矿区地表沉陷预测的KPCA-LSSVM模型
|
摘要
针对矿山地下开采引起的地表沉降问题,考虑到影响地表下沉量的多元因素,将核主成分分析法(Kernel principal component analysis,KPCA)与最小二乘支持向量机(Least squares support vector macine,LSSVM)相结合,构建了矿区地表沉陷预测的KPCA-LSSVM模型。该模型首先采用KPCA法对地表沉陷的影响因素进行分析,然后基于LSSVM理论,根据确定的主成分因子,构建了矿区地表沉陷预测模型。研究表明:(1)煤层赋存深度、煤柱宽度、煤层倾角为影响矿区地表最大下沉量的主要因素;(2)通过将华北某矿区煤层赋存深度、煤柱宽度、煤层倾角作为自变量,地表最大沉陷量作为因变量,构建的矿区地表沉陷KPCA-LSSVM预测模型得出的最大沉陷量与实测值的绝对误差为0.006~0.009 m,远小于FLAC3D模拟值与实测值的误差(0.108~0.217 m),表明该模型可以对矿区地表沉陷进行高精度预测。
Aiming at the problem of surface subsidence caused by underground mining and considering the multiple factors that affect the mining subsidence,a new prediciton model of surface minign subsidence based on kernel principal component analysis(KPCA)and least squares support vector machine(LSSVM)is established.Firslty,the KPCA method ie used to analye the influence factors of surface mining subsidnece,then,based on LSSVM theory,the prediction model of surface mining subsidence is established based on the detremined pricipal component factors.The study results show that:(1)the depth of coal seam,width of coal pillar and dip angle of coal seam are the main factors that affect the maxiumum surface mining subsidence of mining area;(2)taking the depth of coal seam,width of coal pillar and dip angle of coal seam as the independent variables,the maxium surface mining subsidence as the dependent variables,the KPCA-LSSVM prediction model of the surface mining subsidence of the mine is established,the absolute error between the prediction values and the actual measured values is 0.006~0.009 m,whcih is far less than the one(0.108 ~0.217 m) between the FLAC3 Dnumerical simulation values and the actual measured values.The study results further indicated that the KPCA-LSSVM model established in this paper is help for improving the surface mining subsidence prediciton precise of mining area.
Aiming at the problem of surface subsidence caused by underground mining and considering the multiple factors that affect the mining subsidence,a new prediciton model of surface minign subsidence based on kernel principal component analysis(KPCA)and least squares support vector machine(LSSVM)is established.Firslty,the KPCA method ie used to analye the influence factors of surface mining subsidnece,then,based on LSSVM theory,the prediction model of surface mining subsidence is established based on the detremined pricipal component factors.The study results show that:(1)the depth of coal seam,width of coal pillar and dip angle of coal seam are the main factors that affect the maxiumum surface mining subsidence of mining area;(2)taking the depth of coal seam,width of coal pillar and dip angle of coal seam as the independent variables,the maxium surface mining subsidence as the dependent variables,the KPCA-LSSVM prediction model of the surface mining subsidence of the mine is established,the absolute error between the prediction values and the actual measured values is 0.006~0.009 m,whcih is far less than the one(0.108 ~0.217 m) between the FLAC3 Dnumerical simulation values and the actual measured values.The study results further indicated that the KPCA-LSSVM model established in this paper is help for improving the surface mining subsidence prediciton precise of mining area.
引文
[1]阎跃观,戴华阳,王忠武,等.急倾斜多煤层开采地表沉陷分区与围岩破坏机理---以木城涧煤矿大台井为例[J].中国矿业大学学报,2013,42(4):547-553.Yan Yueguan,Dai Huayang,Wang Zhongwu,et al.Ground subsidence zone and surroundingrock failure mechanism due to steep multiple coal seam mining:a case study at Muchengjian Datai Mine[J].Journal of China University of Mining and Technology,2013,42(4)547-553.
