|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Abstract The high residual gas trapping characteristics of low porosity rock are a considerable challenge for recovery of fields with bottom or edge water. Water lifting and handling is often a practical issue. Perforated intervals may suffer from permanent loss of gas mobility related to hysteresis effects due to imbibition of water (either production induced or paleo-residuals). The discrimination of "trapped" versus mobile gas intervals in a low porosity setting is not straight forward because data can be very deceiving: water imbibed zones may look alike free gas zones, but performance and recovery are quite different This paper presents a show case of a clastic low porosity formation buried at 4Km depth. Following disappointing performance of the wells due to unexpected water production, the interpretation of pre-production fluid-fill-cycle revealed that a large percentage of the GIIP had been trapped over geological time due to tilting of the structure. Ignoring paleo-imbibition resulted in over-prediction of the recovery and misplacement of wells. Further imbibition of water into prolific zones during production of the field may result in early irreversible water loading of the wells with initial high rates. A cross-discipline workflow has been put together to quantify field potential and water risk. A fit for purpose reservoir simulation with a focus on the fluid-fill-cycle can explain the performance of the wells to date and predict the expected recovery from paleo-residual volumes. Down flank co-production of gas and water has been proposed as mitigation for water blockage issues: 1) to potentially delay further water cusping through prolific zones and 2) to enhance remobilization of gas trapped in the paleo-residuals zone. The reported example illustrates the risks of water for gas recovery in low porosity formations, as learnt from a field with paleo-imbibition. We describe diagnostic tools for timely discrimination of paleo-trapped gas volumes to be considered prior to investment decisions, well placement and completions; we suggest fit for purpose models that can be constructed to represent rock and fluid physics without embarking into comprehensive full-field full-hysteresis models; we draw attention to specific permeability modeling guidelines for reservoirs containing significant volumes of low porosity rock to avoid overestimation of production plateau time. We recommend the planning of co-production of gas and water in flank wells to safeguard production from prolific zones in low porosity reservoirs with an aquifer.
Abstract In order to better understand reservoir behavior, reservoir engineers make sure that the model fits the data appropriately. The question of how well a model fits the data is described by a match quality function carrying assumptions about data. From a statistical perspective, improper assumptions about the underlying model may lead to misleading belief about the future response of reservoir models. For instance, a simple linear regression model may have a fair fit to available data, yet fail to predict well. On the contrary, a model may perfectly match the data but make poor prediction (i.e. overfitting). In both cases, the regression model mean response will be far from the true response of the reservoir variables and will cause poor decision making. Therefore, a suitable model has to provide balance between the goodness of the fitted model and the model complexity. In the model selection problem, realistic assumptions concerning the details of model specification are the key elements in learning from data. With regard to conventional history match scheme, the data fitting is usually performed by linear least-squares regression model (LSQ) which makes simple, yet often unrealistic, assumptions about the discrepancy between the model output and the measured values. The linear LSQ model ignores any likely correlation structure in discrepancy, changes in mean and pattern similarities reflecting on poor prediction. In this work, we interpret the model selection problem in data-driven settings that enables us to first interpolate the error in history period, and second propagate it towards unseen data (i.e. error generalization). The error models constructed by inferring parameters of selected models can predict the response variable (e.g. oil rate) at any point in input space (e.g. time) with corresponding generalization uncertainty. These models are inferred through training/validation data set and further compared in terms of average generalization error on test set. Our results demonstrate how incorporating different correlation structures of errors improves predictive performance of the model for the deterministic aspect of the reservoir modelling. In addition, our findings based on different inference of selected error models highlight an enormous failure in prediction by improper models.
