural network is selected by grey correlation analysis. The number of nodes in the hidden layer and the output layer of the BP neural network for predicting the residual gas content in the protected layer is 5 / 7 / 3, respectively, and the transfer function is used to determine the input layer of the BP neural network for predicting the residual gas content in the protected layer. The training function and the learning function are logsig function trainlm function and Learngdm function respectively. (4) the realization and result analysis of the fast prediction model of residual gas content in protected bed. Then the neural network is trained by MATLAB software, and the data obtained in the field is detected in the model. The error rate of the model is within the acceptable range through the analysis of the error rate. Thus, a fast prediction model of residual gas content in protected layer is obtained.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD712.3
【参考文献】
中国期刊全文数据库 前10条
1 王利忠;汪开旺;;预抽钻孔改进措施及抽采效果考察[J];煤炭与化工;2017年01期
2 王磊;;高瓦斯矿井瓦斯灾害隐患排查及防控技术研究[J];中国煤炭;2016年10期
3 杨家威;;煤矿瓦斯灾害防治技术的研究[J];能源与节能;2016年10期
4 张建华;;保护层开采效果考察及应用[J];中州煤炭;2016年10期
5 杨军伟;艾德春;邱燕;汤铸;;近距离煤层群上保护层保护范围的数值模拟[J];中国煤炭;2016年09期
6 崔晓松;;用解吸法测试煤层瓦斯含量现状研究[J];科技与创新;2016年03期
7 艾孜海尔·哈力克;热合买提江·依明;开依沙尔·热合曼;买买提明·艾尼;;不同质量粒子分布SPH方法及其应用[J];振动与冲击;2015年22期
8 张浩;倪福生;顾磊;张德义;;ALE方法及SPH方法模拟高速射流破土过程的对比[J];水电能源科学;2015年11期
9 潘建平;曾庆筠;;SPH方法在土工大变形分析中的应用研究进展[J];安全与环境学报;2015年05期
10 李洪生;李树清;谭玉林;;煤层群保护层开采研究的现状与趋势[J];矿业工程研究;2015年03期
中国硕士学位论文全文数据库 前1条
1 钟声;基于遗传算法的BP神经网络的土坝渗流场反演研究[D];西安理工大学;2010年
，本文编号：2205153
本文链接：https://www.wllwen.com/kejilunwen/anquangongcheng/2205153.html
