Application of particle imaging velocimetry (PIV) in centrifuge testing of uniform clay

The Particle Imaging Velocimetry (PIV) technique is being used as an alternative tracking of centroids of markers placed in a regular grid to measure soil displacements. PIV has been used for many years in fluid mechanics; but for physical modelling in geotechnical engineering, this technique is still relatively new. The use of PIV is feasible provided the surface contains sufficient texture. A Cambridge group has shown that natural sand contains enough texture for applying PIV. But when homogeneous clay is used, there is not enough texture on the surface for PIV to be used. Thus to use PIV, a texture has to be marked on the clay surface. This paper addresses specifically the issue of PIV on clay model which does not have natural texture. Two different methods of marking this texture were examined. Then, through a series of calibration tests, errors under 1g and 100g conditions using the PIV method are assessed. Some factors influencing the accuracy of the PIV method were also examined. The results confirmed that without adequate texture, PIV cannot be applied to clay. The calibration exercise shows that the precision of PIV under 1g condition is about 0.008 mm while the precision at 100g condition is about 0.01 mm. It is also shown in this paper that increasing the density of particles per window beyond a threshold value will not contribute significantly to improved accuracy. However, using smaller size markers is desirable, for improved accuracy as well as being able to get more displacement vectors from the same imaging area.