Abstract A workflow applied to achieve a multi-scale characterisation of a carbonate reservoir is presented. Carbonate rocks are strongly heterogeneous due either to complexity of the primary fabric or to diagenetic over-printing. The combination of these features leads to complicated pore systems, thus a proper definition of pore types using either pore size or pore throat size distributions, is important to indirectly capture diagenetic modifications and to get a link to dynamic properties. A new approach was developed in order to define a Rock Type classification (RRT) each time the approaches based on Winland's and Hydraulic Flow Unit methods do not give a reliable core facies characterisation when moving to the log scale. Moreover, the proposed workflow accounts for stratigraphy and seismic since RRT are linked to the elastic properties. In the new MICP-based Rock Typing workflow, RRT are identified by describing dominant pore types using mercury injection (MICP) curves parameterisation and routine core data (RCA). Clustering and subsequent extrapolation of MICP derived RRT to RCA samples, are the first two stages to achieve a predictable classification into the log domain. Log RRT are then defined at the log scale using curves of elastic properties, like Poisson's Ratio (PR), Frame Stiffness (fk) and Flexibility (γk) Factors. These elastic parameters (calculated with the Extended Biot Theory), can capture the effects of pore structure on the petrophysical properties and link RRT prediction at well position to seismic attributes. Since the RRT are characterised in the elastic space, the facies model – properly upscaled – represents the basis to classify elastic attributes from seismic inversion in a Bayesian framework. The seismic classification can then be used as a driver for RRT distribution in the inter-well space into the 3D model. A further benefit is the direct relationship to the original RRT porosity/permeability distributions, when modelling petrophysical properties. This new workflow was a successful solution to define homogeneous reservoir intervals in a carbonate environment characterised by the lack of a significant relationship between depositional facies and petrophysical properties.
Abstract Fractures often influence production behaviour in hydrocarbon reservoirs, yet the pressure transients observed in the wells may not show the conventional well-test signatures. In this case, the effect of fractures on production would be misinterperted or even completely missed. Fracture networks are commonly multi-scale and properties including aperture (or conductivity), length, connectivity and distribution vary greatly within a reservoir. The heterogeneous nature of fractured reservoirs make them very difficult to characterise and develop. In addition, the location of a producer within the fracture network also control flow rates and affect the pressure response; however, conventional well-test analysis assumes that the producer is located in symmetrical fracture networks. To improve our understanding of fracture flow behaviour from well-test data, and in order to better characterise the impact of fractures on reservoir performance, we investigate the effect of variations in fracture conductivity and location of the producer in the fracture network on the pressure transient responses. Naturally fractured reservoirs (NFR) with well-connected fracture networks are traditionally simulated using the Dual-Porosity (DP) model. However, several studies have shown that the classic DP response (V-shape) corresponding to the DP model is an exceptional behaviour applicable only to certain reservoir geology and does not apply to all NFR. To overcome the limitations of the characteristic flow behaviour inherent to this model, we employ Discrete Fracture Matrix (DFM) modelling technique and an unstructured-grid reservoir simulator to generate synthetic pressure transients in all fracture networks that we analysed. Our rigorous and systematic geoengineering workflow enables us to correlate the pressure transients to the known geological features of the simulated reservoir model. We observed that depending on the location of the producer in the fracture network and the properties of the fractures that the producer intercepts, the synthetic pressure transients vary significantly. We therefore use these insights to quantify the impact of variation in fracture conductivity and producer location on fracture flow behaviour and systematically present interpretations to these behaviours. Our findings enable us to interpret some unconventional features of intersecting fractures with variable conductivity. We observed that the behaviour of two intersecting fractures where the well asymmetrically intercepts a finite-conductivity fracture can be similar to that of a well intercepting a fracture in a connected fracture network with uniform fracture conductivity. Furthermore, a well intercepting a finite-conductivity fracture in NFR with both finite- and infinite-conductivity fractures would yield a dual-porosity response that may otherwise be absent if the fracture network is assumed to have uniform conductivity.
