CHARACTERIZATION OF INVERSION ARTIFACTS IN UDAR DATA FOR GEOSTEERING APPLICATIONS

Optimal well placement is important for exploration, appraisal, pilot and production wells. Ultra-deep azimuthal resistivity (UDAR) has been a key tool for mapping reservoir geometry and optimizing well placement for more than a decade. UDAR tools have larger transmitter-receiver spacings and lower transmitting frequencies than previous electromagnetic logging tools. This gives the capability to map reservoir geometries at large distances around the wellbore, but interpreting the data can be challenging due to inversion artifacts. By artifacts we mean apparent resistivity contrasts in the inversion result which do not correspond to actual resistivity contrasts. In this paper we show that significant artifacts can appear even in relatively simple settings; even with simple 2D geometries the assumptions in the 1D inversion are not met. To be able to distinguish commonly seen artifacts from actual resistivity contrasts within the UDAR sensitivity range, a series of simple models have been constructed. The models are variations of simple geometric shapes. This facilitates a quantitative analysis of the shape, size, and magnitude of the observed artifacts. Using one service provider’s cloud-based simulation and inversion software, we have independently carried out simulation and inversion of UDAR data. Results from the simple models have been compared with field data observations. We observe in the synthetic data that the truncation of resistivity contrasts causes strong artifacts in the inversion. They can be observed as resistivity contrasts dipping up or down from the edge of the truncation, or as a pull up or push down of the oil/gas water contact or top reservoir. All these types of artifacts can also be observed in field data. With the use of simple models, we have been able to recreate several artifacts typically observed on actual data from the Norwegian continental shelf. Many artifacts are caused by the use of 1D inversion in inherently 2D or 3D geological settings, but artifacts can also be caused by mechanisms such as inversion converging to a local minimum, unaccounted anisotropy or data windowing. By understanding the root cause of the artifacts and recognizing them in actual data, we increase our confidence in the UDAR interpretation. This knowledge increases the value of information from UDAR measurements for both real-time geosteering and final well completion. It also increases the value of UDAR data for post-well analysis as input to geomodels and further well planning.