煤矿井下多参数突水信息的动态评价方法及系统设计
|
摘要
煤炭行业是诸多能源行业当中的高危行业,其开采一直受到很多条件的限制,其中地下水是威胁煤矿安全生产的重要因素之一。最近,多起重大、特大煤矿突水事件屡见报端,主要是因为煤矿突水形成十分复杂,涉及到水文、地质构造、采动影响等方面的要素,而目前又缺乏在采掘过程中的实时动态监测,因此给人民的生命财产造成了极大的危害。
针对煤矿开采过程中缺乏有效的动态监测手段、预测方法单一等缺点,研究神经网络和D-S证据理论相结合的两级融合算法,提出了煤矿井下多参数突水信息的动态评价方法。此方法将传感器技术、数据融合技术和自动控制技术相结合,实现了煤矿井下传感器数据的动态采集、实时传输及突水等级的确定,为预防水害的发生提供决策依据。
目前关于突水的数据还很少,而突水预测需要大量的数据。因此,提出了在采掘工作面区域内进行分布式多方位传感器布置的方案。首先确定影响工作面突水的诸多因素,通过分析煤矿顶板、底板突水机理,得出影响突水的主要因素,如巷道导水沟水位、含水层水位、顶、底板压力和掘进面前方的水量等。其次,选取了山西省漳村、沙坪两个煤矿进行实例分析,根据两个煤矿不同的水文、地质结构条件,对应的进行了传感器的布置,从而进行水情动态多参数测量。
提出了建立井下现场动态水情数据库,填补了煤矿井下动态水情信息数据的空白。系统以VC++作为软件开发平台,Microsoft SQL Server为数据库平台,采用类三层C/S结构的开发模式,结合以太网标准PROFINET通讯技术,实现对煤矿井下的压力、水位、温度、流量等数据的实时传输,将采集到的数据预处理、传输到井下变电所的防爆PC机整理、分类、存储。然后,将采集到的数据进一步分析,应用两级融合算法,计算得出关于煤矿井下工作面突水的安全等级状态。
对山西漳村煤矿3#煤层进行了实例分析,从开采、探测资料中提取10组已采点作为构建预测模型的样本,运用两级融合算法对煤矿井下突水等级评估。结果表明,该两级融合算法能够准确评估出煤矿井下突水的程度,有较强的抗干扰能力和较高的准确度,从而验证了系统的有效性和可行性。
The coal industry is one of the most dangerous energy industries, its exploitation has been subjected to many conditions, the threat of groundwater which is an important factor in coal mine production safety. Recently, several important, large water inrush events are often found in newspapers. The water inrush of coal mines causes by some complex elements, involving hydrology, geology, mining effect etc.. At present, it is short of real-time dynamic monitoring in the excavating process, which jeopardized to people’s lives and wealth.
According to the problems of the lack of real-time monitoring and single detection method during the excavating process, combined with neural network and the D-S evidence theory fusion algorithm, it is proposed that the predictive method of coal mining water inrush based on multi-parameter is studied. This method is the integration of sensing technology, data fusion technology and automation technology. It achieves the aim of the mine sensor data collection in time, real-time transmission and determination of the water inrush level ,to provide the basis for preventing water damage.
At present,the datas about water inrush are still short, while large amounts of datas are required in water inrush prediction. The method is proposed that distributed multi-directional sensors are arranged in the mining working face. Firstly,many factors affecting workface inrush are determined.Through analysis of the mechanism of coal top and bottom board inrush, the main factors are obtained,such as roadway gutter water level, aquifer water level, top, bottom board pressure and the amount of water before driving side and so on.Secondly, the ShaPing and ZhangCun coal in Shanxi Province are analyzed, according to two different hydrological, geological structure conditions, correspondingly arranging the sensors to measure the dynamic parameter.
The dynamic water regimen database is established, filling the blank of the dynamic hydrologic data. The system takes VC++ as a software development platform and choose Microsoft SQL Server as the database system,adopting structure of C/S as development model,combing with the Ethernet standard PROFINET Communication technology. It realizes the real-time data transimission about the pressure on the coal mine,water level,temperature,flow.Then the collected datas are pretreated and transmitted to the underground substation flameproof computer putting into the database. At the same time, the collected datas are analyzed further, using the fusion algorithm, and finally the safety level about coal water inrush is reached.
