改进Kriging插值的地面沉降监控模型的研究

Application of Improved Kriging Time Interpolation in Land Subsidence Monitoring

随着城市化进程的加快,地下水超采、高层建筑的兴建、建筑容积率和建筑密度的增加以及特殊的地质环境所引起的城市地面沉降日趋严峻。因此,建立一种通用性更强、效率更高的沉降监控模型是十分必要的。Kriging插值以其独特的建模机制,在区域空间插值中已得到广泛应用,而在时域插值中却鲜有研究,借鉴Kriging空间插值模型,研究其在时间域插值的应用具有重要意义。为保证时域模型数据的时间相关性,减少扰动信息累积,需删除早期数据;而为了保证变异函数能真实反映数据的时间分布特征,模型又需要足够的样本数据。为此,探讨了一种基于灰色理论的等维动态Kriging时域插值模型,采用等维动态预测,保证数据时间相关性,减少信息扰动;采用对建模数据量要求较少的灰理论模型来拟合变异函数曲线,解决样本不足的问题,并结合工程实例,验证了该算法的实用性、可靠性和有效性。

With the acceleration of urbanization,the urban land subsidence are becoming more serious each day due to the over exploitation of groundwater,the construction of the high-rise buildings,the increase of building volume ratio and density,as well as the specific geological conditions. So it is necessary to establish a subsidence monitoring model which is more universal and efficient. Kriging has been widely used in regional spatial interpolation with its unique modeling mechanism,but so far the time domain interpolation hasn＇t been researched. Therefore it is of great significance to study the application of the time domain interpolation combining Kriging spatial interpolation. On one hand,in order to ensure the time correlation of data in time-domain model and reduce the disturbance information accumulation,the early data should be deleted; on the other hand,in order to make sure that the variation function will reflect the characteristics of time distribution,sufficient samples are required. To this end,a dynamic Kriging time-domain interpolation model based on grey theory was established,which designed with dynamic forecast of equal dimension to ensure the time correlation of data and reduce the disturbance information accumulation. Furthermore,the curve of variation function was fitted by using the gray theory model which required less modeling data quantity to solve the problem of insufficient samples. This algorithm is proved practicable,reliable and efficient in the application to the actual engineering case of this study.