残差灰色预测模型在地面沉降监测中的应用

Residual Error Modified Model Based on GM(1,1)and Its Application in Monitoring Land Subsidence

抽取地下液体资源引起的地面沉降是一种严重的地质灾害。对天津市某一地区的地面沉降非线性特征进行了探讨,得到了经数据处理后的地面沉降特征曲线。在地面沉降这一复杂系统中,既含有已知的又含有未知的或非确定的信息,可以作为一个灰色系统来研究。针对地面沉降的下沉曲线非线性特征,提出用一种基于残差灰色预测模型对地面沉降量时间序列进行研究。结果表明,通过残差修正,预测精度得到提高,利用此模型预测地面沉降可减少地面沉降监测经费和实时提供预警信息。

Land subsidence caused by pumping groundwater is a severe geological hazard in Tianjin.The nonlinear characteristics of land subsidence in an area of Tianjin were discussed and the characteristic curve of land subsidence was obtained based on data operation.In the complex system of land subsidence,including definite and indefinite conditions,so it could be researched as a grey system.To the nonlinear characteristics of land subsidence curve,a residual error modification method was presented based on GM(1,1) to study the time subsidence sequence.The results showed that the residual error modification based on GM(1,1) could raise forecasting precision,it was a practical method to forecast land subsidence.This method could reduce the monitoring fee of land subsidence and provide the early-warning timely.