It is made a case study which used the data of 3 # coal in ZhangCun of Shanxi Province,10 points has been taken as the group sample for building prediction model from the mining and exploration data.And the fusion algorithm is used to evaluate water inrush level in mining.The results show that the fusion algorithm can accurately assess the situation of underground mine water inrush ,it has anti-jamming capability and high accuracy and it’s proofed that the system is effective and feasible.
针对煤矿开采过程中缺乏有效的动态监测手段、预测方法单一等缺点,研究神经网络和D-S证据理论相结合的两级融合算法,提出了煤矿井下多参数突水信息的动态评价方法。此方法将传感器技术、数据融合技术和自动控制技术相结合,实现了煤矿井下传感器数据的动态采集、实时传输及突水等级的确定,为预防水害的发生提供决策依据。
目前关于突水的数据还很少,而突水预测需要大量的数据。因此,提出了在采掘工作面区域内进行分布式多方位传感器布置的方案。首先确定影响工作面突水的诸多因素,通过分析煤矿顶板、底板突水机理,得出影响突水的主要因素,如巷道导水沟水位、含水层水位、顶、底板压力和掘进面前方的水量等。其次,选取了山西省漳村、沙坪两个煤矿进行实例分析,根据两个煤矿不同的水文、地质结构条件,对应的进行了传感器的布置,从而进行水情动态多参数测量。
提出了建立井下现场动态水情数据库,填补了煤矿井下动态水情信息数据的空白。系统以VC++作为软件开发平台,Microsoft SQL Server为数据库平台,采用类三层C/S结构的开发模式,结合以太网标准PROFINET通讯技术,实现对煤矿井下的压力、水位、温度、流量等数据的实时传输,将采集到的数据预处理、传输到井下变电所的防爆PC机整理、分类、存储。然后,将采集到的数据进一步分析,应用两级融合算法,计算得出关于煤矿井下工作面突水的安全等级状态。
对山西漳村煤矿3#煤层进行了实例分析,从开采、探测资料中提取10组已采点作为构建预测模型的样本,运用两级融合算法对煤矿井下突水等级评估。结果表明,该两级融合算法能够准确评估出煤矿井下突水的程度,有较强的抗干扰能力和较高的准确度,从而验证了系统的有效性和可行性。
The coal industry is one of the most dangerous energy industries, its exploitation has been subjected to many conditions, the threat of groundwater which is an important factor in coal mine production safety. Recently, several important, large water inrush events are often found in newspapers. The water inrush of coal mines causes by some complex elements, involving hydrology, geology, mining effect etc.. At present, it is short of real-time dynamic monitoring in the excavating process, which jeopardized to people’s lives and wealth.
According to the problems of the lack of real-time monitoring and single detection method during the excavating process, combined with neural network and the D-S evidence theory fusion algorithm, it is proposed that the predictive method of coal mining water inrush based on multi-parameter is studied. This method is the integration of sensing technology, data fusion technology and automation technology. It achieves the aim of the mine sensor data collection in time, real-time transmission and determination of the water inrush level ,to provide the basis for preventing water damage.
At present,the datas about water inrush are still short, while large amounts of datas are required in water inrush prediction. The method is proposed that distributed multi-directional sensors are arranged in the mining working face. Firstly,many factors affecting workface inrush are determined.Through analysis of the mechanism of coal top and bottom board inrush, the main factors are obtained,such as roadway gutter water level, aquifer water level, top, bottom board pressure and the amount of water before driving side and so on.Secondly, the ShaPing and ZhangCun coal in Shanxi Province are analyzed, according to two different hydrological, geological structure conditions, correspondingly arranging the sensors to measure the dynamic parameter.
The dynamic water regimen database is established, filling the blank of the dynamic hydrologic data. The system takes VC++ as a software development platform and choose Microsoft SQL Server as the database system,adopting structure of C/S as development model,combing with the Ethernet standard PROFINET Communication technology. It realizes the real-time data transimission about the pressure on the coal mine,water level,temperature,flow.Then the collected datas are pretreated and transmitted to the underground substation flameproof computer putting into the database. At the same time, the collected datas are analyzed further, using the fusion algorithm, and finally the safety level about coal water inrush is reached.
It is made a case study which used the data of 3 # coal in ZhangCun of Shanxi Province,10 points has been taken as the group sample for building prediction model from the mining and exploration data.And the fusion algorithm is used to evaluate water inrush level in mining.The results show that the fusion algorithm can accurately assess the situation of underground mine water inrush ,it has anti-jamming capability and high accuracy and it’s proofed that the system is effective and feasible.
