面向透明工作面的地质建模插值误差分析  

Research on interpolation error analysis of geological modeling of intelligent working face

作　　者：安林 韩保山 李鹏 代振华[1,3] 王新苗 AN Lin;HAN Baoshan;LI Peng;DAI Zhenhua;WANG Xinmiao(Xi’an Research Institute Co.Ltd.,China Coal Technology and Engineering Group Corp.,Xi’an 710077,China;College of Geology and Environment,Xi’an University of Science and Technology,Xi’an 710054,China;China Coal Research Institute,Beijing 100013,China)

机构地区：[1]中煤科工集团西安研究院有限公司,陕西西安710077 [2]西安科技大学地质与环境学院,陕西西安710054 [3]煤炭科学研究总院,北京100013

出　　处：《煤田地质与勘探》2022年第6期184-189,共6页Coal Geology & Exploration

基　　金：天地科技有限公司科技创新创业资金专项重点项目(2019-TD-ZD003,2020-TD-ZD002)。

摘　　要：工作面建模一般通过三维地震、井巷和钻孔测量等探测手段获取工作面的实际展布情况,然后利用插值算法建立相应模型。其中,采样数据是工作面建模的基础,插值算法是工作面模型实现的必经途径。插值算法和采样数据量的大小在不同程度上影响工作面模型的精确性,定量研究工作面模型精确度影响因素将对插值算法优选和采样数据获取量提供重要的参考价值。在工作面探测资料的基础上,通过交叉验证的方法,分别计算对比函数插值、光滑离散插值(DSI插值)和克里金插值的插值误差。为了解决透明工作面建模采样量大的问题,提出相对间距误差,并计算得到13组不同采样比例时模型的相对间距误差。结果表明:(1)透明工作面模型构建过程中,DSI插值、克里金插值和函数插值的平均绝对误差分别为0.0155、0.0225、0.2312,因此,DSI插值算法构建的模型精确度最高,克里金插值算法次之,函数插值算法最差。(2)随着采样数据量的增加,模型的误差逐渐减小,当采样数据量小于10%时,插值误差下降幅度很大;但当采样数据量大于10%时,其下降幅度趋于平缓,建议在构建工作面模型时采样数据量大于10%。(3)在透明工作面模型实际构建过程中,建议采用DSI插值算法;同时根据最低采样数据量分析得到的最佳更新距离和最佳采样间距进行采样,提高工作面局部数据量。Working face modeling generally obtains the actual distribution of the working face through detection meth-ods such as three-dimensional seismic,tunnel and borehole surveys,and then it establishes the corresponding model us-ing interpolation algorithms.During the modeling process,sampling data are the foundation and interpolation is the ne-cessary way to realize the working face model.The interpolation algorithm and the amount of sampled data affect the ac-curacy of the face model to varying degrees.Quantitative research on the factors affecting the accuracy of the face mod-el will provide important reference value for the optimization of the interpolation algorithm and the amount of sampled data.On the basis of the detection data of the working face,firstly,the cross-validation method is used to calculate the interpolation errors of the contrast function interpolation,DSI interpolation and Kriging interpolation.Then,in order to solve the problem of large sampling amount of transparent working surface modeling,the relative spacing error is pro-posed,and the relative spacing error of the models with 13 groups of different sampling scales is calculated.The results show that:(1)The average absolute errors of DSI interpolation,Kriging interpolation and function interpolation are re-spectively in the process of constructing a transparent face model.They are 0.0155,0.0225,and 0.2312,so the model constructed according to the DSI interpolation algorithm has the highest accuracy,followed by the Kriging interpolation algorithm,and the function interpolation algorithm is the worst.(2)As the amount of sampled data increases,the error of the model gradually decreases.When the amount of sampled data is less than 10%,the interpolation error decreases greatly;but when the amount of sampled data is greater than 10%,the decrease tends to be gentle.It is recommended that the sample data volume could be greater than 10%for constructing the working face model.(3)In the actual con-struction process of the transparent working face,it i

关 键 词：工作面建模 透明工作面 插值算法 误差分析 光滑离散插值 

分 类 号：TD713[矿业工程—矿井通风与安全]