|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Johansen, Yngve Bolstad (Aker BP ASA) | Christoffersen, Kjell (Aker BP ASA) | Chatterjee, Amitabha (Aker BP ASA) | Elfenbein, Carsten (Aker BP ASA) | Olsborg, Lodve Hugo (Aker BP ASA) | Kvilaas, Geir Frode (Aker BP ASA) | Baig, Mirza Hassan (Schlumberger) | Datir, Harish (Schlumberger) | Bachman, Nate (Schlumberger) | Kausik, Ravinath (Schlumberger) | Hurlimann, Martin (Schlumberger)
The Ivar Aasen (IA) oilfield is located on the Gudrun Terrace on the eastern flank of the Viking Graben in the Norwegian North Sea. The field was discovered in 2008. The reservoir is located within a sedimentary sequence of Mid-Jurassic to Late-Triassic age, which consists of shallow marine to fluvial, alluvial, floodplain and lacustrine deposits overlying a regionally extensive, fractured calcrete interval. The sequence exhibits a complex mineral composition and is heterogeneous at a scale below that of a logging sensor. Shale layers, re-deposited shale and calcrete fragments are present in various forms throughout the sequence.
The formation evaluation in early field evaluation was based on traditional logs such as gamma ray, neutron, bulk density, and resistivity. The evaluation based on these logs was associated with large uncertainty, particularly in the more complex sediments. This because it is difficult to estimate basic properties such as net fraction, porosity, saturation and permeability when dealing with heterogeneities at a scale lower than the resolution of the sensors. Estimated in place volumes and reserves can be significantly inaccurate in such cases. It is also hard to optimize well placement and drainage strategy when there are uncertainties to where the target is. Due to this, it was decided in 2014 to reduce the uncertainty when an opportunity emerged. The rig to be used for the Ivar Aasen drilling campaign, arrived early. Instead of planning for idle rig time, a more aggressive strategy was chosen. The Ivar Aasen team with partners decided to address the reservoir uncertainty by drilling three data wells with option for sidetracks in the available time window.
In 2015, three “geopilots” were drilled with oil-based mud (OBM), cored, and subsequently logged with advanced wireline tools. The cores were used to improve the understanding of the IA sediments, and its impact on the reservoir performance. Specialized wireline logs were used to characterize mineralogy (capture and inelastic neutron-induced spectroscopy), saturation in laminated shale/sand sections (tri-axial induction), fluid identification and pore volumes (well formation testing/sampling, nuclear magnetic resonance), and near wellbore saturation using non-resistivity based methods (nuclear magnetic resonance, dielectric dispersion). A thin bed formation evaluation method based on the Thomas-Stieber method was applied to interpret the above-mentioned data. Older existing wells were reinterpreted by the same method, also benefiting from the parameters measured in the geopilots. The improved understanding of the reservoir properties based on the interpretations of the advanced wireline data contributed to a significant increase of the in-place volumes in the Triassic Skagerrak 2 reservoir zone. These results led to important modifications of the drainage strategy and the development campaign to be executed in 2016.
Abstract The Ivar Aasen (IA) oilfield is located on the Gudrun Terrace on the eastern flank of the Viking Graben in the Norwegian North Sea. The field was discovered in 2008. The reservoir is located within a sedimentary sequence of Mid-Jurassic to Late-Triassic age, which consists of shallow marine to fluvial, alluvial, floodplain and lacustrine deposits overlying a regionally extensive, fractured calcrete interval. The sequence exhibits a complex mineral composition and is heterogeneous at a scale below that of a logging sensor. Shale layers, re-deposited shale and what was first believed to be redeposited calcrete fragments present in various forms throughout the sequence. Looking more in depth to XRD and XRF data and contrasting Fe concentration in the dolomite, it is also possible to explain some of the carbonate deposits through other processes. Extensive data acquisition in the form of advanced wireline logs and coring with analysis performed in “geopilot” wells before production start, enabled a novel thin bed formation evaluation technique based on the modified Thomas-Stieber method (Johansen et al. 2018). The method increased the in-place oil volumes within the Triassic reservoir zone internally named Skagerrak 2. This led to several improvements and a modified drainage strategy of Ivar Aasen. Several good producers were placed in the complex net of the Skagerrak 2 Formation. Results from these producers have encouraged development of an even more marginal and complex net, deeper into the Triassic sedimentary sequence. Therefore, another “geopilot” was drilled into the deeper Triassic sediments, internally named as the Alluvial Fan. This zone exhibits conglomerate clasts in a matrix varying between clay, silt, feldspars, and very fine to very coarse sand fractions, grading towards gravel. Previously, this zone was considered to be mostly non-net. Applying the same interpretation method as for Skagerrak 2, the Alluvial Fan promised economic hydrocarbon volumes. The latest geopilot proved producible hydrocarbons, and subsequently a producer was also successfully placed in this part of the reservoir. Production data and history matching from the beginning of production have for a long while established the previous increase of IA Triassic oil volumes published in 2018. Advanced studies of mineralogy and spectroscopy (Johansen et al. 2019) have indicated that a significant amount of the previously interpreted dolomite, could be reinterpreted as ferroan dolomite. The latter is a heavier mineral that increases the matrix density, hence also the total porosity. The additional findings described provided another necessary first-order correction to further enhance the evergreen geomodel. This paper describes this methodology which resulted in improved petrophysics and reservoir properties of the Alluvial Fan, yet again demonstrating the value of advanced wireline logs and detailed analysis that in total impacts the IA reserve volumes in a significant manner. Repeated success with the applied spectroscopy data and the thin bed methodology used today (Johansen et al. 2018), has resulted in even the deeper Braid Plain Formation becoming of economic interest. It is expected to lie within the oil zone in an upthrow block in the northern part of the IA field and could be developed into the next target.
