Item

Abstract

Geoelectrical methods are particularly suited for CO2 injection monitoring due to their high sensitivity to fluid displacement processes in porous rock formations. The use of borehole electrodes is favorable for deep storage horizons. Yet data acquisition based on permanently installed borehole electrodes can be challenged by the finite extent of the electrodes, unintended borehole deviation and complex borehole completion. Such conditions can lead to systematic errors in the electrical data sets, distortions of tomograms, and ultimately misinterpretations. We systematically analyze the effects of different borehole related error sources on tomographic inversion results and present respective methods for mitigation. Specifically, we incorporate the finite extent of the ring electrodes and the borehole completion into the electrical finite-element models and discuss the opportunity to infer borehole deviations solely based on geoelectrical data by means of a coupled inversion. While the finite extent of ring electrodes can be neglected if the electrode spacing is sufficiently large (> 5 m), different borehole completion materials used to fill the well annulus can cause potentially strong resistivity contrasts between the borehole completion and the rock formation, i.e., close to the electrodes. Resulting inversion artifacts are generally less severe when the borehole completion is more resistive compared to the surrounding rock. It is also shown that 2.5D inversion approaches are not adequate for imaging injection experiments in the presence of borehole completion. Unintended borehole deviation can result in geometric errors. Especially, vertical electrode shifts cause strong and localized inversion artifacts. Coupled inverse schemes potentially provide the opportunity to infer electrode shifts solely based on geoelectrical data provided the availability of high quality measurements (< 5% data error). After discussing the effects of the different borehole related error sources, the mitigation methods are validated using synthetic data sets. Subsequently, relevant methods are applied to a field data set from the Ketzin CO2 storage site, Germany, where crosshole electrical resistivity imaging is used for CO2 migration monitoring. The mitigation methods presented can improve estimates of the subsurface resistivity distribution, which, in our particular example, is an essential basis for the quantification of CO2 saturation from time-lapse geoelectrical measurements.