期刊文献+
共找到4篇文章
< 1 >
每页显示 20 50 100
采空区煤自燃预测指标及其预警值的研究 被引量:18
1
作者 刘志强 《煤矿开采》 北大核心 2013年第2期99-102,63,共5页
通过探索崔家寨矿采空区自然发火指标气体变化规律,研究得出自燃预测预报的关键参数,确定了自然发火预测预报指标及其临界值,并在实际中验证了研究的正确性,进行了有效地预测预报,有效减少和控制煤层自燃火灾的发生,使煤炭生产安全可靠... 通过探索崔家寨矿采空区自然发火指标气体变化规律,研究得出自燃预测预报的关键参数,确定了自然发火预测预报指标及其临界值,并在实际中验证了研究的正确性,进行了有效地预测预报,有效减少和控制煤层自燃火灾的发生,使煤炭生产安全可靠,提高煤矿经济效益。 展开更多
关键词 煤自燃指标 预警值 临界值
在线阅读 下载PDF
基于随机森林算法的煤自燃温度预测模型研究 被引量:25
2
作者 郑学召 李梦涵 +2 位作者 张嬿妮 姜鹏 王宝元 《工矿自动化》 北大核心 2021年第5期58-64,共7页
针对传统煤自燃温度预测模型预测精度较差、基于支持向量机(SVM)的预测模型对参数的选取要求较高和基于神经网络的预测模型测试时易出现过拟合的问题,提出了一种基于随机森林算法的煤自燃温度预测模型。利用煤自燃程序升温实验选取O2浓... 针对传统煤自燃温度预测模型预测精度较差、基于支持向量机(SVM)的预测模型对参数的选取要求较高和基于神经网络的预测模型测试时易出现过拟合的问题,提出了一种基于随机森林算法的煤自燃温度预测模型。利用煤自燃程序升温实验选取O2浓度、CO浓度、C2H4浓度、CO/ΔO2比值、C2H4/C2H6比值作为煤自燃预警指标数据,并对指标数据进行处理,将数据分为学习集和测试集;对学习集抽样形成决策树并按决策树最优特征分裂形成随机森林;采用均方误差值和判定系数(R2)优化随机森林算法的参数,进而构建随机森林模型;将测试集数据输入已训练好的随机森林模型,得到煤自燃温度预测结果。模型对比结果表明:与基于粒子群优化反向传播(PSO-BP)神经网络算法和基于SVM算法的煤自燃温度预测模型相比,随机森林测试阶段的R2为0.8697,PSO-BP测试阶段的R2为0.7836,SVM测试阶段的R2为0.8350,说明基于随机森林算法的煤自燃温度预测模型能够较为准确地对煤自燃温度进行预测,具有较强的鲁棒性和普适性,解决了基于PSO-BP神经网络算法的煤自燃温度预测模型和基于SVM算法的煤自燃温度预测模型容易出现过拟合的问题。 展开更多
关键词 自燃温度预测 随机森林 指标气体 自燃预警指标
在线阅读 下载PDF
Application of a novel detection approach based on non-dispersive infrared theory to the in-situ analysis on indicator gases from underground coal fire 被引量:12
3
作者 TIAN Fu-chao LIANG Yun-tao +2 位作者 ZHU Hong-qing CHEN Ming-yi WANG Jin-cheng 《Journal of Central South University》 SCIE EI CAS CSCD 2022年第6期1840-1855,共16页
Coal mine fires,which can cause heavy casualties,environmental damages and a waste of coal resources,have become a worldwide problem.Aiming at overcoming the drawbacks,such as a low analysis efficiency,poor stability ... Coal mine fires,which can cause heavy casualties,environmental damages and a waste of coal resources,have become a worldwide problem.Aiming at overcoming the drawbacks,such as a low analysis efficiency,poor stability and large monitoring error,of the existing underground coal fire monitoring technology,a novel monitoring system based on non-dispersive infrared(NDIR)spectroscopy is developed.In this study,first,the measurement principle of NDIR sensor,the gas concentration calculation and its temperature compensation algorithms were expounded.Next,taking CO and CH_(4) as examples,the liner correlation coefficients of absorbance and the temperature correction factors of the two indicator gases were calculated,and then the errors of concentration measurement for CO,CO_(2),CH_(4) and C_(2)H_(4) were further analyzed.The results disclose that the designed NDIR sensors can satisfy the requirements of industrial standards for monitoring the indicator gases for coal fire hazards.For the established NDIR-based monitoring system,the NDIRbased spectrum analyzer and its auxiliary equipment boast intrinsically safe and explosion-proof performances and can achieve real-time and in-situ detection of indicator gases when installed close to the coal fire risk area underground.Furthermore,a field application of the NDIR-based monitoring system in a coal mine shows that the NDIR-based spectrum analyzer has a permissible difference from the chromatography in measuring the concentrations of various indicator gases.Besides,the advantages of high accuracy,quick analysis and excellent security of the NDIR-based monitoring system have promoted its application in many coal mines. 展开更多
关键词 indicator gas coal spontaneous combustion infrared spectrometry bundle tube monitoring intrinsically safe
在线阅读 下载PDF
Experiment study of optimization on prediction index gases of coal spontaneous combustion 被引量:5
4
作者 NIU Hui-yong 《Journal of Central South University》 SCIE EI CAS CSCD 2016年第9期2321-2328,共8页
The coal of Anyuan Mine has the characteristic of easy spontaneous combustion. Conventional method is difficult to predict it. Coal samples from this mine were tested in laboratory. The data obtained from laboratory d... The coal of Anyuan Mine has the characteristic of easy spontaneous combustion. Conventional method is difficult to predict it. Coal samples from this mine were tested in laboratory. The data obtained from laboratory determination were initialized for the value which was defined as "K". The ratio of each index gas and value of "K", and the ratio of combination index gases and value of "K", were analyzed simultaneously. The research results show that for this coal mine, if there is carbon monoxide in the gas sample, the phenomenon of oxidation and temperature rising for coal exists in this mine; if there is C_2H_4 in the gas sample, the temperature of coal perhaps exceeds 130 °C. If the coal temperature is between 35 °C and 130 °C, prediction and forecast for coal spontaneous combustion depend on the value of Φ(CO)/K mainly; if the temperature of coal is between 130 °C and 300 °C, prediction and forecast for coal spontaneous combustion depend on the value of Φ(C_2H_6)/Φ(C_2H_2) and Φ(C_2H_6)/K. The research results provide experimental basis for the prediction of coal spontaneous combustion in Anyuan coal mine, and have better guidance on safe production of this coal mine. 展开更多
关键词 COAL spontaneous combustion index gases PREDICTION INITIALIZATION
在线阅读 下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部