[2]胡丰,李云安,刘海鳌,等.采矿顶板最小安全厚度及地面沉陷变形研究[J].金属矿山,2013(8):117-119.Hu Feng,Li Yunan,Liu Hai Ao,et al.Study of minimum safe thickness and ground subsidence of the stope roof[J].Metal Mine,2013(8):117-119.
[3]彭林军,赵晓东,李术才.深部开采地表沉陷规律模拟研究[J].岩土力学,2011,32(6):1910-1914.Peng Linjun,Zhao Xiaodong,Li Shucai.Simulating research on rules of surface subsidence due to deep mining[J].Rock and Soil Mechanics,2011,32(6):1910-1914.
[4]郭健,查吕应,庞有超,等.基于小波分析的深基坑地表沉降预测研究[J].岩土工程学报,2014,36(S2):343-347.Guo Jian,Zha Lvying,Pang Youchao,et al.Prediction for ground settlement of deep excavations based on wavelet analysis[J].Chinese Journal of Geotechnical Engineering,2014,36(S2):343-347.
[5]王军保,刘新荣,李鹏,等.MMF模型在采空区地表沉降预测中的应用[J].煤炭学报,2012,37(3):411-415.Wang Junbao,Liu Xinrong,Li Peng,et al.Study on prediction of surface subsidence inmined-out region with the MMF model[J].Journal of China Coal Society,2012,37(3):411-415.
[6]贺桂成,丁德馨,刘永,等.衡山石膏矿老采空区地表沉陷的ANFIS预测[J].采矿与安全工程学报,2012,29(6):876-881.He Guicheng,Ding Dexin,Liu Yong,et al.ANFIS prediction of surface subsidence of the old goaf of the gypsum mine in Hengshan[J].Journal of Mining&Safety Engineering,2012,29(6):876-881.
[7]王军保,刘新荣,刘小军.开采沉陷动态预测模型[J].煤炭学报,2015,40(3):516-521.Wang Junbao,Liu Xinrong,Liu Xiaojun.Dynamic prediction model of mining subsidence[J].Chinese Journal of Coal Science,2015,40(3):516-521.
[8]张兵,崔希民.开采沉陷动态预计的分段Knothe时间函数模型优化[J].岩土力学,2017,38(2):541-548.Zhang Bing,Cui Ximin.Optimization of segmented Knothe timefunction modelfor dynamic prediction of mining subsidence[J].Rock and Soil Mechanics,2017,38(2):541-548.
[9]刘玉成,曹树刚,刘延保.可描述地表沉陷动态过程的时间函数模型探讨[J].岩土力学,2010,31(3):925-931.Liu Yucheng,Cao Shugang,Liu Yanbao.Discussion on some time functions for describing dynamic course of surface subsidence due to mining[J].Rock and Soil Mechanics,2010,31(3):925-931.
[10]蒋建平,陈功奇,章杨松.偏最小二乘回归在地表沉陷预测中的应用[J].重庆大学学报,2010,33(9):92-97.Jiang Jianping,Chen Gongqi,Zhang Yangsong.Application of partial least-squares regression in the bforecast of ground subsidence[J].Journal of Chongqing University,2010,33(9):92-97.
[11]彭帅英,李广杰,彭文,等.基于改进遗传算法的Holt-Winters模型在采空沉陷预测中的应用[J].吉林大学学报:地球科学版,2013,43(2):515-520.Peng Shuaiying,Liang Guangjie,Peng Wen,et al.Mining subsidence forecast method basedon improved genetic algorithm and HoltWinters mode[lJ].Journal of Jilin University:Earth Science Edition,2013,43(2):515-520.
[12]郭启琛,李文平,王启庆,等.薄基岩浅埋深矿区地表开采沉降预测及影响因素分析[J].金属矿山,2015(12):161-164.Guo Qichen,Li Wenping,Wang Qiqing,et al.Prediction and analysis to affecting factors of surface mining subsidence in the mining area of shallow coal seam with thin bedrock[J].Metal Mine,2015(12):161-164.
