Interwell Saturation Prediction by Artificial Intelligence Analysis of Well Logs

Well log analysis, through deploying advanced artificial intelligence (AI) algorithms, is key for wellbore geological studies. By analyzing different well characteristics with modern AI tools it becomes possible to estimate interwell saturation with improved accuracy, outlining primary fluid channels and saturation propagations in the reservoirs interwell region. The development of modern deep learning and artificial intelligence methods allows analysts to predict interwell saturation as a function of observed data in the near wellbore logged geological layers. This work addresses the use of deep neural network architectures as well as tensor regression models for predicting interwell saturation from other well characteristics, such as resistivity and porosity, as well as local near-well saturation. Several algorithms are compared in terms of both accuracy and computational efficiency. Sensitivity analysis for model parameters is carried out, which is based on the wells’ geometry, radius, and multiple sampling techniques. Additionally, the impact of local saturation prior knowledge on the model accuracy is analyzed. A reservoir box model encompassing volumetric interwell porosity, resistivity and saturation data was utilized for the validating and testing of the AI algorithms. A prototype is developed with Python 3.6 programming language.

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.

Hybrid Physics-Constrained and Data-Driven Approach for Interwell Saturation Estimation from Well Logs

Combining physics-based models for well log analysis with artificial intelligence (AI) advanced algorithms is crucial for wellbore studies. Data-driven methods do not generalize well and lack theoretical knowledge accumulated in the field. Estimating well saturation significantly improves if predictions from physical models are used to constrain data-driven algorithms in outlined primary fluid channels and other important points of interest. Saturation propagations in the reservoirs interwell region also generalize better under using combination of models. This work addresses combined usage of theoretical and data-driven models by aggregating them into single hybrid model. Multiple physical and data-driven models are under study, their parameters are optimized using observations. Weighted sum is used to predict water saturation at every point with weights being recomputed at each step. Model outputs are compared in terms accuracy and cumulative loss. A synthesized reservoir box model encompassing volumetric interwell porosity, resistivity and saturation data is used for the validation of the algorithms. Aggregated model for estimating interwell saturation shows improved prediction accuracy compared both to physics-based or data-driven approaches separately.

Application of Acoustic Impedance and Reflectivity in Reservoir Characterization

Well log analysis is one of the methods for reservoir characterization, in the oil and gas industry. Logs are used for subsurface formation evaluation. They are useful in hydrocarbon zone identification and volume calculation. Interpretation of well log involves sequential steps, which are lithology, shale volume, porosity and saturation determination. It is unwise to analyze well log without following the logical steps, as this could introduce errors in the result. Petrophysical and Geomechanical properties are two classes of properties for reservoir characterization. The computed volume of shale in the reservoir was 10%, the average water saturation was 30%, and the average porosity was 25pu. The bulk density decreased from 2.15g/cc to 1.95g/cc and there is a considerably lower acoustic impedance in the hydrocarbon bearing sands. In challenging reservoirs, where traditional petrophysical methods do not give definitive results, the use of geomechanical methods will improve interpretation certainty and help to clear doubts in the interpreted results.

Optimal Drawdown for Woodford and Mayes in the Anadarko Basin Using Data Analytics

Summary One of the enduring pieces of the jigsaw puzzle for all unconventional plays is drawdown (DD), a technique for attaining optimal return on investment. Assessment of the DD from producing wells in unconventional resources poses unique challenges to operators; among them the fact that many operators are reluctant to reveal the production, pressure, and completion data required. In addition to multiple factors, various completion and spacing parameters add to the complexity of the problem. This work aims to determine the optimum DD strategy. Several DD trials were implemented within the Anadarko Basin in combination with various completion strategies. Privately obtained production and completion data were analyzed and combined with well log analysis in conjunction with data analytics tools. A case study is presented that explores a new strategy for DD producing wells within the Anadarko Basin to optimize a return on investment. We use scatter-plot smoothing to develop a predictive relationship between DD and two dependent variables—estimated ultimate recovery (EUR) and initial production (IP) for 180 days of oil—and introduce a model that evaluates horizontal well production variables based on DD. Key data were estimated using reservoir and production variables. The data analytics suggested the optimal DD value of 53 psi/D for different reservoirs within the Anadarko Basin. This result may give professionals additional insight into more fully understanding the Anadarko Basin. Through these optimal ranges, we hope to gain a more complete understanding of the best way to DD wells when they are drilled simultaneously. Our discoveries and workflow within the Woodford and Mayes Formations may be applied to various plays and formations across the unconventional play spectrum. Optimal DD techniques in unconventional reservoirs could add billions of dollars in revenue to a company’s portfolio and dramatically increase the rate of return, as well as offer a new understanding of the respective producing reservoirs.

