Monitoring hydraulic processes with automated time-lapse electrical resistivity tomography (ALERT) |
| |
| Authors: | Oliver Kuras Jonathan D Pritchard Philip I Meldrum Jonathan E Chambers Paul B Wilkinson Richard D Ogilvy Gary P Wealthall |
| |
| Institution: | British Geological Survey, Kingsley Dunham Centre, Keyworth, Nottingham, NG12 5GG, United Kingdom |
| |
| Abstract: | Hydraulic processes in porous media can be monitored in a minimally invasive fashion by time-lapse electrical resistivity tomography (ERT). The permanent installation of specifically designed ERT instrumentation, telemetry and information technology (IT) infrastructure enables automation of data collection, transfer, processing, management and interpretation. Such an approach gives rise to a dramatic increase in temporal resolution, thus providing new insight into rapidly occurring subsurface processes. In this paper, we discuss a practical implementation of automated time-lapse ERT. We present the results of a recent study in which we used controlled hydraulic experiments in two test cells at reduced field scale to explore the limiting conditions for process monitoring with cross-borehole ERT measurements. The first experiment used three adjacent boreholes to monitor rapidly rising and falling water levels. For the second experiment, we injected a saline tracer into a homogeneous flow field in freshwater-saturated sand; the dynamics of the plume were then monitored with 2D measurements across a 9-borehole fence and 3D measurements across a 3 × 3 grid of boreholes. We investigated different strategies for practical data acquisition and show that simple re-ordering of ERT measurement schemes can help harmonise data collection with the nature of the monitored process. The methodology of automated time-lapse ERT was found to perform well in different monitoring scenarios (2D/3D plus time) at time scales associated with realistic subsurface processes. The limiting factor is the finite amount of time needed for the acquisition of sufficiently comprehensive datasets. We found that, given the complexity of our monitoring scenarios, typical frame rates of at least 1.5–3 images per hour were possible without compromising image quality. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