[13]李春意,高永格,崔希民.基于正态分布时间函数地表动态沉陷预测研究[J].岩土力学,2016,37(S1):108-116.Li Chunyi,Gao Yongge,Cui Ximin.Progressive subsidence prediction of ground surface based on the normal distribution time function[J].Rock and Soil Mechanics,2016,37(S1):108-116.
[14]王宁,吴侃,刘锦,等.基于Boltzmann函数的开采沉陷预测模型[J].煤炭学报,2013,38(8):1352-1356.Wang Ning,Wu Kan,Liu Jin,et al.Model for mining subsidence prediction based on Boltzmann function[J].Journal of the China Coal Society,2013,38(8):1352-1356.
[15]孟万利,蔡来良,杨望山,等.矿区开采沉陷预计的Geomagic法[J].金属矿山,2017(1):126-131.Meng Wanli,Cai Lailiang,Yang Wangshan,et al.Mining subsidence prediction method based on Geomagic[J].Metal Mine,2017(1):126-131.
[16]谭鹏,曹平.基于灰色关联支持向量机的地表沉降预测[J].中南大学学报:自然科学版,2012,43(2):632-637.Tan Peng,Cao Ping.Predicting surface settlement of tunnel using grey relational-support vector machine[J].Journal of Central South University:Science and Technology Edition,2012,43(2):632-637.
[17]王正帅,邓喀中.老采空区残余沉降的离散灰色预测模型[J].煤炭学报,2010,35(7):1084-1088.Wang Zhengshuai,Deng Kazhong.Discrete grey prediction model of old goaf residual subsidence[J].Journal of China Coal Society,2010,35(7):1084-1088.
[18]王正帅,邓喀中.采动区地表动态沉降预测的Richards模型[J].岩土力学,2011,32(6):1664-1668.Wang Zhengshuai,Deng Kazhong.Richards model of surfacedynamic subsidence prediction in mining area[J].Rock and Soil Mechanics,2011,32(6):1664-1668.
[19]张子月,邹友峰,陈俊杰,等.采煤塌陷区土地动态沉降预测模型[J].农业工程学报,2016,32(21):246-251.Zhang Ziyue,Zou Youfeng,Chen Junjie,et al.Prediction model of land dynamic settlement in coal mining subsidence area[J].Transactions of the Chinese Society of Agricultural Engineering,2016,32(21):246-251.
[20]葛新民,范卓颖,范宜仁,等.基于核主成分-小波能谱分析的复杂储层油水界面预测[J].中南大学学报:自然科学版,2015,46(5):1747-1753.Ge Xinmin,Fan Zhuoying,Fan Yiren,et al.Oil/water contact prediction of complex reservoir using kernel principal component analysis and wavelet power spectrum analysis[J].Journal of Central South University:Science and Technology Edition,2015,46(5):1747-1753.
[21]陈斌,陆从德,刘光鼎.基于核主成分分析的时间域航空电磁去噪方法[J].地球物理学报,2014,57(1):295-302.Chen Bin,Lu Congde,Liu Guangding.A denoising method based on kernel principal component analysis for airborne time domain electromagnetic data[J].Chinese Journal of Geophysics,2014,57(1):295-302.
[22]向小东,宋芳.基于核主成分与加权支持向量机的福建省城镇登记失业率预测[J].系统工程理论与实践,2009,29(1):73-80.Xiang Xiaodong,Song Fang.Forecasting of the urban registered unemployment rate in Fujian province based on kernel principal component analysis and weighted support vector machine[J].Systems Engineering-Theory&Practice,2009,29(1):73-80.
[23]龚承柱,李兰兰,杨娟,等.基于EMD-PSR-LSSVM的城市燃气管网短期负荷预测[J].系统工程理论与实践,2014,34(11):3001-3008.Gong Chengzhu,Li Lanlan,Yang Juan,et al.An integrated short-term load forecasting approach for urban gas pipeline network based on EMD,PSR and LSSVM[J].Systems Engineering-Theory&Practice,2014,34(11):3001-3008.
