Existing technology used to measure stress in granular materials is susceptible to water ingress and resulting damage to the electrical components, which limits this technology's use in long-term monitoring of soil structures. The connections of these instruments to readout or recording devices are also fragile and easily damaged. In order to explore the efficacy of fibre optic technology not just as a replacement component, but rather as a material of robust elastic properties, a series of experiments was devised to test the possibility of applying a transverse lateral stress to the fibre and measuring its transformed longitudinal strain. The fibre was inscribed with a 5 mm long Bragg grating and encapsulated in a softer material of high Poisson's ratio in an attempt to enhance the longitudinal strain developed in the fibre. This encapsulated device was then subjected to one-dimensional stress in a standard sand, and the strain in the fibre measured. The experiment showed that satisfactory correlation exists between the measured strain output when converted to an applied stress and the mathematically (and numerically) derived stresses. The fibre/silicone bonding was also modelled and results showed that slippage on that interface could be considered insignificant. In both theoretical and practical applications, this experiment can be regarded as successfully validating the principle of deriving stress from a longitudinal elastic strain measured normal to the applied stress. Hence, development can move towards both miniaturization (for research) and more robust construction (to withstand field conditions). Further research will encompass investigating the response of the cell to water and saturated soil conditions, particularly the device sensitivity to transient stress conditions. In addition, development of the sensor to read the complete three-dimensional state of stress in a soil remains the ultimate goal.

No abstract

This paper reports on the results of an investigation into the feasibility of a fibre Bragg grating stress cell for use in a geotechnical environment. To increase the transverse sensitivity of the sensor, we encapsulated it in silicon rubber. Tests were performed in two types of singular grained compacted silica sands. The data obtained with the calibrated stress cell agreed well with the Boussinesq stress, and it is shows a fair degree of repeatability.

A case history is presented in which movements around an underground powerhouse were predicted using two dimensional boundary element stress analysis. The powerhouse, for the Kiambere Hydro-Electric Project in Kenya, is situated in fresh, very widely jointed migmatite at a depth of 128 to 162 m below ground level.Results of early analyses which employed the vertical and horizontal in situ stresses are presented. These are compared with the results of analyses employing an inclined stress field more closely approximating to the true three dimensional conditions. Shear stresses, which had to be ignored in the earlier analysis, were thus reduced. Finally, the effects of minor excavations around the cavern are assessed and included into the results of the cavern analyses, producing a pseudo three dimensional analysis. The absolute values of the deflections were extremely limited but are of relatively little significance in themselves. More importantly, the case study illustrates the high degree of precision which extensometer monitoring can attain, and the level of correlation achieved between predicted and observed deformations. The in situ stress testing carried out for the study is put into its tectonic context adjacent to the East African Rift Valley.