Automatic inversion of deep-directional-resistivity measurements for well placement and reservoir description

Abstract

Deep-directional-resistivity (DDR) logging-while-drilling (LWD) technology has gained acceptance as an efficient method for horizontal well landing, geosteering, and reservoir and fluid-contact mapping. The largest part of the value lies in the automatic inversion of the measurements, improving the detection of geologic and fluid boundaries within a radius of investigation of more than 30 m around the wellbore in real time. The tool and its sensors were designed to acquire a data set with sensitivity to a number of formation properties ensuring the concurrent deployment of a robust inversion. Based on this, an inversion method implemented on computer clusters statistically samples (or explores) the multidimensional inversion space, returning the distribution of formation models that fit the available data. The most noticeable aspect to operators is that this is done without introducing bias. There is no need to commit to a number of layers or to the position, thickness, resistivity, or dip of the layers to generate the continuous images on which all DDR interpretation is based, which translates into several theoretical and practical advantages. In practice, this gives users a higher level of quality control and more confidence in interpreting the subsurface within a larger diameter around the wellbore, which is useful for integrating the information into a geomodel. This derisks current and future operation, improves real-time and planning decisions, and ultimately drives better well placement, completion, production, and return on investment.