[24]王新,孟玲玲.基于EEMD-LSSVM的超短期负荷预测[J].电力系统保护与控制,2015,43(1):61-66.Wang Xin,Meng Lingling.Ultra-short-term load forecasting based on EEMD-LSSVM[J].Power System Protection and Control,2015,43(1):61-66.
[25]龙文,梁昔明,龙祖强,等.PSO-LSSVM灰色组合模型在地下水埋深预测中的应用[J].系统工程理论与实践,2013,33(1):243-248.Long Wen,Liang Ximing,Long Zuqiang,et al.LSSVM gray combined forecasting model based on PSO and its application in groundwater dynamic prediction[J].Systems Engineering-Theory&Practice,2013,33(1):243-248.
[26]曹净,丁文云,赵党书,等.基于LSSVM-ARMA模型的基坑变形时间序列预测[J].岩土力学,2014,35(S2):579-586.Cao Jing,Ding Wenyun,Zhao Dangshu,et al.Time series forecast of foundation pit deformation based on LSSVM-ARMA model[J].Rock and Soil Mechanics,2014,35(S2):579-586.
[27]陈健峰.矿山深部开采覆岩沉陷预测的改进粒子群优化BP模型[J].金属矿山,2017(7):128-132.Chen Jianfeng.Overlying strata subsidence prediction of deep mining based on improved particle swarm optimization BP model[J].Metal Mine,2017(7):128-132.
[28]侯得峰,李德海,许国胜,等.厚松散层下采高对地表动态沉降特征的影响[J].煤炭科学技术,2016,44(12):191-196.Hou Defeng,Li Dehai,Xu Guosheng,et al.Impact of mining thickness on dynamic subsidence characteristics in condition of mining under thick unconsolidated layers[J].Coal Science and Technology,2016,44(12):191-196.
[2]胡丰,李云安,刘海鳌,等.采矿顶板最小安全厚度及地面沉陷变形研究[J].金属矿山,2013(8):117-119.Hu Feng,Li Yunan,Liu Hai Ao,et al.Study of minimum safe thickness and ground subsidence of the stope roof[J].Metal Mine,2013(8):117-119.
[3]彭林军,赵晓东,李术才.深部开采地表沉陷规律模拟研究[J].岩土力学,2011,32(6):1910-1914.Peng Linjun,Zhao Xiaodong,Li Shucai.Simulating research on rules of surface subsidence due to deep mining[J].Rock and Soil Mechanics,2011,32(6):1910-1914.
[4]郭健,查吕应,庞有超,等.基于小波分析的深基坑地表沉降预测研究[J].岩土工程学报,2014,36(S2):343-347.Guo Jian,Zha Lvying,Pang Youchao,et al.Prediction for ground settlement of deep excavations based on wavelet analysis[J].Chinese Journal of Geotechnical Engineering,2014,36(S2):343-347.
[5]王军保,刘新荣,李鹏,等.MMF模型在采空区地表沉降预测中的应用[J].煤炭学报,2012,37(3):411-415.Wang Junbao,Liu Xinrong,Li Peng,et al.Study on prediction of surface subsidence inmined-out region with the MMF model[J].Journal of China Coal Society,2012,37(3):411-415.
[6]贺桂成,丁德馨,刘永,等.衡山石膏矿老采空区地表沉陷的ANFIS预测[J].采矿与安全工程学报,2012,29(6):876-881.He Guicheng,Ding Dexin,Liu Yong,et al.ANFIS prediction of surface subsidence of the old goaf of the gypsum mine in Hengshan[J].Journal of Mining&Safety Engineering,2012,29(6):876-881.
[7]王军保,刘新荣,刘小军.开采沉陷动态预测模型[J].煤炭学报,2015,40(3):516-521.Wang Junbao,Liu Xinrong,Liu Xiaojun.Dynamic prediction model of mining subsidence[J].Chinese Journal of Coal Science,2015,40(3):516-521.
