|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
The aim of this work is to study shale gas production subject to water blocking in compressible shale. Water blocking is a capillary pressure end-effect causing the wetting phase (e.g. water) to accumulate near the transition from a porous medium to an open medium; in this context, a transition from shale matrix to a hydraulic fracture. Shale is considered a tight porous medium with ultralow permeability, and hydraulic fracturing is essential to obtain economical production. Water is frequently used as a fracturing fluid, but its accumulation at the matrix end-face reduces the gas mobility and can lead to rapid decline of gas production rate.
The tight nature of the shale as a porous medium also introduces non-standard flow and storage mechanisms. This work develops a mathematical model that accounts for apparent permeability, compressibility of gas and shale, gas adsorption, Forchheimer gas flow, and multiphase flow parameters like relative permeability and capillary pressure, which depend on wettability. The behavior of the model at steady state production is explored to understand the impact of the various mechanisms.
Kalam, Shams (King Fahd University of Petroleum & Minerals) | Khan, Mohammad (Schlumberger) | Mahmoud, Mohamed (King Fahd University of Petroleum & Minerals) | Khan, Rizwan Ahmed (King Fahd University of Petroleum & Minerals) | Abu-Khamsin, Sidqi A. (King Fahd University of Petroleum & Minerals)
Artificial intelligence (AI) has proven to be the smartest predicting tool in the oil and gas industry. In this paper, Artificial Neural Network (ANN) algorithm was applied to build two new empirical correlations to predict relative permeability profiles of oil-water two phase flow in the reservoir for both sandstone and carbonate reservoirs. The proposed model evaluates the relative permeability as a function of porosity, rock absolute permeability, initial water saturation, residual oil saturation, wettability index and water saturation. Accordingly, relative permeability to water and oil are respective outputs. Real data of both sandstone and carbonate reservoirs taken from literature were used in the development of the new empirical correlations. Multiple realizations with various hidden layer neurons were run to find the best scenario; and maximum coefficient of determination (R2) was designated as the finest case. The weights and biases values were found for the models of relative permeability to water and oil after proper training and are presented in this paper. Tan-sigmoid and linear transfer functions were utilized in the hidden and output layers, respectively. Neural Network was trained using Levenberg-Marquardt back-propagation algorithm. The novel ANN model was able to accurately estimate relative permeability to oil and water for an unseen data set of 319 real data points. Root mean squared error for both models are near to zero, while R2 for relative permeability to oil and water is 0.92 and 0.98, respectively. The relative permeability models are presented in the form of an actual mathematical correlation. The use of the developed ANN models significantly saves time and cost for conducting experiments for relative permeability measurements.
The Ichthys gas-condensate field is situated in the Browse Basin, North West Shelf of Australia, and the field production commenced in July 2018. The Brewster Member, one of the two reservoirs in the field, is a liquid-rich sandstone reservoir. One of the major uncertainties is the degree of well productivity impairment, caused by condensate banking since the dew point pressure is close to the initial reservoir pressure. The objectives of this study are to evaluate the condensate banking impact on well production performance, and to establish a modelling methodology to consider the condensate banking effect in a full-field simulation model, based on the field production data.
Permanent downhole gauges are deployed in the field, and thus, downhole pressure can be monitored continuously. We conducted high rate tests for selected wells to monitor well productivity impairment from the condensate banking. This production data was history-matched with a compositional sector model by applying Local Grid Refinement (LGR) and Velocity-Dependent Relative Permeability (VDRP) to account for more accurate physics in the near-well region. With the tuned VDRP model, skin trends were predicted to increase with various gas rates, and a skin correlation was established as a function of this gas rate. This correlation is applied to the full-field simulation where LGR and VDRP cannot be applied due to a simulation time constraint.
The skin correlation was validated through the history matching, using the full-field model and was used to predict the future field production performance. We need continuous monitoring of the condensate banking effect, to further validate the correlation, because the production data used in this study is less than one-year duration. The correlation is then flexible enough to tune the history matching when necessary.
We present the monitoring and modelling of the condensate banking effect with the actual production data. The implementation of the proposed well modelling will help reservoir engineers in considering the condensate banking effect in the field production forecast.
A multi-chamber, finite-difference dynamic model assumed pore-scale reservoir properties and replicated laboratory processes to determine oil-water transition zone characteristics: (1) imbibition residual oil saturation to water flood (distributed as a function of rock properties); (2) imbibition oil-water relative permeability (according to both steady-state and unsteady-state calculations); and, (3) microscopic oil recovery factor as a function of height above free water level. A priori, scaled capillary pressure hysteresis and segregated flow conditions governed pore-scale flow behaviour (initial mobile saturation scaled 0-1) assuming a pipeline network analogue. Although an oil-water system was modelled the process could be extended to other fluid phases.