Kausik, Ravinath (Schlumberger-Doll Research Center) | Prado, Augustin (Schlumberger-Doll Research Center) | Gkortsas, Vasileios-Marios (Schlumberger-Doll Research Center) | Venkataramanan, Lalitha (Schlumberger-Doll Research Center) | Datir, Harish (Schlumberger) | Johansen, Yngve Bolstad (Aker BP)
ABSTRACT The computation of permeability is vital for reservoir characterization because it is a key parameter in the reservoir models used for estimating and optimizing hydrocarbon production. Permeability is routinely predicted as a correlation from near-wellbore formation properties measured through wireline logs. Several such correlations, namely SDR permeability and Timur-Coates permeability models using nuclear magnetic resonance (NMR) measurements, K-lambda using mineralogy, and other variants, have often been used, with moderate success. In addition to permeability, the determination of the uncertainties, both epistemic (model) and aleatoric (data) are important for interpreting variations in the predictions of the reservoir models. In this paper we demonstrate a novel dual deep neural network framework encompassing a Bayesian neural network (BNN) and an artificial neural network (ANN) for determining accurate permeability values along with associated uncertainties. Deep learning techniques have been shown to be effective for regression problems, but quantifying the uncertainty of their predictions and separating them into the epistemic and aleatoric fractions is still considered challenging. This is especially vital for petrophysical answer products because these algorithms need the ability to flag data from new geological formations that the model was not trained on as "out of distribution" and assign them higher uncertainty. Additionally, the model outputs need sensitivity to heteroscedastic aleatoric noise in the feature space arising due to tool and geological origins. Reducing these uncertainties is key to designing intelligent logging tools and applications, such as automated log interpretation. In this paper we train a BNN with NMR and mineralogy data to determine permeability with associated epistemic uncertainty, obtained by determining the posterior weight distributions of the network by using variational inference. This provides us the ability to differentiate in- and out-of-distribution predictions, thereby identifying the suitability of the trained models for application in new geological formations. The errors in the prediction of the BNN are fed into a second ANN trained to correlate the predicted uncertainty to the error of the first BNN. Both networks are trained simultaneously and therefore optimized together to estimate permeability and associated uncertainty. The model is trained on a "ground-truth" core database representing samples from different geology formations. The application of the machine learning permeability model demonstrates a greater than 50% reduction of the mean square error in comparison to traditional SDR and Timur-Coates permeability models (KSDR and KTIM, respectively) on wells from the Ivar Aasen Field. We also demonstrate how the machine learning workflow enables us to understand the value of information (VOI) of different logging measurements, by replacing the logs with their median values from nearby wells during model inference, and studying the increase of the mean square error in the permeability predictions.
Davis, Graham (Premier Oil) | Newbould, Rob (Premier Oil) | Lopez, Aldo (Premier Oil) | Hadibeik, Hamid (Halliburton) | Guevara, Zunerge (Halliburton) | Engelman, Bob (Halliburton) | Balliet, Ron (Halliburton) | Ramakrishna, Sandeep (Halliburton) | Imrie, Andrew (Halliburton)
The oil and gas potential of the basins surrounding the Falkland Islands has attracted exploration drilling that resulted in discovering the Sea Lion Field in the North Falkland Basin in May 2010. Recent exploration drilling has resulted in new oil discoveries to the south of the Sea Lion Complex that has not only confirmed the area as a significant hydrocarbon province but has also enhanced the likelihood of future commercial development of resources. Primary oil targets are stacked and amalgamated deepwater lacustrine turbidite fans comprising multiple lobes. In exploration and appraisal wells, porosity characterization, permeability assessment, pressure measurements, and hydrocarbon fluid identification are essential input data for robust reservoir characterizations and resource estimations.
A comprehensive suite of advanced logging measurements, in addition to conventional log measurements, have been used to facilitate data analysis and calibration to laboratory core measurements. The pressure gradients and fluid samples obtained from formation testing when combined with the wireline log measurements are fundamental when determining the thickness, quality, and connectivity of hydrocarbon zones, which, in turn, impact the commercial evaluation of the well. In these remote offshore basins where rig costs are high and the ability to focus data acquisition in specific zones of interest and minimize logging time whilst identifying and reacting early in real time to data points that lie off the expected trends can add significant value to the operating company.
Formation evaluation challenges include hydrocarbon identification and resolving fluid contact uncertainties. In addition, establishing whether there are any baffles or barriers in the system or significantly varying reservoir properties as a consequence of facies changes has the potential to complicate the evaluation in respect to permeability characterization and volume estimation.
A method of facies classification using a combination of resistivity-based borehole imaging data and nuclear magnetic resonance (NMR) data is outlined in this paper. This method, when combined with conventional log data, has exhibited encouraging results in terms of identifying lithofacies and determining a rock quality index (RQI). The mud logs and gamma ray logs were interpreted with the borehole image logs in these turbidite reservoirs, which resulted in identifying four distinct depositional lithofacies. These lithofacies were integrated with the free fluid index (FFI) to bulk volume irreducible (BVI) ratio determined from the NMR data. The FFI to BVI ratio was used as an index for RQI classification, which was then subsequently used to optimize formation pressure testing and sampling points.
The contribution and importance of lithofacies identification is typically ignored when optimizing formation pressure depths and interpreting the results. The methodology presented in this paper uses an integrated workflow jointly developed by the operator and service company that allows detailed reservoir evaluation in the zones of interest and real-time adjustments to optimize the data acquisition programme that potentially enables rig-time savings and, consequently, reduces overall formation evaluation costs.