关键词: 迭代协同克里金, 反演方法, 非均质地基, 固结沉降预测, 敏感性

Abstract: Currently, the prediction of ground consolidation settlement through inversion is primarily based on the assumption of soil homogeneity or stratification. However, it is a well-established fact that the hydrological and mechanical parameters of natural ground exhibit spatial variability. To address this issue, an iterative co-Kriging inversion method that utilizes settlement and pore pressure monitoring data for high-resolution inversion is introduced, aiming to depict the spatially non-uniform distribution of ground soil parameters. The resolution of the data employed for inversion characterization is coupled with sensitivity analysis to elucidate the underlying mechanism. The findings indicate that the parameter field obtained through this method represents the optimal unbiased estimate. Employing both settlement and pore pressure observation data for inversion to characterize the heterogeneous ground proves more effective than utilizing solely settlement or pore pressure data for such analysis. Furthermore, a higher resolution leads to improved predictive accuracy of ground consolidation and settlement. The inversion resolution of various observation information types for different parameters is positively associated with the sensitivity magnitude of observation information to parameters.

Key words: iterative co-Kriging, inversion method, heterogeneous ground, prediction of consolidation settlement, sensitivity