2种灰色模型在矿区地表沉陷预计中的适用性

目前矿区地表单点沉陷动态预计方法主要基于传统的水准测量数据,监测方法单一,成本高,观测点易破坏,不能保证地表形变信息的实时性,且采用灰色模型进行地表沉陷预计时只针对单一模型的应用,没有结合模型自身特点分析其适用性。以袁店二矿7221工作面为试验区域,采用合成孔径雷达差分干涉测量技术监测矿区地表沉陷量,分别建立了描述沉陷量与时间关系的GM(1,1)与灰色Verhulst模型进行地表沉陷量预计,实现了矿区地表沉陷监测与动态预计一体化。通过比较、分析GM(1,1)与灰色Verhulst模型对地表沉陷量的拟合及预计结果,得出了2种灰色模型在矿区地表沉陷预计中的适用性:在矿区开采沉陷开始至活跃前期,若地表单点沉陷量曲线呈近似单峰型,则宜采用GM(1,1)进行短期预计;当矿区地表沉陷进入衰退阶段,单点沉陷量曲线呈平底饱和状态,则宜采用灰色Verhulst模型进行中长期预计。

Applicability of two grey models in earth-surface subsidence prediction of mining area

Existing dynamic single-point subsidence prediction methods of mining area earth-surface are mainly based on traditional leveling measurement data,which have single monitoring mode,high cost and easily destroyed observation points,and cannot guarantee real-time of earth-surface deformation information.Moreover,application of grey model for earth-surface subsidence prediction is only for a single model and its applicability is not analyzed according to characteristics of the model itself.Taking 7221 working face of Yuandian No.2 Mine as an experimental zone,earth-surface subsidence value is measured by differential interferometry synthetic aperture radar technology,and then GM(1,1)and grey Verhulst model that describe the relationship between subsidence and time are built separately for predicting earth-surface subsidence value,so as to realize the integration of earth-surface subsidence monitoring and dynamic prediction of mining area.Through comparing and analyzing fitting and prediction results of earth-surface subsidence value between GM(1,1)and grey Verhulst model,the applicability of the two grey models in earth-surface subsidence prediction of mining area is obtained:GM(1,1)should be used for short-term prediction from beginning of mining subsidence to early active stage if subsidence value curve of single point is approximately unimodal;when earth-surface subsidence enters into decline stage and subsidence value curve of single point is in flat-bottom saturation,gray Verhulst model should be used for medium and long-term prediction.