半参数模型在开采沉陷预计中的应用  

Application of Semi-parametric Model in Prediction of Mining Subsidence

作　　者：成枢[1] 朱玉明 马卫骄 高秀明 刘忠伟 Cheng Shu;Zhu Yuming;Ma Weijiao;Gao Xiuming;Liu Zhongwei(College of Geomatics,Shandong University of Science and Technology,Qingdao 266590,China;Shangdong Provincial Nuclear Industry 248 Geological Brigade,Qingdao 266041,China;Wangjiazhai Coal Mine,Shangdong Taifeng Holding Group Co.,Ltd.,Tai'an 271204,China)

机构地区：[1]山东科技大学测绘科学与工程学院,山东青岛266590 [2]山东省核工业二四八地质大队,山东青岛266041 [3]山东泰丰控股集团有限公司王家寨煤矿,山东泰安271204

出　　处：《金属矿山》2020年第3期154-158,共5页Metal Mine

摘　　要：半参数模型作为发展于20世纪80年代用于参数研究的重要统计模型在许多领域得到了广泛应用。当观测值中存在无法消除的模型误差时,模型误差既无法与观测值建立函数关系,也不能将其归入参数模型误差项或采用非参数模型处理。半参数模型在平差模型中将这种模型误差采用非参数向量表示,使平差模型兼顾参数模型与非参数模型的特点。在处理模型误差不可忽略的测量数据时,半参数模型在数学模型方面更加贴近实际问题,在数值求解方面可以同时求解模型误差与偶然误差。针对半参数模型的理论,采用了矿区实测数据验证半参数模型在数据处理中的有效性与优越性。对矿区地表沉降实测数据建立了半参数灰色Verhulst模型,并设计了不同方案进行计算。试验结果表明:当观测量中模型误差不可忽略时,半参数模型的计算结果优于参数模型。As an important statistical model developed in 1980 s,semi-parametric model has been widely used in many fields.When there are irreconcilable model errors in the observed values,the model errors can neither establish a functional relationship with the observed values,nor can they be classified as parametric model errors or treated by non-parametric models.In the adjustment model of semi-parametric model,the model error is expressed as non-parametric vector,which makes the adjustment model take into account the characteristics of both parametric model and non-parametric model.When dealing with the measurement data whose model error cannot be ignored,semi-parametric model is closer to practical problems in mathematical model,and can solve both model error and accidental error in numerical solution.According to the theory of semi-parametric model,the validity and superiority of semi-parametric model in data processing are verified by the measured data of mining area.The semi-parametric grey Verhulst model was established to calculate the measured data of surface subsidence.The test results show that the semi-parametric model is better than the parametric model when the model error is not negligible.

关 键 词：开采沉陷 半参数模型 补偿最小二乘估计 正则矩阵 灰色VERHULST模型 

分 类 号：TD325[矿业工程—矿井建设]