基于灰色模型的矿区地表沉降预测研究

Study on Prediction of Surface Subsidence in Mining Area Based on Grey Model

煤矿过度开采会引起一系列灾害问题,为防止灾害事故造成严重损失,预测矿区地表沉降刻不容缓。本研究利用三次样条插值法处理得到的等时间序列沉降数据,建立了GM(1,1)模型、加权参数估计模型和灰色Verhulst模型,对矿区地表沉降变形进行数据拟合和预测。对比曲线和精度检验结果表明GM(1,1)模型与加权参数估计模型不适用于沉降突然增大的情况,宜用其进行前期沉降预测;灰色Verhulst模型拟合和预测精度最高,与实际沉降变化一致,适用于矿区地表沉降预测。

Overexploitation of coal mine will cause a series of disaster problems.In order to prevent seri⁃ous losses caused by disaster accidents,it is urgent to predict the surface subsidence in the mining area.Based on the equal time series settlement data processed by cubic spline interpolation method,GM(1,1)model,weighted parameter estimation model and grey Verhulst model are established to fit and predict the surface settlement deformation in the mining area.The comparison curve and accuracy test results show that GM(1,1)model and weighted parameter estimation model are not suitable for the sudden in⁃crease of settlement,so they should be used for early settlement prediction;the grey Verhulst model has the highest fitting and prediction accuracy,which is consistent with the actual settlement change,and is suitable for the prediction of surface settlement in the mining area.