Erzuah, Samuel (National IOR Centre of Norway, University of Stavanger) | Fjelde, Ingebret (International Research institute of Stavnger, UiS, National IOR Centre of Norway) | Voke Omekeh, Aruoture (IRIS, National IOR Centre of Norway)
Abstract The wetting properties of the reservoir rocks are governed by the tendency of the individual minerals constituting the reservoir rock to adsorb oil during crude oil/brine/rock (COBR) interactions. To explore the oil adhesion kinetics during COBR interactions, one approach is to assess the oil adhesion tendencies of the individual minerals. The aim of this presented study was to characterize the wettability by determining the oil adhesion tendencies of the minerals using Quartz Crystal Microbalance with Dissipation (QCM-D). The kinetics of the mass (Δmads) and the thickness (Δt) of the adsorbed film were modelled mathematically using the Sauerbrey relation with the QCM-D output as input. In addition, we present Surface Complexation modelling (SCM) evaluation of possible electrostatic linkages of the studied COBR system. The kinetics of oil adsorption during COBR interactions were prominent during Formation Water (FW)/Stock Tank Oil (STO)/FW injection sequence with kaolinite sensor as compared to that of quartz. This was depicted by the relatively high change in the FW frequency signal (Δf) before and after the injection of STO with kaolinite sensor as compared to quartz. Negligible change in the frequency signal (Δf≈ 0) was observed during the various injection sequence with quartz sensor. This suggested that minor adsorption has taken place, thus confirming the hydrophilic nature of the quartz sensor. The mathematical modelling of the thickness (Δt) and the mass (Δmads) of the adsorbed film also reveals that kaolinite is more oil wet than quartz. This is portrayed by the relatively high magnitude of the adsorbed oil on kaolinite (Δt = 6nm - 14nm and Δm = 1600ng - 3500ng). The SCM results also confirm negligible (≈ 0.008) electrostatic pair linkage for the quartz sensor as compared to kaolinite (≈ 0.3). This shows that the tendency for oil to be adsorbed onto kaolinite sensors were relatively high as compared to quartz. The electrostatic pair linkages reveal that the dominant electrostatic pair linkage existing between the mineral- brine and the oil-brine interface was cation bridging by divalent cations such as Ca and Mg. Hence, it was not surprising that the FW/STO/FW injection sequence for all the three (3) methods were relatively oil-wet as compared to similar sequence of optimum LSW composition. This was attributed to the abundance of Ca and Mg to bridge the two negatively charged surfaces in the former than in the latter.
Abstract Harmonic Pulse testing was developed as a form of well testing that can be applied during ongoing production or injection operations. A pulsed signal is superimposed to the background pressure trend thus no interruption of well and reservoir production is required before and during the test. The pulsed pressure and rate signal analysis is performed in the frequency domain; to this end, the pressure and the rate signals need to be decomposed into harmonic components. The derivative of the harmonic components in the frequency domain can then be analyzed similarly to a conventional well test. In practice the interpretability of the derivative of the harmonic components can be significantly improved if the pressure trend to which the pressure pulses are superimposed is removed, i.e. a detrending of the pressure data is performed prior to well test interpretation. In the present paper, the results obtained after applying different detrending methodologies to pressure data recorded during pulse tests in different reservoir conditions are presented and discussed. Analyses on synthetic test data proved that polynomial detrending is effective in removing the pressure trend induced by field depletion and constant well interference but cannot deal with transient effect related to preexisting rate history or ongoing production changes. Conversely, some of the detrending algorithms based on a heuristic approach are very effective to remove both. Moreover, detrended data can be further regularized by excluding anomalous cycles from the analysis, i.e. cycles that do not respect the designed test periodicity, such as in the case of well interference and/or temporary interruption of the pressure pulses during the execution of the test. The adoption of an effective detrending strategy can considerably improve the quality of the pressure data obtained from harmonic pulse tests and thus the test interpretability. Therefore, it offers the possibility of overcoming the limitation of applicability due to the difficulty of imposing a regularly pulsing rate for the whole test duration (typically lasting several days). This makes harmonic pulse tests very attractive for well performance monitoring, as in gas storage fields.
