Modern geosteering is heavily dependent on real-time interpretation of deep electromagnetic (EM) measurements. We present a methodology to construct a deep neural network (DNN) model trained to reproduce a full set of extra-deep EM logs consisting of 22 measurements per logging position. The model is trained in a 1D layered environment consisting of up to seven layers with different resistivity values.A commercial simulator provided by a tool vendor is utilized to generate a training dataset.The dataset size is limited because the simulator provided by the vendor is optimized for sequential execution.Therefore,we design a training dataset that embracesthe geological rules and geosteering specifics supported by the forward model. We use this dataset to produce an EM simulator based on a DNN without access to the proprietary information about the EM tool configuration or the original simulator source code.Despite employing a relatively small training set size, the resulting DNN forward model is quite accurate for the considered examples: a multi-layer synthetic case and a section of a published historical operation from the Goliat Field.The observed average evaluation time of 0.15 milliseconds per logging position makes it also suitable for future use as part of evaluation-hungry statistical and/or Monte-Carlo inversion algorithms within geosteering workflows.

Hydrocarbon reservoirs are often located in spatially complex and uncertain geological environments, where the associated costs of drilling wells for exploration and development are notoriously high. These costs may be reduced with an optimized well-placement strategy based on real-time geological information, known as well geosteering. To effectively place the well in an updated geomodel and support well geosteering decisions in real-time, we apply an iterative inversion approach based on the Levenberg-Marquardt form of the Ensemble Randomized Maximum Likelihood method. The method estimates geomodel properties together with their uncertainties by reducing the statistical misfit between the measurements acquired with well-logging tools and the predicted measurements from numerical simulations. Analyses of synthetic cases indicate that the method's reliability and computational speed depend on the distance from the logging tool to formation boundaries, the contrast of model properties, and the thickness of formation layers. The proposed method delivers reliable estimates of model properties with only 40 ensemble members and 2--10 iterations, hence it is approximately 10--125 times faster than Metropolis-Hastings Monte Carlo, which we use as a baseline condition given its proven track record. Likewise, the developed method is amenable to parallelization to further reduce computational times.Implementation of the method with a synthetic example inspired by a historical well geosteering operation yielded accurate formation evaluation and verified its accurate and reliable performance under complex geological conditions.

The costs for drilling offshore wells are high and hydrocarbons are often located in complex reservoir formations. To effectively produce from such reservoirs and reduce costs, optimized well placement in real-time (geosteering) is crucial. Geosteering is usually assisted by an updated formation evaluation obtained by well-log interpretation while drilling. A reliable, computationally efficient, and robust workflow to interpret well logs and capture uncertainties in real-time is necessary for this application. An iterative ensemble-based method, namely the approximate Levenberg Marquardt form of the Ensemble Randomized Maximum Likelihood (LM-EnRML) is integrated in our formation evaluation workflow. We estimate model parameters, resistivity and density in addition to boundary locations, and related uncertainties by reducing the statistical misfit between the measurements from the well logging tools and the theoretical measurements from the forward tool simulators. The results of analyzing several synthetic cases with several types of logs verified that the proposed method can give good estimate of model parameters by employing as few as 40 ensemble members and 2-10 iterations. By comparing the CPU time, we conclude that the proposed method has at least about 10-125 times lower computational time compare to a common statistical method, such as Metropolis-Hastings Monte Carlo. In addition, the ensemble-based method can run in parallel on multiple CPUs. The reliability and speed of well-log interpretation is normally sensitive to several parameters such as the distances between the formation boundaries and the logging tool, model parameter's contrast, formation layer thickness and well inclination. Testing the method on a case inspired from a real field also yielded accurate formation evaluation. Thus, the proposed ensemble-based method has been proven robust and computationally efficient to estimate petrophysical formation properties, layer boundaries and their uncertainties, indicating that it is suitable for geosteering.

The majority of geosteering operations rely on traditional shallow sensing logging tools as sources of information. Many such operations rely on stratigraphic-based steering when the logs from the drilled well are matched to logs from an offset well by modifying the lateral shape of stratigraphy. The match of the logs indicates a plausible interpretation, but due to the scarcity of log data in many situations, this interpretation is not unique. In manual workflows maintaining several likely interpretations is not realistic and in automated workflows, multiple interpretations are seldom used. We describe a deep neural network (DNN) that outputs a selected number of stratigraphic interpretations using a single evaluation of the input log data in two milliseconds. The input data defined prior to training consists of one or several log pairs consisting of one current lateral and one offset-well log. For each of the interpretations, the DNN also estimates the respective probability and can be configured to produce likely ahead-of-data predictions of the geology, which are based on the data mismatches and the likelihood of geological configurations with respect to the training dataset. The described probabilistic interpretation and prediction is enabled by the supervised training of a mixture density DNN (MDN) with a stable multiple-trajectory-prediction loss function. In this paper, we apply the MDN for the sequential interpretation of well logs. We use the interpretations and the probabilities from the previous interpretation step as starting points for the probabilistic interpretations and predictions for the current step. We avoid the curse of dimensionality by discarding the unlikely starting points. The batchable MDN evaluation enables tracking of hundreds of solutions while still maintaining sub-second performance, compared to minute(s) reported in other recent papers. The performance of the method is verified on synthetic test data as well as the realistic well data from the Geosteering World Cup 2020 (based on the Middle Woodford formation, located in the South Central Oklahoma Oil Province in the United States) and stratigraphic configurations provided by geologists. In all cases, the method manages to capture likely interpretations. At the same time, the accuracy of predictions deteriorates for the configurations which were not typical for the training dataset.

No abstract