引文
[1]张英梅,汤红颖.基于LabView的煤矿井下排水监控系统的设计[J].煤炭科学技术,2008,36(2):68-70.
[2]靳德武.我国煤层底板突水问题的研究现状及展望[J].煤炭科学技术,2002:30(6):1-4.
[3]靳德武.华北型煤田煤层底板突水预测信息分析理论、方法及应用[J].中国岩溶,2003,22(1):41-46.
[4]王延福,靳德斌,曾艳京.矿井煤层底板突水预测新方法研究[J].水文地质工程地质,1999(4):33-37.
[5]毛红川,刘志,邓春涛等.煤矿突水预测方法探讨[J].河北工业科技,2008,25(4):197-199.
[6]吴迪,张涛,刘广亮等.瑞利波勘探技术在煤矿陷落柱探查中的应用[J] .中州煤炭,2009(10):50-52.
[7]于景邨.矿井瞬变电磁法勘探[M].徐州:中国矿业大学出版社,2007.
[8]王扬州,于景邨,刘建等.瞬变电磁法矿井超前探测[J].工程地理物理学报,2009,6(1):29-32.
[9]何友,王国宏.多传感器信息融合及应用[J].电子工业出版社,2000(7):35-40.
[10] Bar-Shalom Y,Fortmann TE.Tracking and Data Association,NewYork,Academic Press Inc ,1988.
[11] Bar-Shalom Y,Extension of the Probabilistic Data Association Filter in Multi-target Tracking,Proceedings of the 5' 0Symp .on Nonlinear Estimation,1974.
[12] Reid D B ,An Algoirthm for Tracking Multiple Targets,IEEE Trans,Automatic Control,1978.
[13]罗志增,蒋静坪.基于D-S理论的多信息融合方法及应用[J].电子学报,1999,27(9):100-102.
[14]肖斌.多传感器信息融合及其在工业中的应用[D].太原:太原理工大学,2009.
[15]藏大进,严宏凤,王跃才.多传感器信息融合技术综述[J].工矿自动化,2006(06): 29-32. 72
[16] Nageswara S V R,David BR,Jacob B.Fusion method for Physical systems based On Physical laws[C] . Proceedings of 2000 International Conference on Information Fusion.France:Paris,2000:89-95.
[17] Le Moinge J,Smith J.Image registration and fusion in remote sensing for NASA [C]proceedings of 2000 International Conference on Information Fusion.France:Paris,2000:375-382.
[18]程珍珍.多传感器数据融合技术在煤矿水害预防中的应用[D].太原:太原理工大学,2009.
[19]姜成志,张绍兵.建立在神经网络基础上的煤矿突水评价模型[J].黑龙江科技学院学报,2006,116(1):9-11.
[20]吴景龙,王巨,王春山.人工神经网络系统在预防煤矿突水中应用[J].煤矿安全.2008.4:55-57.
[21]王怀洪.利用物探信息预测煤矿水害的方法探讨[J].山东煤炭科技,2006,4(12):50-51.
[22]吴慈生,白勤虎,王新元.基于模糊聚类分析的矿井突水预测及危险性评价[J] .中国安全科学学报,1995,5(12):80-82.
[23]李丽,程久龙.基于信息融合的矿井底板突水预测[J] .煤炭学报,2006,31(5):623-626.
[24]廖巍,周荣义,李树清.基于小波神经网络的煤层底板突水非线性预测方法研究[J].中国安全科学学报,2006,16(11):25-28.
[25]张心彬,程久龙,李丽等.多源信息融合技术在矿井奥灰突水预测中的应用[J].测绘科学,2006,31(6):146-148.
[26] Reboul S,Brige D,Benjelloun M.Optimal segmentation by random process Fusion[C].Proceedings of 2000 International Conference Information Fusion.France,Paris,2000:344-348.
[27]江东,王建华,陈佩佩等.基于神经网络的煤矿底板突水预测[J].灾害学,1999,14(1):28-32.
[28]张艳红,孙晓光.煤层底板突水预测及防治[J].陕西煤炭,2008,6,73-75.
[29] Tang X.Multiple Competitive Learning network fusion for object classification[J].IEEETransactions on System,Man,and Cybernetic,1998,28(4):532-543.
[30]徐建国,冯增强等.矿井防止水综合技术[M].北京:中国矿业大学出版社,2007:25-39.