The dynamic model was divided into five chambers replicating two experiments. Flow between chambers was regulated by transmissibility multipliers (open/shut). Replicating a core displacement experiment, an oil source rock "drained" (flooded) the central core chamber to form the oil-water transition zone; water then imbibed into the core chamber displacing mobile oil into a shallower effluent chamber; capillary pressure hysteresis defined the imbibition residual oil saturation to water flood. In a second experiment, the central core chamber was subdivided into zones, each comprising five model layers, and the imbibition oil-water relative permeability was determined by flooding the quiesced core chamber situated between high/low-pressure chambers (source/sink). Both steady-state and unsteady-state relative permeability were determined across the oil-water transition zone.
The Digital Core Laboratory (DCL) determined the distribution of imbibition residual oil saturation as a function of: wettability assumption; rock properties; and, proximity to the free water level. For a range of wettability assumptions, imbibition oil-water relative permeability was determined across the transition zone, on a zonal basis, from top core chamber (top reservoir) down to the free water level.
Summary of Results: (1) The imbibition residual oil saturation (ISorw) decreased with depth toward the free water level, while the drainage irreducible water saturation (Swir) increased with depth; (2) ISorw was found to be strongly influenced by the wettability assumption, contrasting with the uniformity of Swir; (3) Microscopic oil recovery factor was determined across the transition zone as a function of both proximity to the free water level and the wettability assumption; except for the oil-wet case, microscopic oil recovery factor increased with depth toward the free water level.
The novel approach was the multi-chambered design of the finite difference model forming a Digital Core Laboratory (DCL). Tuning the model to a Special Core Analysis (SCAL) data set, with the objective of filling SCAL data gaps, is expected to be one application.
Behrenbruch, Peter (Bear and Brook Consulting) | Quoc Doan, Truc (Bear and Brook Consulting) | Triet Do Huu, Minh (Bear and Brook Consulting) | Duy Bui, Khang (Bear and Brook Consulting) | Kennaird, Tony (Bear and Brook Consulting)
A detailed comparison is made of the more recently developed phenomenological 2-phase Modified Carman-Kozeny (2pMCK) relative permeability formulation with that of the industry standard, the Modified Brooks-Corey (MBC) formulation. The purpose is to show the strengths and weaknesses of the two formulations and to demonstrate how their combined use can yield the most consistent overall result, the optimum choice as input to reservoir simulation.
A brief overview of the two relative permeability formulations is given first. Several laboratory data sets are reviewed by deploying the two models, validating the data and pinpointing potential problems. In some cases, both methods are used to extrapolate lab results to determine a more realistic residual oil saturation value and corresponding water relative permeability endpoint, particularly for samples involving fines movement. The apparent increased degree of curvature of the oil relative permeability is also often problematic and is typically related to pore-fill, requiring modification to laboratory defined relationships. A clear workflow is outlined on how to derive overall optimal results. Both methods show that if additional information such as independent wettability measurements are available, there is more confidence in final relationships derived. As evident, problems are typically associated with the second data point and the final one, start and finish of multi-phase flow measurements, necessitating adjustment in oil relative permeability curvature and the final endpoint data, residual oil saturation and associated water relative permeability. If the curvature of the oil relative permeability is excessive, MBC extrapolation is prone to failure. While the 2pMCK model does not show such shortcoming, the model cannot currently handle very large exponents. However, for most realistic situations, such limitation is not a problem. Another advantage of the 2pMCK model is its ability to pinpoint laboratory artefacts.
The concurrent use of the two relative permeability formulations gives a new perspective of relative permeability modelling and is particularly suitable for quality checking and analysing more challenging laboratory results. The approach has been computerised, allowing for ease of data handling, model comparison and consistency.
Wu, Shuhong (RIPED, Petrochina) | Fan, Tianyi (RIPED, Petrochina) | Zhao, Lisha (RIPED, Petrochina) | Peng, Hui (RIPED, Petrochina) | Wang, Baohua (RIPED, Petrochina) | Ma, Xuesong (RIPED, Petrochina)
CO2 injection has proven to be a promising technology in enhancing oil recovery in low-permeablility carbonate reservoirs, especially in miscible flooding. How to demonstrate near-miscible/miscible mechanisms and their influence on production performance is a difficult issue to deal with in a compositional simulation. In this paper, considerable research has been conducted on compositional models, especially miscible gas flooding models, by identifying the near-miscible/miscible state and its influence on relative permeability.
