Currently, a significant amount of logging-while-drilling (LWD) borehole image processing is associated with the correction of acquisition-related artifacts such as depth tracking, truncations and tool eccentering. However, the process of true dip computation is still typically based on the routines developed in the 1960s. This approach was formulated for wireline images acquired in vertical wells where formation boundaries and fractures were assumed to be a planar surface, and could be presented as a sinusoid on the borehole image log. This assumption satisfies the majority of wireline cases due to the very limited depth of investigation of wireline images, and the relatively short exposure length of the geological feature along the borehole.
A different scenario often exists for an LWD borehole image in a high-angle or horizontal well. Due to the low well-to-formation incidence angle (relative dip), the boundary can be potentially observed along the borehole wall for tens of meters, and occasionally greater. Within these distances planar geological boundaries are rarely found. The corresponding borehole image response does not have an ideal sinusoid shape. Further, it is unlikely that a well trajectory has a constant inclination and azimuth value over the interval penetrating the geological boundary. The conventional 2-dimensional approach to the interpretation results in an inaccurate true dip in these cases. The variable depth of investigation (DOI) for density and gamma ray LWD imaging tools causes additional uncertainty. Advanced software packages enable consideration of a variable DOI, but the concept of a single DOI value (although variable in depth) is acceptable only for the planar intersections. A final challenge is the manual (interactive) picking of every sinusoid that can be a considerable issue in performing image interpretation.
A recently developed, robust method for true dip calculation considers these parameters but requires minimal user intervention. This workflow relies on a 3-dimensional approach to determine bedding geometry by describing geological features with contours on a 2-dimensional image and incorporating 3-dimensional Cartesian coordinates. The method forms a set of contours where relevant formation boundaries are selected using a set of fuzzy logic rules. Special attention to noise removal contributes to the successful application of the method. An additional capability handles borehole images that contain borehole breakouts and washouts.
A study was conducted over a number of wells with a range of trajectories including high-angle, horizontal and build-up sections. True dips calculated with the developed workflow appeared consistent within the entire borehole. Furthermore, the proposed approach removed human bias, promoting accurate reservoir characterization.