[31]焦作矿业学院,阜新矿业学院编.采煤概论[M].北京:煤炭工业出版社,1978:197-199.
[32]付瑞锋.可视化多参数水文动态监测系统研究[D] .西安:西安科技大学,2008.
[33]王齐峰.多参数水文实时监测智能预警系统[D].西安:西安科技大学,2008.
[34]沈炜,徐慧.Visual C++数据库编程技术与实例[M] .北京:人民邮电出版社,2005:245-270.
[35]张英梅,程珍珍.D-S证据理论在煤矿水害预测中的应用[J].太原理工大学学报2008.11,39(6):589-591.
[36]董长虹.matlab神经网络与应用[M].北京:国防工业出版社,2005:65-85.
[37]程珍珍.多传感器数据融合技术在煤矿水害预防中的应用[D].太原:太原理工大学,2009.
[38] Deneeux T.An Evidenee-Theoretic Neural network Classifier[C].IEEE.Int. Conf.on Syst.,Man and Cybern.Americal:Newyork,1995,25(3):712-714.
[39] Yager R R.On the Dempster-Shafer Framework and New Combination Rules [J].Information Science,1998,4:93-138.
[40] T L Hunts Herger,et al.A framework for Multi sensor Fusion in the Presence of Uncertainty[C] . Proe . Of Workshop on Spatial Reasoning and Multi-sensor Fusion.St.CharlesⅢoct.1987:345-350.
[41] J Llines,et al.Data Fusion Technology forecast for C3MIS[C].Proe.of the 3rd International Conference on Command,Control,Communications and MIS.Bournemouth:UK,IEE,London,March 1989(16):521-525.
[42] EL Waltz,D M Buede.Data Fusion and Decision Support for Command and Control[J].IEEE Trans.On SMC,1986:865-879.
[43] S K Rogers.Et al.Artificial Neural Network Technology for Automation Target Recognition[J].SPIE Institute Series, Vol.157,1990:231-247.
[44] L.F.Pau.Fusion of Multi sensor Data in Pattern Recognition Theory andApplication[J].Journal of sensor Fusion,1982:189-201.
[45]徐田高.煤矿突水危险度研究[J].中国煤炭地质,2009,21(2):43-45.
[46]许学汉,王杰.煤矿突水预报研究[M].北京:地质出版社,1991:2-14.
[2]靳德武.我国煤层底板突水问题的研究现状及展望[J].煤炭科学技术,2002:30(6):1-4.
[3]靳德武.华北型煤田煤层底板突水预测信息分析理论、方法及应用[J].中国岩溶,2003,22(1):41-46.
[4]王延福,靳德斌,曾艳京.矿井煤层底板突水预测新方法研究[J].水文地质工程地质,1999(4):33-37.
[5]毛红川,刘志,邓春涛等.煤矿突水预测方法探讨[J].河北工业科技,2008,25(4):197-199.
[6]吴迪,张涛,刘广亮等.瑞利波勘探技术在煤矿陷落柱探查中的应用[J] .中州煤炭,2009(10):50-52.
[7]于景邨.矿井瞬变电磁法勘探[M].徐州:中国矿业大学出版社,2007.
[8]王扬州,于景邨,刘建等.瞬变电磁法矿井超前探测[J].工程地理物理学报,2009,6(1):29-32.
[9]何友,王国宏.多传感器信息融合及应用[J].电子工业出版社,2000(7):35-40.
[10] Bar-Shalom Y,Fortmann TE.Tracking and Data Association,NewYork,Academic Press Inc ,1988.
[11] Bar-Shalom Y,Extension of the Probabilistic Data Association Filter in Multi-target Tracking,Proceedings of the 5' 0Symp .on Nonlinear Estimation,1974.
[12] Reid D B ,An Algoirthm for Tracking Multiple Targets,IEEE Trans,Automatic Control,1978.
[13]罗志增,蒋静坪.基于D-S理论的多信息融合方法及应用[J].电子学报,1999,27(9):100-102.
[14]肖斌.多传感器信息融合及其在工业中的应用[D].太原:太原理工大学,2009.
[15]藏大进,严宏凤,王跃才.多传感器信息融合技术综述[J].工矿自动化,2006(06): 29-32. 72
[16] Nageswara S V R,David BR,Jacob B.Fusion method for Physical systems based On Physical laws[C] . Proceedings of 2000 International Conference on Information Fusion.France:Paris,2000:89-95.