Well Log Analysis of Nahr Umr, Shuaiba and Zubair Formations in EB-4 well of East Baghdad Oil Field, Iraq, using Rock Physics Templates (RPTs)

RPT is a method used for classifying various lithologies and fluids from data of well logging or seismic inversion. Three Formations (Nahr Umr, Shuaiba, and Zubair Formations) were selected in the East Baghdad Oil field within well EB-4 to test the possibility of using this method. First, the interpretations of the well log and Density – Neutron cross plot were used for lithology identification, which showed that Nahr Umr and Zubair formations consist mainly of sandstone and shale, while the Shuaiba Formation consists of carbonate (dolomite and limestone). The study was also able to distinguish between the locations of hydrocarbon reservoirs using RPT. Finally, a polynomial equation was generated from the cross plot domain (AI versus Vp/Vs) to estimate one parameter from the other in these formations, and vice versa.

Porosity and Permeability Measurements Integration of The Upper Cretaceous in Balad Field, Central Iraq

The corrected porosity image analysis and log data can be used to build 3D models for porosity and permeability. This can be much realistic porosity obtainable because the core test data is not always available due to high cost which is a challenge for petroleum companies and petrophysists. Thus, this method can be used as an advantage of thin section studies and for opening horizon for more studies in the future to obtain reservoir properties. Seventy-two core samples were selected and the same numbers of thin sections were made from Khasib, Sa`di, and Hartha, formations from Ba-1, Ba-4, and Ba-8 wells, Balad Oilfield in Central Iraq to make a comprehensive view of using porosity image analysis software to determine the porosity. The petrophysical description including porosity image analysis was utilized and both laboratory core test analysis and well log analysis were used to correct and calibrate the results. The main reservoir properties including porosity and permeability were measured based on core samples laboratory analysis. The results of porosity obtained from well log analysis and porosity image analysis method are corrected by using SPSS software; the results revealed good correlation coefficients between 0.684 and 0.872. The porosity range values are 9-16% and 9-27% for Khasib and Sa’di in Ba-1 Well, respectively; 10-21%, 9-25%, and 16-27% for Khasib, Sa’di and Hartha in Ba-4 Well, respectively; and 11-24% and 15-24% for Khasib and Hartha in Ba-8 Well, respectively according to petrographic image analysis. By using the laboratory core analysis, the porosity range values are 12-26% and 17-24% for Khasib and Sa’di in Ba-1 Well, respectively; 6-28% and 14-27% for Sa’di and Hartha in Ba-4 Well, respectively; and 17-19% and 15-24% for Sa’di and Hartha in Ba-8 Well, respectively. Finally, the well log analysis showed that the porosity range values are 11-16% and 7-27% for Khasib and Sa’di in Ba-1 Well, respectively; 4-18%, 21-26%, and 16-19% for Khasib, Sa’di and Hartha in Ba-4 Well, respectively; and 9-24% and 15-23% for Khasib and Hartha in Ba-8 Well, respectively. The permeability range values based on laboratory core analysis are 1.51-8.97 md and 0.29-2.77 md for Khasib and Sa’di in Ba-1 Well, respectively; 0.01-24.5 md and 0.28-6.47 md for Sa’di and Hartha in Ba-4 Well, respectively; and 0.86-2.25 md and 0.23-3.66 for Sa’di and Hartha in Ba-8 Well, respectively.