Abstract: Traditional methods for the quantification of subgrade compaction uniformity are all based on the univariate parameters, such as the deviation coefficient. However, these parameters are unable to consider the spatial distribution of the collected data and evaluate the non-uniformity of the compaction state. The semivariogram is modeled to characterize the spatial variation of the compaction values(CV) based on the geostatistics. Then the exponential model is utilized to fit the calculated semivariogram curve. Finally the sill value C is obtained as the uniformity indicator to ensure the subgrade compaction quality. Both the suggested geostatistical method and traditional statistical method are employed to assess the compaction uniformity of the experimental subgrades in Zhijiang construction site of Shanghai-Kunming Passenger Dedicated Line. The results demonstrate that, different from the univariate parameter Cv, the proposed parameter C can not only eliminate the influence of the uncertain factors such as systematic error, but also provide a more objective evaluation on the spatial non-uniform distribution of the subgrade compaction state. The compaction uniformity evaluation specification can be significantly improved by the suggested method.

Key words: railway subgrade, compaction quality, uniformity evaluation, geostatistics