Laboratory Investigation of Pipeline/Soil Interactions Using X-Ray Computer Assisted Tomography

X-ray Computer Assisted Tomography (CAT or CT) Scanning has been used successfully for the determination of physical properties of porous rocks and for flow visualization in a variety of porous media including soils. CAT scanning has also been demonstrated as an effective tool for the visualization of different stress conditions in porous rocks with greater accuracy when dealing with unconsolidated media. This success was the motivation in trying to expand the diagnostic capabilities of CAT scanning in the phenomena associated with pipeline / soil interactions. For this purpose, a physical model was built. The model consisted of a x-ray transparent holder which had two pistons, one at each end. The pistons were mounted on a cylindrical rod of variable diameter. The holder was then filled with sand. The entire apparatus was placed in the CAT scanner gantry. A series of experiments were performed whereby the rod was forced through the sand pack. The holder was scanned from end to end and the images of various cross-sections were acquired and analyzed for bulk density and porosity. The experiments were coupled with calibration experiments where a uniformly packed sand was loaded under hydrostatic load in given increments. As the sand pack compacted, its bulk density increased. The normalized change in density (strain) was monitored as a function of the pressure load (stress). The results of the calibration tests were used to identify the levels of stress on the sand surrounding the moving rod. It was discovered that areas of compaction ahead of the moving rod and dilation behind the moving rod could be successfully identified and mapped. The stress / strain calibration data allowed the translation of the bulk density images into stress maps around the pipeline. Although the system and materials used in this work were utilized only for demonstration, it was evident that this type of experimental work could be successfully used to calibrate complicated field scale computer models that are very difficult to tune because of the lack of experimental data.