Discrete element method has been widely adopted to simulate processes that are challenging to continuum-based approaches. However, its computational efficiency can be greatly compromised when large number of particles are required to model regions of less interest to researchers. Due to this, the application of DEM to boundary value problems has been limited. This paper introduces a three-dimensional discrete element–finite difference coupling method, in which the discrete–continuum interactions are modeled in local coordinate systems where the force and displacement compatibilities between the coupled subdomains are considered. The method is validated using a model dynamic compaction test on sand. The comparison between the numerical and physical test results shows that the coupling method can effectively simulate the dynamic compaction process. The responses of the DEM model show that dynamic stress propagation (compaction mechanism) and tamper penetration (bearing capacity mechanism) play very different roles in soil deformations. Under impact loading, the soil undergoes a transient weakening process induced by dynamic stress propagation, which makes the soil easier to densify under bearing capacity mechanism. The distribution of tamping energy between the two mechanisms can influence the compaction efficiency, and allocating higher compaction energy to bearing capacity mechanism could improve the efficiency of dynamic compaction.
Modern meteorological observations in South China from 1960 to 2009 show a strong correlation between winter temperatures and two snowfall parameters, the southern boundary of the snow and the number of snowy days. Based on this relationship, the variation in annual winter mean temperature in South China from 1736 to 2009 was reconstructed using data acquired from Chinese historical documents dating from the Qing dynasty, such as memos and local gazettes. The reconstructed time series were used to analyse variations in winter temperature in South China. Significant interannual and interdecadal changes were found. The maximum temperature difference between neighbouring years was 3.1 °C for 1958–2009 and 3.0 °C for 1736–1957, whereas the maximum temperature difference between adjacent decades was 0.8 °C for the 1960s–2000s and 0.6 °C for the 1740s–1950s. The 2000s was the warmest decade; the mean temperature was 1.6 °C higher than that of the 1870s, which was the coldest decade between the 1740s and the 2000s. The mean winter temperature was warmer in the 18th and 20th centuries and coldest in the 19th century. 相似文献