Abstract This paper presents the results of an experimental study on the effects of various CO2-injection modes on immiscible flooding performance in heterogeneous sandstone porous media. Core flooding experiments were conducted for n-Decane – synthetic brine – CO2 systems at 9.6 MPa backpressure and 343 K to attain immiscible flooding condition (Minimum miscibility pressure (MMP) of CO2 in n-Decane is 12.4 MPa). For this purpose, two sets of heterogeneous sandstone core samples were assembled with heterogeneity in either parallel (layered samples) or perpendicular (composite samples) to the flow. The results obtained for both composite and layered core samples indicated that heterogeneity tremendously influences the outcome of the CO2 enhanced oil recovery (CO2–EOR). Oil recovery decreases dramatically with increase in the heterogeneity level or permeability ratio (PR). For instance, a recovery of 76.04% in a homogenous sample decreases to 72.15%, 65.97% and 54.71% when in layered samples the PR increases to 2.5, 5, and 12.5, respectively. In addition, the crossflow in layered core sample is found to have a noticeable effect on the ultimate oil recovery (increasing oil recovery up to 5%). It is also worth noting that for the composite samples when we arranged the plugs by putting the low permeability segments closer to the sample outlets the recovery factor increased (e.g. the recovery of 68.32% for Low-High-Low arrangement versus 64.37% for High-Low-High configuration). However, regardless of the segment arrangements, the recoveries in composite cores are less than that obtained from the homogeneous core sample (76.04%)). Reservoir heterogeneity plays a critical role in determining the successes of the EOR processes, but its effect has rarely been comprehensively quantified in the laboratory. The limited experimental studies conducted to date seem to suffer from a number of deficiency mainly associated with sample preparation and experimental setup. In the present work, in addition to investigating a number of factors rarely studied experimentally before (e.g. effect of crossflow), attempts have been made to overcome the deficiencies of previous studies. Thus the results of this study can be insightful in overcoming the current challenges in capturing the importance of geological uncertainties in the current and future EOR projects.
Abstract Analysing well test data from gas-condensate reservoirs can be challenging, particularly when pressure drops below the dew point and condensate forms near the well. Usually, a single-phase pseudo-pressure approach is used to analyse the corresponding pressure transient well test data in combination with a composite reservoir model, but this approach has its shortcomings. A two-phase pseudo-pressure approach has been shown to give better results, but it requires relative permeability (kr) data which are not readily available. The paper demonstrates the reliability of calculating a pressure versus radius profile from pressure versus time build up data from tests on wells in gas-condensate reservoirs. The calculation is based on the probe radius concept originally proposed for single-phase systems. It also evaluates the possibility of obtaining, analytically, kr data affected by pressure and velocity in single-rate and multi-rate tests. The calculation is based on Darcy's law using the pressure derivative with respect to distance obtained from the pressure versus radius profile. The theory of these calculations is verified by comparing the results with the output from a numerical simulator based on a single-layer single-well radial reservoir model. A simple binary rich gas condensate fluid was used in the numerical simulation. The results demonstrate that the pressure versus radius profile using probe radius theory reasonably matches the corresponding pressure versus radius profile output from the numerical simulation. This also shows that the method can reliably predict the condensate bank radius. At first, the analytical approach used to calculate kr, as a function of radius, resulted in oscillations due to the nature of derivative calculation involved. The source of the kr noise was treated following two separate approaches that can be used depending on the level of noise versus ease of use. The calculated kr versus radius data showed good agreement with the corresponding grid block kr versus radius data predicted by the simulator for various rock types and flow rates. The method described in this paper provides a practical tool to estimate kr data that can be used to calculate two-phase pseudo pressure and hence potentially improve the analysis of gas-condensate test data affected by condensate banking and velocity. The derived kr data may also have other applications.