[8]张兵,崔希民.开采沉陷动态预计的分段Knothe时间函数模型优化[J].岩土力学,2017,38(2):541-548.Zhang Bing,Cui Ximin.Optimization of segmented Knothe timefunction modelfor dynamic prediction of mining subsidence[J].Rock and Soil Mechanics,2017,38(2):541-548.
[9]刘玉成,曹树刚,刘延保.可描述地表沉陷动态过程的时间函数模型探讨[J].岩土力学,2010,31(3):925-931.Liu Yucheng,Cao Shugang,Liu Yanbao.Discussion on some time functions for describing dynamic course of surface subsidence due to mining[J].Rock and Soil Mechanics,2010,31(3):925-931.
[10]蒋建平,陈功奇,章杨松.偏最小二乘回归在地表沉陷预测中的应用[J].重庆大学学报,2010,33(9):92-97.Jiang Jianping,Chen Gongqi,Zhang Yangsong.Application of partial least-squares regression in the bforecast of ground subsidence[J].Journal of Chongqing University,2010,33(9):92-97.
[11]彭帅英,李广杰,彭文,等.基于改进遗传算法的Holt-Winters模型在采空沉陷预测中的应用[J].吉林大学学报:地球科学版,2013,43(2):515-520.Peng Shuaiying,Liang Guangjie,Peng Wen,et al.Mining subsidence forecast method basedon improved genetic algorithm and HoltWinters mode[lJ].Journal of Jilin University:Earth Science Edition,2013,43(2):515-520.
[12]郭启琛,李文平,王启庆,等.薄基岩浅埋深矿区地表开采沉降预测及影响因素分析[J].金属矿山,2015(12):161-164.Guo Qichen,Li Wenping,Wang Qiqing,et al.Prediction and analysis to affecting factors of surface mining subsidence in the mining area of shallow coal seam with thin bedrock[J].Metal Mine,2015(12):161-164.
[13]李春意,高永格,崔希民.基于正态分布时间函数地表动态沉陷预测研究[J].岩土力学,2016,37(S1):108-116.Li Chunyi,Gao Yongge,Cui Ximin.Progressive subsidence prediction of ground surface based on the normal distribution time function[J].Rock and Soil Mechanics,2016,37(S1):108-116.
[14]王宁,吴侃,刘锦,等.基于Boltzmann函数的开采沉陷预测模型[J].煤炭学报,2013,38(8):1352-1356.Wang Ning,Wu Kan,Liu Jin,et al.Model for mining subsidence prediction based on Boltzmann function[J].Journal of the China Coal Society,2013,38(8):1352-1356.
[15]孟万利,蔡来良,杨望山,等.矿区开采沉陷预计的Geomagic法[J].金属矿山,2017(1):126-131.Meng Wanli,Cai Lailiang,Yang Wangshan,et al.Mining subsidence prediction method based on Geomagic[J].Metal Mine,2017(1):126-131.
[16]谭鹏,曹平.基于灰色关联支持向量机的地表沉降预测[J].中南大学学报:自然科学版,2012,43(2):632-637.Tan Peng,Cao Ping.Predicting surface settlement of tunnel using grey relational-support vector machine[J].Journal of Central South University:Science and Technology Edition,2012,43(2):632-637.
[17]王正帅,邓喀中.老采空区残余沉降的离散灰色预测模型[J].煤炭学报,2010,35(7):1084-1088.Wang Zhengshuai,Deng Kazhong.Discrete grey prediction model of old goaf residual subsidence[J].Journal of China Coal Society,2010,35(7):1084-1088.
[18]王正帅,邓喀中.采动区地表动态沉降预测的Richards模型[J].岩土力学,2011,32(6):1664-1668.Wang Zhengshuai,Deng Kazhong.Richards model of surfacedynamic subsidence prediction in mining area[J].Rock and Soil Mechanics,2011,32(6):1664-1668.
