Formation evaluation in high-angle / horizontal (HAHz) wells is particularly challenging in complex reservoirs such as channelized sands. Data acquired in such con-text cannot be fully representative of the reservoir physical properties if the reservoir architecture is not properly considered in forward modeling and inversion.
This paper presents a case study from a channelized sand reservoir where a combined interpretation of LWD images and resistivity inversion delivered corrected formation resistivity values for a water saturation estimate. Formation evaluation data from different logging-while-drilling (LWD) tools were used to land a horizontal well in the channelized sand reservoir (Barents Sea). The inversion of resistivity data was performed to obtain a real-time reservoir model and to support steering decisions. The real-time inversion results were then improved in a post-well analysis using memory data with higher resolution. Inversion intervals with poor data match were re-processed, enabling assessment according to quality control knowledge. An additional benefit from LWD resistivity memory data is that more components are available, facilitating a more substantiated evaluation of formation properties. The inversion process includes constraints from image interpretation: formation dips, interpreted from borehole images at best resolution, were used to limit the inversion parameters and improve the underlying earth model for saturation estimates. In a further step, the entire reservoir architecture was used to constrain the inversion process in structurally complex intervals. Constrained inversion delivered a geologically sound model of the formation resistivity, wherein bedding structures are better resolved when compared to unconstrained inversion.
Post-well analysis of this dataset also includes different inversion options like interval lengths and automation settings. The results of the different approaches will be discussed in detail, including the final resistivity map for parts of the channelized sand reservoir. Analysis of differences between synthetic and measured data and the integration with other well data such as borehole images was utilized to refine the underlying earth mod-el in challenging intervals. The final target was to obtain a representative model of the reservoir physical properties, allowing a robust and consistent reservoir characterization.