Advances in electromagnetic imaging in the presence of well casings: algorithms and field experiments

Controlled-Source Electromagnetic (CSEM) methods have the potential to be powerful geophysical tools for imaging and monitoring the distribution of electrically resistive fluids, such as freshwater aquifers, CO₂ injected into the subsurface or hydrocarbons during oil and gas production.  However, the presence of metallic infrastructure (steel well casings, pipelines etc.) presents an enormous challenge, because the highly conductive metal masks the electromagnetic response of subsurface geology and distorts any associated time-lapse changes. Therefore, numerical techniques to predict and mitigate the contamination caused by pipelines and casings on CSEM surveys are critical for real world imaging and 4D applications near any such metal objects.

In a collaborative project between the Colorado School of Mines and Shell, we have developed CSEM modeling and inversion tools that can handle realistic scenarios with multiple vertical as well as deviated casings and complex pipeline networks, as will be encountered in mature oil field environments. First, we implemented a forward modeling code based on the Method of Moments technique, which effectively turns the casings into extra sources, such that we do not need to discretize them into excessive numbers of very small model cells. We used this modeling tool to demonstrate quantitatively how steel casings impact synthetic and real time-lapse EM data. The forward modeling code was then combined with a newly developed Gauss-Newton inversion engine, which by itself has been demonstrated to provide images of superior resolution, depth penetration and data fit with less dependency on initial conditions compared to previous quasi-Newton inversion engines.

In this contribution, we first demonstrate on synthetic data that the combination of these two algorithms provides high-quality electrical resistivity images in the immediate vicinity of well casings. Then, we show encouraging results of applying the new tools to field trial data acquired over known casings under semi-controlled conditions. The images obtained are nearly free of casing imprint and subsurface geology could be recovered. These results suggest that this technology may enable us to explain severely distorted field data that were previously uninterpretable.