Abstract In this paper we describe new and innovative flow diagnostics tools for dual porosity models for naturally fractured reservoirs. Our new diagnostic tools allow us to compare and rank large numbers of geological models based on their approximate dynamic response in almost negligible time. Fast ranking methods allow us to select a representative ensemble of models that quantify geological uncertainty for robust production forecasting via full physics reservoir simulation. Reliable production forecasting for fractured carbonate reservoirs is a challenge. Natural fractures, adverse wettability and complex matrix heterogeneity are all highly uncertain and can all negatively impact upon recovery. Ideally we should consider a large and diverse ensemble of reservoir models to quantify the impact of geological uncertainty on reservoir performance. However, the computational cost can be significant, especially for dual porosity/permeability models. A brute force approach using powerful workstations, clusters or cloud computing can be taken to reduce the time investment. But this is not always possible, rendering robust uncertainty quantification impractical for many asset teams. Often only a small subset of scenarios is considered which may collapse into a single base case, from which development decisions are made. Base cases often fail to predict future production, need frequent modifications, lack geological realism and provide incomplete risk assessments, often causing asset teams to miss economic opportunities. Flow diagnostics can provide dynamic reservoir information in a fraction of the time for full physics simulation. We propose a workflow where we utilise flow diagnostics as a ranking tool to complement forecasting using reservoir simulation throughout. Our approach addresses the model run time, allowing us to use standard hardware. Flow diagnostics solve simplified physics to approximate the dynamic response of the reservoir, from this we can calculate and visualize key dynamic properties (e.g., time-of-flight, drained and swept reservoir volumes, time-to-breakthrough, decline rates, sweet spots, well-allocation factors). Flow diagnostics provide robust indicators of dynamic heterogeneity that allow us to select a diverse ensemble of models that captures the range of uncertainty. In this work, novel diagnostics utilising physically based transfer models have been developed to account for the fracture-matrix exchange, which otherwise could only be obtained from lengthy simulation. A new Damköhler number based metric DaDP links the advective time-of-flight in the fractures to the transfer from the matrix. DaDP identifies fast and slow draining regions of the matrix, stagnant regions within the fracture network and wells at risk of water breakthrough. This information can subsequently be used to optimise well placement and rates to maximise production and delay water breakthrough.
Tian, Wei (China University of Petroleum) | Liu, Huiqing (China University of Petroleum) | He, Shunli (China University of Petroleum) | Wang, Jing (China University of Petroleum) | Zhang, Zequan (China University of Petroleum) | Zheng, Majia (PetroChina Southwest Oil&Gas Company Limited) | Ning, Zhengfu (China University of Petroleum)
Abstract Shale gas is considered very important unconventional hydrocarbon resources. Due to the technological advances of hydraulic fracturing, the development of shale gas has become the main focus in recent years. Porosity and permeability are the most important petrophysical parameters during the production of shale gas. A considerable amount of research work has been carried out on stress law of porosity and permeability. However, nearly none of them considered the effects of methane adsorption. This paper utilizes the simplified local-density (SLD) theory to study adsorption of supercritical gas in shale gas reservoirs. On the basis of the basic features of high pressure supercritical adsorption of shale gas, Peng—Robinson equation is used to describe adsorbed fluid. The interaction between the gas molecules and porewalls of shale is considered using Lennard-Jones potential. Finally, we establish the SLD model to do regression analysis for the adsorption experiments data. The density of adsorbed phase and free phase density could be obtained applying SLD model and then the amount of gas adsorption can be determined. The Gibbs adsorption amount calculated using the SLD model is used to establish matrix strain model. Finally, the strain model is incorporated into widely used analytical porosity and permeability models to develop the coupled model with consideration of the coupled effect of gas adsorption and stress on the porosity and permeability of shale gas reservoirs. And the trend of variation of porosity and permeability of shale rocks taking account of the effects of stress and gas adsorption can be obtained. Lab experiments of gas adsorption of methane gas are made on three shale samples. The developed SLD model is applied to describe gas adsorption data. The outcome indicates that SLD model can properly analyze and fit the experimental data. From the results calculated by the new developed model of porosity and permeability, we can conclude that porosity ratios and permeability ratios of gas shales decrease with the increase of pore pressure, which is contrary to the tendency of changes in porosity and permeability only taking account of the effects of effective stress. This result demonstrates that gas adsorption has very large impact on pore volume, therefore the deformation of matrix induced by the adsorption of methane gas cannot be neglected. The proposed model could further be used for the accurate evaluation of storage capacity of shale gas reservoirs and gas production of wells.