New Software for Processing of LWD Extradeep Resistivity and Azimuthal Resistivity Data

Abstract

Summary In petroleum exploration, reservoir navigation is used for reaching a productive reservoir and placing the borehole optimally inside the reservoir to maximize production. For proper well placement, it is necessary to calculate in real-time the parameters of the formation we are drilling in and the parameters of formations we are approaching. On the basis of these results, a decision to change the direction of drilling could be made. Modern logging-while-drilling (LWD) extra-deep and azimuthal resistivity tools acquire multicomponent, multispacing, and multifrequency data that provide sufficient information for resolving the surrounding formation parameters. These tools are generally used for reservoir navigation and real-time formation evaluation. However, real-time interpretation software is very often based on simplified resistivity models that can be inadequate and lead to incorrect geosteering decisions. The core of the newly developed software is an inversion algorithm based on transversely isotropic layered Earth with an arbitrary number of layers. The following model parameters are determined in real time: horizontal and vertical resistivities and thickness of each layer, formation dip, and azimuth. The inversion algorithm is based on the method of the most-probable parameter combination. The algorithm has good performance and excellent convergence because of its enhanced capability of avoiding local minima. This capability enables interpretation of real-time resistivity data, including azimuthal and extra-deep measurements. A graphical user interface (GUI) was developed to provide an interactive environment for each stage of the resistivity data interpretation process: preview of input resistivity logs, initial preprocessing and filtering of raw data, creation of initial guess, running inversion and viewing inversion results, and quality-control indicators. Applications of the developed software will be shown on a series of synthetic examples and field data from the North Sea and Gulf of Mexico (GOM). This newly developed software is currently in use for real-time reservoir navigation and post-well analysis.

Extradeep-Resistivity Application in Brazil Geosteering Operations Enables Successful Well Landing

“…Any parameters of this model and tool position relative to layer boundaries can be recovered. More details on the inversion algorithm are in a previous paper (Sviridov et al 2014). This multilayer-inversion software is designed to process real-time data for an effective solution of three main tasks: landing, geosteering, and detecting remote pay zones; this is achieved through the recovery of layer resistivity, anisotropy ratio, thickness, and position of the tool relative to the boundaries.…”

The Importance of Extra-Deep Azimuthal Resistivity in Reservoir Development: An Update on Recent Field Experiences

“…The computational part is described in the literature (Sviridov et al, 2014) and based on the application of the most-probable parameter combination method. The aim of the method is finding the minimum of the objective function comprising three summands: the data misfit term, the regularization term, and the term of parameter constraints.…”

Modeling extra-deep electromagnetic logs using a deep neural network

“…Taking advantage of such logging instruments requires automated workflows. Traditionally for EM logging, deterministic inversions are used (Sviridov et al, 2014;Pardo et al, 2015). Such processes require the solution of possibly thousands of forward problems under strict time constraints (Pardo et al, 2015;Shahriari et al, 2018).…”

“…To take full advantage of the depth of investigation, we train the model to respond to up to seven layers (three above and three below the logging instrument). For the training (offline) phase, we generate a data set using a commercial simulator software (Sviridov et al, 2014) that offers no access to the source code. Critical to the success of the proposed methodology is the design of a relatively small application-specific and geologically consistent training data set, for which we demonstrate that it yields accurate approximation in the modeled geosteering scenarios.…”