[19]张子月,邹友峰,陈俊杰,等.采煤塌陷区土地动态沉降预测模型[J].农业工程学报,2016,32(21):246-251.Zhang Ziyue,Zou Youfeng,Chen Junjie,et al.Prediction model of land dynamic settlement in coal mining subsidence area[J].Transactions of the Chinese Society of Agricultural Engineering,2016,32(21):246-251.
[20]葛新民,范卓颖,范宜仁,等.基于核主成分-小波能谱分析的复杂储层油水界面预测[J].中南大学学报:自然科学版,2015,46(5):1747-1753.Ge Xinmin,Fan Zhuoying,Fan Yiren,et al.Oil/water contact prediction of complex reservoir using kernel principal component analysis and wavelet power spectrum analysis[J].Journal of Central South University:Science and Technology Edition,2015,46(5):1747-1753.
[21]陈斌,陆从德,刘光鼎.基于核主成分分析的时间域航空电磁去噪方法[J].地球物理学报,2014,57(1):295-302.Chen Bin,Lu Congde,Liu Guangding.A denoising method based on kernel principal component analysis for airborne time domain electromagnetic data[J].Chinese Journal of Geophysics,2014,57(1):295-302.
[22]向小东,宋芳.基于核主成分与加权支持向量机的福建省城镇登记失业率预测[J].系统工程理论与实践,2009,29(1):73-80.Xiang Xiaodong,Song Fang.Forecasting of the urban registered unemployment rate in Fujian province based on kernel principal component analysis and weighted support vector machine[J].Systems Engineering-Theory&Practice,2009,29(1):73-80.
[23]龚承柱,李兰兰,杨娟,等.基于EMD-PSR-LSSVM的城市燃气管网短期负荷预测[J].系统工程理论与实践,2014,34(11):3001-3008.Gong Chengzhu,Li Lanlan,Yang Juan,et al.An integrated short-term load forecasting approach for urban gas pipeline network based on EMD,PSR and LSSVM[J].Systems Engineering-Theory&Practice,2014,34(11):3001-3008.
[24]王新,孟玲玲.基于EEMD-LSSVM的超短期负荷预测[J].电力系统保护与控制,2015,43(1):61-66.Wang Xin,Meng Lingling.Ultra-short-term load forecasting based on EEMD-LSSVM[J].Power System Protection and Control,2015,43(1):61-66.
[25]龙文,梁昔明,龙祖强,等.PSO-LSSVM灰色组合模型在地下水埋深预测中的应用[J].系统工程理论与实践,2013,33(1):243-248.Long Wen,Liang Ximing,Long Zuqiang,et al.LSSVM gray combined forecasting model based on PSO and its application in groundwater dynamic prediction[J].Systems Engineering-Theory&Practice,2013,33(1):243-248.
[26]曹净,丁文云,赵党书,等.基于LSSVM-ARMA模型的基坑变形时间序列预测[J].岩土力学,2014,35(S2):579-586.Cao Jing,Ding Wenyun,Zhao Dangshu,et al.Time series forecast of foundation pit deformation based on LSSVM-ARMA model[J].Rock and Soil Mechanics,2014,35(S2):579-586.
[27]陈健峰.矿山深部开采覆岩沉陷预测的改进粒子群优化BP模型[J].金属矿山,2017(7):128-132.Chen Jianfeng.Overlying strata subsidence prediction of deep mining based on improved particle swarm optimization BP model[J].Metal Mine,2017(7):128-132.
[28]侯得峰,李德海,许国胜,等.厚松散层下采高对地表动态沉降特征的影响[J].煤炭科学技术,2016,44(12):191-196.Hou Defeng,Li Dehai,Xu Guosheng,et al.Impact of mining thickness on dynamic subsidence characteristics in condition of mining under thick unconsolidated layers[J].Coal Science and Technology,2016,44(12):191-196.