Search for a command to run...
This paper describes a series of experiments that tested the ability of the electrical resistivity tomography (ERT) method to locate correctly wet and dry fractures in a meso-scale model. The goal was to develop a method of monitoring the flow of water through a fractured rock matrix. The model was a four by six array of limestone blocks equipped with 28 stainless steel electrodes. Dry fractures were created by placing pieces of vinyl between one or more blocks. Wet fractures were created by injecting tap water into a joint between blocks. In electrical terms, the dry fractures are resistive and the wet fractures are conductive. The quantities measured by the ERT system are current and voltage around the outside edge of the model. The raw ERT data were translated to resistivity values inside the model using a three-dimensional Occam's inversion routine. This routine was one of the key components of ERT being tested. The model presented several challenges. First, the resistivity of both the blocks and the joints was highly variable. Second, the resistive targets introduced extreme changes the software could not precisely quantify. Third, the abrupt changes inherent in a fracture system were contrary to the smoothly varying changes expected by the Occam's inversion routine. Fourth, the response of the conductive fractures was small compared to the background variability. In general, ERT was able to locate correctly resistive fractures. Problems occurred, however, when the resistive fracture was near the edges of the model or when multiple fractures were close together. In particular, ERT tended to position the fracture closer to the model center than its true location. Conductive fractures yielded much smaller responses than the resistive case. A difference-inversion method was able to correctly locate these targets.