[17] Le Moinge J,Smith J.Image registration and fusion in remote sensing for NASA [C]proceedings of 2000 International Conference on Information Fusion.France:Paris,2000:375-382.
[18]程珍珍.多传感器数据融合技术在煤矿水害预防中的应用[D].太原:太原理工大学,2009.
[19]姜成志,张绍兵.建立在神经网络基础上的煤矿突水评价模型[J].黑龙江科技学院学报,2006,116(1):9-11.
[20]吴景龙,王巨,王春山.人工神经网络系统在预防煤矿突水中应用[J].煤矿安全.2008.4:55-57.
[21]王怀洪.利用物探信息预测煤矿水害的方法探讨[J].山东煤炭科技,2006,4(12):50-51.
[22]吴慈生,白勤虎,王新元.基于模糊聚类分析的矿井突水预测及危险性评价[J] .中国安全科学学报,1995,5(12):80-82.
[23]李丽,程久龙.基于信息融合的矿井底板突水预测[J] .煤炭学报,2006,31(5):623-626.
[24]廖巍,周荣义,李树清.基于小波神经网络的煤层底板突水非线性预测方法研究[J].中国安全科学学报,2006,16(11):25-28.
[25]张心彬,程久龙,李丽等.多源信息融合技术在矿井奥灰突水预测中的应用[J].测绘科学,2006,31(6):146-148.
[26] Reboul S,Brige D,Benjelloun M.Optimal segmentation by random process Fusion[C].Proceedings of 2000 International Conference Information Fusion.France,Paris,2000:344-348.
[27]江东,王建华,陈佩佩等.基于神经网络的煤矿底板突水预测[J].灾害学,1999,14(1):28-32.
[28]张艳红,孙晓光.煤层底板突水预测及防治[J].陕西煤炭,2008,6,73-75.
[29] Tang X.Multiple Competitive Learning network fusion for object classification[J].IEEETransactions on System,Man,and Cybernetic,1998,28(4):532-543.
[30]徐建国,冯增强等.矿井防止水综合技术[M].北京:中国矿业大学出版社,2007:25-39.
[31]焦作矿业学院,阜新矿业学院编.采煤概论[M].北京:煤炭工业出版社,1978:197-199.
[32]付瑞锋.可视化多参数水文动态监测系统研究[D] .西安:西安科技大学,2008.
[33]王齐峰.多参数水文实时监测智能预警系统[D].西安:西安科技大学,2008.
[34]沈炜,徐慧.Visual C++数据库编程技术与实例[M] .北京:人民邮电出版社,2005:245-270.
[35]张英梅,程珍珍.D-S证据理论在煤矿水害预测中的应用[J].太原理工大学学报2008.11,39(6):589-591.
[36]董长虹.matlab神经网络与应用[M].北京:国防工业出版社,2005:65-85.
[37]程珍珍.多传感器数据融合技术在煤矿水害预防中的应用[D].太原:太原理工大学,2009.
[38] Deneeux T.An Evidenee-Theoretic Neural network Classifier[C].IEEE.Int. Conf.on Syst.,Man and Cybern.Americal:Newyork,1995,25(3):712-714.
[39] Yager R R.On the Dempster-Shafer Framework and New Combination Rules [J].Information Science,1998,4:93-138.
[40] T L Hunts Herger,et al.A framework for Multi sensor Fusion in the Presence of Uncertainty[C] . Proe . Of Workshop on Spatial Reasoning and Multi-sensor Fusion.St.CharlesⅢoct.1987:345-350.
[41] J Llines,et al.Data Fusion Technology forecast for C3MIS[C].Proe.of the 3rd International Conference on Command,Control,Communications and MIS.Bournemouth:UK,IEE,London,March 1989(16):521-525.
[42] EL Waltz,D M Buede.Data Fusion and Decision Support for Command and Control[J].IEEE Trans.On SMC,1986:865-879.
[43] S K Rogers.Et al.Artificial Neural Network Technology for Automation Target Recognition[J].SPIE Institute Series, Vol.157,1990:231-247.
[44] L.F.Pau.Fusion of Multi sensor Data in Pattern Recognition Theory andApplication[J].Journal of sensor Fusion,1982:189-201.
[45]徐田高.煤矿突水危险度研究[J].中国煤炭地质,2009,21(2):43-45.
[46]许学汉,王杰.煤矿突水预报研究[M].北京:地质出版社,1991:2-14.