Connecting the Dots between Geology and Seismic to Mitigate Drilling Risks: Mapping & Characterization of the High Pressure High Temperature Gotnia Formation in Kuwait

As Kuwait focuses on developing the deep Jurassic reservoirs, the Gotnia Formation presents significant drilling challenges. It is the regional seal, consisting of alternating Salt and Anhydrite cycles, with over-pressured carbonate streaks, which are also targets for future exploration. The objective of this study was to unravel the Gotnia architecture, through detailed mapping of the intermediate cycles, mitigating drilling risks and characterizing the carbonate reservoirs. A combination of noise attenuation, bandwidth extension and seismic adaptive wavelet processing (SAWP)) was applied on the seismic data, to improve the signal-to-noise ratio of the seismic data between 50Hz to 70Hz and therefore reveal the Anhydrite cycles, which house the carbonate streaks. The Salt-Anhydrite cycles were correlated, using Triple Combo and Elastic logs, in seventy-six wells, and spatially interpreted on the band-limited P-impedance volume, generated through pre-stack inversion. Pinched out cycles were identified by integrating mud logs with seismic data and depositional trends. Pre-stack stochastic inversion was performed to map the thin carbonate streaks and characterize the carbonate reservoirs. The improved seismic resolution resulted in superior results compared to the legacy cube and aided in enhancing the reflector continuity of Salt-Anhydrite cycles. In corroboration with the well data, three cycles of alternating salt and anhydrite, with varying thickness, were mapped. These cycles showed a distinctive impedance contrast and were noticeably more visible on the P-impedance volume, compared to the seismic amplitude volume. The second Anhydrite cycle was missing in some wells and the lateral extension of the pinch-outs was interpreted and validated based on the P-impedance volume. As the carbonate streaks were beyond the seismic resolution, they were not visible on the Deterministic P-impedance. The amount of thin carbonate streaks within the Anhydrite cycles could be qualitatively assessed based on the impedance values of the entire zone. Areas, within the zone, with a higher number of and more porous carbonate streaks displayed lowering of the overall impedance values in the Anhydrite zones, and could pose drilling risks. This information was used to guide the pre-stack stochastic inversion to populate the thin carbonate streaks and generate a high-resolution facies volume, through Bayesian Classification. Through this study, the expected cycles and over-pressured carbonate layers in the Gotnia formation were predicted, which can be used to plan and manage the drilling risks and reduce operational costs. This study presents an integrated and iterative approach to interpretation, where the well log analysis, seismic inversion and horizon interpretation were done in parallel, to develop a better understanding of the sub-surface. This workflow will be especially useful for interpretation of over-pressured overburden zones or cap rocks, where the available log data can be limited.

Integrated Stochastic Inversion Approach Assessing Thin Reservoirs Full Potential in the Prolific Oil-Prone Periphery of Hassi Messaoud Field, Algeria

Nowadays it become harder, risker and more expensive understanding the reservoir potentiality and design the optimum development plan for challenging thin reservoirs. In a geological complex area, integrated seismic reservoir characterization approach was crucial to unlock the potentiality for the study area, which is located at Oued Mya basin, SE Saharan platform Algeria. Seismic data analysis is one of the key procedures for characterizing reservoirs and monitoring subsurface properties. Integration of seismic stochastic inversion and geological model will help to demonstrate the link between seismic and reservoir properties more quantitative. The seismic data of region were challenging, with around 30 wells drilled over 400 Km2. To overcome the challenges, the available geological and geophysical data were integrated to construct the reservoir characterization study and reduce drilling uncertainty. Ensure the reservoir characterization process was constrained by a robust workflow and consistent with the available geophysical, geological, and petrophysical data. Petrophysical interpretation, seismic interpretation, rock physics analysis and Stochastic Inversion processes were carried out. These processes were integrated to characterize the lateral and vertical extent of the lithofacies in five stacked reservoirs across the area of interest to identify the potential delineation of thin reservoirs of nine-meter thickness.

Uncertainty quantification in stochastic inversion with dimensionality reduction using variational autoencoder

Estimating rock and fluid properties in the subsurface from geophysical measurements is a computationally and memory intensive inverse problem. For nonlinear problems with non-Gaussian variables, analytical solutions are generally not available, and the solutions of those inverse problems must be approximated using sampling and optimization methods. To reduce the computational cost, model and data can be re-parameterized into low-dimensional spaces where the solution of the inverse problem can be computed more efficiently. Among the potential dimensionality reduction methods, deep learning algorithms based on deep generative models provide an efficient approach to reduce the dimension of the model and data vectors. However, such dimension reduction might lead to information loss in the reconstructed model and data, reduction of the accuracy and resolution of the inverted models, and under or overestimation of the uncertainty of the predicted models. To comprehensively investigate the impact of model and data dimension reduction with deep generative models on uncertainty quantification, we compare the prediction uncertainty in nonlinear inverse problem solutions obtained from Markov chain Monte Carlo and ensemble-based data assimilation methods implemented in lower dimensional data and model spaces using a deep variational autoencoder. The proposed workflow is applied to two geophysical inverse problems for the prediction of reservoir properties: pre-stack seismic inversion and seismic history matching. The inversion results consist of the most likely model and a set of realizations of the variables of interest. The application of dimensionality reduction methods makes the stochastic inversion more efficient.

Implementation of Vendor-Independent Stochastic Inversion for Improving Quality and Efficiency of Well Placement on the Field of Novatek Company

The article describes the application of a new stochastic inversion of the deep-azimuthal resistivity data, independent from the tool vendor. The new model was performed on the data from several wells of the PAO «Novatek», that were drilled using deep-azimuthal resistivity tools of two service companies represented in the global oilfield services market. This technology allows to respond in a timely manner when the well approaches the boundaries with contrasting resistivity properties and to avoid exit to unproductive zones. Nowadays, the azimuthal resistivity data is the method with the highest penetration depth for the geosteering in real time. Stochastic inversion is a special mathematical algorithm based on the statistical Monte Carlo method to process the readings of resistivity while drilling in real time and provide a geoelectrical model for making informed decisions when placing horizontal and deviated wells. Until recently, there was no unified approach to calculate stochastic inversion, which allows to perform calculations for various tools. Deep-azimuthal resistivity logging tool vendors have developed their own approaches. This article presents a method for calculating stochastic inversion. This approach was never applied for this kind of azimuthal resistivity data. Additionally, it does not depend on the tool vendor, therefore, allows to compare the data from various tools using a single approach.

Conditional stochastic inversion of common-offset GPR reflection data

We present a stochastic inversion procedure for common-offset ground-penetrating radar (GPR) reflection measurements. Stochastic realizations of subsurface properties that offer an acceptable fit to GPR data are generated via simulated annealing optimization. The realizations are conditioned to borehole porosity measurements available along the GPR profile, or equivalent measurements of another petrophysical property that can be related to the dielectric permittivity, as well as to geostatistical parameters derived from the borehole logs and the processed GPR image. Validation of our inversion procedure is performed on a pertinent synthetic data set and indicates that the proposed method is capable of reliably recovering strongly heterogeneous porosity structures associated with surficial alluvial aquifers. This finding is largely corroborated through application of the methodology to field measurements from the Boise Hydrogeophysical Research Site near Boise, Idaho, USA.