Effect of depositional method on the undrained behavior and microstructure of sand with silt

It has been known for many years that depositional method plays an important part in the results of laboratory testing of clean sands. In this paper, undrained triaxial compression tests were performed on specimens composed of Nevada sand with 20% non-plastic silt content using a variety of depositional methods that include slurry deposition, water sedimentation, air pluviation, mixed dry deposition, and dry funnel deposition. It was found that there was a marked difference in the undrained behavior even though the density and stress conditions were identical. The conclusion was that the soil fabric was responsible for this result. Using the scanning electron microscope, a method was developed to examine and quantify the microstructure of silty sand. These results appear to indicate that there are internal particle structures in sand with silt and their relative quantities correlate with the observed macroscopic undrained behavior.     

Reconstitution de données climatiques pour l’Algérie du Nord : application des réseaux neuronaux

In the present context of climate change and preservation of biodiversity, the appreciation of the vulnerability of the natural ecosystems and their capacity of adaptation appears among the main preoccupations to the world level (GIEC, 2007). This assessment of the ecosystems requires the availability of climatic data, what is often made difficult by the weak density or even the absence of meteorological stations notably, to the level of the mountains zones. In order to study the climate–vegetation relationship in North Algeria, we use an automatic interpolation method, the neural network method, for the reconstitution of climatic data of the sampled sites, (1035 phytoecological samples), from the existing meteorological network (269 stations). This method is characterized by a great suppleness of non-linearity and by its capacity for reconstituting information from partial and not well-defined indications such as the case of data provided from meteorological networks. In order to reconstitution of climatic data, we use the explicate variables, longitude, latitude and altitude, the variables to explain being the rainfall and temperatures. To define the best approach, the network calibration has been activated on climatic parameters taken globally or solely, for the whole of study zone, and by geographical sector. The results of the interpolation are expressed through a climatic parameter cartography, released automatically by the MapInfo software. The reliability results obtained by this method can be appreciated by elaboration of errors maps comparing to reference data.