This paper describes a multi-component, three-phase, compositional model for simulating miscible CO2 flooding problems. An EOS equation is constructed for phase equilibrium and property calculations. Gibbs stability model is developed to determine whether a hydrocarbon mixture at a particular temperature and pressure is more stable in a single-phase state or in a two-phase state. An interfacial tension (IFT) weighted relatively permeability model is developed to calculate the relative permeability in a near-miscible state. With the help of the Gibbs and Kr models, miscible CO2 flooding in a carbonate reservoir is simulated to evaluate the miscible and immiscible state and demonstrate the production performance.
The Gibbs stability model determines the tangent plane of the Gibbs energy surface at the mixture composition and parallel tangent planes at possible incipient phase compositions. If any of the parallel tangent planes lie below the tangent plane of the mixture composition, a two-phase state will exist. The simulation shows that a Gibbs stability model in CO2 injection can identify whether the hydrocarbon mixture (oil) and CO2 at reservoir pressure and temperature is in one phase state (miscible) or in a two-phase state (immiscible). The IFT weighted Kr model determines an interpolation between immiscible and miscible states using a weighting function, which is a function of IFT with a specified critical gas/oil IFT to control the contribution of the near-miscible effect on kr. The compositional simulation shows that CO2 becomes miscible with a hydrocarbon mixture under the condition of CO2 multi-contact with the hydrocarbon mixture in the reservoir pressure and temperature. Three areas exist between the injector and the producer which are the immiscible area (IFT: about initial value), near-miscible area (IFT: 0.5~2) and miscible area (IFT: lower than 0.5). The residual oil saturation of the miscible CO2 flooding area is around 5%. Oil recovery is enhanced compared with the previous hydrocarbon gas injection.
CO2 injection represents a promising technology to improve production performance and to enhance oil recovery for the carbonate reservoir mentioned in the paper.
In a newly discovered field in the south east of Abu Dhabi, ten Thammama reservoirs were penetrated using a single well and how accurate an understanding of a reservoir and the uncertainties remaining can be assessed by a single well penetrating an accumulation of Hydrocarbon is a frequent question for companies planning to develop new, small reservoirs. The answer depends on how much data the single well got and the other information available about the reservoir and analogue reservoirs. When it comes to Modelling a reservoir that has only been pierced by one well, the assurance on the robustness of the available information is necessary in order to identify gaps and areas of uncertainty and then seek the means to minimize them.
X-1 well drilled at the Crest of the structure and the location defined by utilizing leatest 3D seismic cube and the total depth 11500 ft RTKB. It was penetrated in the Lower Cretaceous sequence including ten reservoirs which are trapped by 4way dip closure. All the main reservoirs were tested in X-1 well and produced with an oil rate of 370 - 4,370 BOPD (30° - 32° API)
An analytical "Lambda" Saturation Height Function was developed and implemented in order to overcome the absence of MICP data for initialization stage of Dynamic Modelling. Numerous choices made to develop a family of models that represents the range of reservoir possibilities. Uncertainty analysis including Analog Data from nearby fields utilized to rank the range of outcomes and prioritize obtaining additional information.
This paper highlight a case study of a three onshore Abu Dhabi discovered structures addressing mitigations toward the challenges to draw a solid conclusion with single well modelling when it comes for development decision by Stakeholders at later stage. Further, the methodology as applied to the case study in this paper, allows identification of the range of outcomes and ranks the additional requirements, ranging from acquiring new Seismic data to drilling new wells, which guide for the next steps to a management decision to develop the reservoir or not.
Hernando, Louis (Poweltec) | Martin, Nicolas (Poweltec) | Zaitoun, Alain (Poweltec) | Al Mufargi, Hilal (Daleel Petroleum LLC) | Al Harthi, Hamood (Daleel Petroleum LLC) | Al Naabi, Ahmed (Daleel Petroleum LLC) | Al Subhi, Khamis (Daleel Petroleum LLC) | Al Harrasi, Mohammed Talib (Daleel Petroleum LLC)
Many horizontal wells in GCC (Gulf Cooperation Council) are producing from fractured carbonate reservoirs under aquifer pressure support. After water breakthrough, these wells suffer from high water cut rise due to water channeling through the fractures, inducing a strong and quick loss of well productivity. The wells being completed horizontal open hole, zonal isolation is very difficult to isolate the water producing zone. Moreover, mechanical water shut off like using liners or open hole packers is normally costly and challenging as well. Therefore, bullheading chemicals is often the only remaining option. The fractures being producing mixture of both oil and water, water shutoff in this type of well is very challenging.
