开拓前煤与瓦斯突出危险性区域预测技术研究

发布时间：2018-10-23 09:30

【摘要】：煤与瓦斯突出是威胁煤矿安全生产的最严重自然灾害之一。目前我国煤与瓦斯突出区域预测主要集中在开拓后依靠井下巷道、实测瓦斯参数为主的点预测,现场工程实践遇到的开拓前煤与瓦斯突出危险性预测难题,一直没有得到切实可行的解决方法。支持向量机作为一种可以很好的解决小样本、非线性、高维数等实际问题的模式识别方法,已被广泛应用到众多生产领域预测预报中,为此本文引入支持向量机建立学习模型对煤矿开拓前的突出危险进行预测,期望提高开拓前煤与瓦斯突出预测准确性。山西吕梁某矿9号煤层尚未揭露,无井下瓦斯参数实测条件,对矿井瓦斯赋存规律进行研究,收集矿井地勘时期所测的瓦斯参数并进行分析。采用钻孔煤屑解吸法实测同属一个地质单元邻近矿井瓦斯含量,对地勘瓦斯含量进行修正;根据瓦斯含量和瓦斯压力之间的关系,利用间接法反算煤层瓦斯压力。煤与瓦斯突出与其各个因素之间存在着复杂的非线性关系,本文分析支持向量机预测煤与瓦斯突出的可行性,引入粒子群优化算法(PSO),利用粒子群算法对支持向量机参数进行优化,建立开拓前煤与瓦斯突出PSO-SVM预测模型。收集同属一个地质单元的邻近矿井突出样本作为训练样本,吕梁某矿地勘修正后数据作为测试样本,利用MATLAB编写煤与瓦斯突出预测的PSO-SVM算法程序对其9号煤层进行煤与瓦斯突出预测,与单项指标法和综合指标D与K法预测结果一致。基于粒子群优化的支持向量机用在开拓前煤与瓦斯突出危险预测中是可行的,为矿井开拓前煤与瓦斯突出预测提供方向。
[Abstract]:Coal and gas outburst is one of the most serious natural disasters threatening coal mine safety. At present, the regional prediction of coal and gas outburst in China is mainly focused on the underground roadway after development, the point prediction based on the measured gas parameters, and the difficult problem of prediction of coal and gas outburst danger encountered in the field engineering practice before the development of coal and gas outburst. There has been no practical solution. Support vector machine (SVM), as a pattern recognition method which can solve many practical problems such as small sample, nonlinear and high dimension, has been widely used in many production fields. In order to improve the accuracy of coal and gas outburst prediction, this paper introduces support vector machine to establish a learning model to predict the outburst danger before the development of coal mine. Coal seam No. 9 in a certain mine of Lv Liang, Shanxi Province, has not been exposed, and there is no condition of underground gas parameter measurement. The law of gas occurrence in coal mine is studied, and the gas parameters measured in geological prospecting period of mine are collected and analyzed. The gas content in coal mine adjacent to the same geological unit is measured by the method of coal cuttings desorption by borehole, and the gas content in geological prospecting is corrected. According to the relationship between gas content and gas pressure, the indirect method is used to calculate the gas pressure in coal seam. There is a complex nonlinear relationship between coal and gas outburst and its various factors. This paper analyzes the feasibility of predicting coal and gas outburst by support vector machine. The particle swarm optimization (PSO),) algorithm is introduced to optimize the parameters of support vector machine (SVM), and the PSO-SVM prediction model of coal and gas outburst is established. The outburst samples of adjacent mines belonging to the same geological unit were collected as training samples, and the revised geological prospecting data of a certain mine of Lv Liang were used as test samples. The PSO-SVM algorithm program of coal and gas outburst prediction is compiled by MATLAB to predict coal and gas outburst in coal seam No. 9, which is consistent with the prediction results of single index method and synthetic index D and K method. The application of support vector machine based on particle swarm optimization is feasible in predicting the danger of coal and gas outburst before exploitation, which provides a direction for the prediction of coal and gas outburst before mine development.
【学位授予单位】：华北科技学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD713.2

【参考文献】