To solve this problem, an original method was used consisting of microgel and gel injections. The strategy is to create a flow barrier deep in the fracture while preserving oil production from the matrix. The optimized procedure enables control of chemical placement by bullhead injection. Laboratory tests were conducted to optimize Microgel and gel chemistry (injectivity, gelation time and gel consistency/stability, return oil permeability).
The treated candidate was 530m long horizontal well and produced at around 600 bpd with water cut of 92%. Temperature and salinity were 62°C and around 100,000ppm respectively. Well behavior indicated existence of fractures along the wellbore.
The treatment consisted of 230m3 of successive Microgel/Gel injection at an average rate of 2.5 BPM. The treatment proceeded bullhead through annulus. WHP remained low throughout injection (150 psi). After curing time of 10 days for gel set, production was released with ramp up rate. The well, which was facing strong production decline, stopped declining, with a water cut drop of around 10%. Incremental daily oil rate was around 50 bopd with cumulative oil for the first 4 months of around 4500 bbls of oil.
This technology may have many applications in fractured carbonate fields and may lead to better oil potential from such high water cut highly fractured reservoir wells, with strong aquifer.
Two of the most critical properties for multiphase flow in a reservoir are relative permeability (Kr) and capillary pressure (Pc). To determine these parameters, a meticulous interpretation of coreflooding and centrifuge experiments is necessary. In this work, a machine learning technique was incorporated to assist in the determination of these parameters quickly and synchronously.
A state of the art framework was developed where a large database of Kr and Pc curves were generated based on existing mathematical models. This database was used to perform thousands of coreflood simulation runs representing drainage steady state experiments. The results obtained from the corefloods including pressure drop and water saturation profile along with other conventional core analysis data were fed as features into the machine learning model. The entire data set was split into 70% for training, 15% for testing, and the remaining 15% for the model validation. The 70% of the training data teaches the model to capture fluid flow behavior inside the core. K-fold cross validation technique was also utilized to increase the accuracy of the model. The trained/tested model was thereby employed to estimate Kr and Pc curves based on available experimental results.
The values of the coefficient of determination (R2) was used to assess the accuracy and efficiency of the developed model. The respective cross plots indicate that the model is capable of making accurate predictions with error percentage less than 2% on history matching experimental data. Furthermore, the latter implies that the AI-based model is capable of determining Kr and Pc curves with less effort and better reliability as opposed to the conventional way of creating an entire simulation model. Additionally, the results include multiple options for Kr and Pc curves from which the best solution can be determined using engineering judgement. This is unlike solutions from existing commercial software, which usually provides only a single solution. The model currently focusses on the prediction of Kr and Pc curves for drainage steady state experiments; however, the work can be extended to capture the imbibition cycle as well.
Kalam, Shams (King Fahd University of Petroleum & Minerals) | Khan, Mohammad (Schlumberger) | Khan, Rizwan Ahmed (King Fahd University of Petroleum & Minerals) | Alam, Mir Muhammad Mansoor (King Fahd University of Petroleum & Minerals) | Sadeed, Ahmed (ENPRO) | Mahmoud, Mohamed (King Fahd University of Petroleum & Minerals) | Abu-Khamsin, Sidqi A. (King Fahd University of Petroleum & Minerals)
Availability of large amounts of data helps in developing data-driven models using state of the art Artificial intelligence (AI) methodologies. Relative permeability is an important parameter used by reservoir engineers and are usually accurately obtained from laboratory experiments, which are relatively expensive. Therefore, AI can play an important role in developing models to predict relative permeability accurately without extensive lab procedures. Accordingly, this work presents application of two AI algorithms namely, Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). Furthermore, two novel mathematical correlations are extracted from the ANN model to predict relative permeability of oil/water in oil- and water-wet environments. The input data, obtained from literature, for the development of AI models include porosity, rock absolute permeability, initial water saturation, residual oil saturation, wettability index and water saturation.
A customized workflow is applied to ensure proper data is fed into the AI models. In addition, a rigorous sensitivity analysis is performed within the framework of this workflow. This analysis involves running multiple realizations with varying number of neurons, resulting in various weights and bias for the ANN model. Moreover, ANFIS model is tuned using various cluster sizes to result in the most optimum value. Finally, the optimized ANN and ANFIS models are compared using the Root Mean Squared Error (RMSE) and correlation coefficient (R2) analysis when applied to a blind dataset comprising of more than 300 data points. The analysis illustrates that the ANN model is relatively better in predicting relative permeability values to both, oil, and water. On the other hand, analysis of the ANFIS model shows that it yields high error values when tested on unseen dataset. Also, unlike the ANN mode, it does not provide an actual mathematical correlation. This work presents alternate data-driven artificial intelligence models which will lead to quicker and cheaper relative permeability estimates.