Babadagli

The August issue of SPE Res Eval & Eng includes eleven papers focusing on three specific areas of petroleum reservoir engineering: reservoir modeling/simulation, reservoir characterization, and recovery mechanisms. A summary of the papers is given below.

Reservoir Modeling and Simulation

The first five papers in this issue cover different areas of reservoir modeling and simulation practices with practical application examples and validations using field cases. The first paper in this category, An Improved Triple Porosity Model for Evaluation of Naturally Fractured Reservoirs, proposes a modification to the existing triple porosity model and provides a field case verification. The proposed model improves the classical triple porosity algorithm by handling the scale associated with the matrix, fractures, and vugs. This allows a more realistic estimation of the cementation or porosity exponent (m) for the composite system. The paper also illustrates the possible errors involved in reserves calculations due to inaccurate estimation of the (m) exponent using two carbonate field examples from the Middle East.

The next paper, Estimates of Fracture Lengths in an Injection Well by History Matching Bottomhole Pressures and Injection Profile, introduces an interesting approach to estimate the fracture length using a reservoir simulation exercise. The authors use the bottom-hole pressures and injectivity of wells in the history matching exercise, also taking into account formation plugging due to suspended solids in the injected water, poro and thermo elastic stresses changes, injection rate changes, shut-downs and restarts, and average reservoir pressure. They finally apply the proposed and tested approach to a field case in Columbia (Guando field). They recommend conducting microseismic surveys, standard well tests, and injection well profiles at different rates and times for a better understanding of injection processes as these data would delineate both induced and other types of fractures (tensile or shear, single or multiple, planar or curved) in injection wells.

Integrated Reservoir Modeling of a Large Sour-Gas Field with High Concentrations of Inerts introduces an interesting case study on the largest acid gas reinjection project in the world. The paper discusses the integration of reservoir characterization studies, geochemical analysis, surveillance data, reservoir simulation, and surface facility network models of the Madison reservoir in the LaBarge field, which is a high percentage sour-gas field. The integrated studies are used to optimize the drilling and depletion plan to maximize methane production with current facilities and to analyze acid gas injection into the aquifer. The detailed exercise presented in this paper enabled the authors to incorporate the design of the next-generation model, which uses a single black-oil simulation over the production and injection areas, and local grid refinement around an increased number of wells.

A new approach, proposed in the paper Candidate Selection Using Stochastic Reasoning Driven by Surrogate Reservoir Models, provides a tool for candidate reservoir selection for improved recovery satisfying physical, financial, geopolitical, and human constraints. After screening more than 1,500 reservoirs for additional recovery potential with waterflooding operations, the authors created a fully stochastic workflow that included stochastic back-population of incomplete datasets, stochastic proxy models over time series, and stochastic ranking methods using Bayesian belief networks. For the sensitivity and uncertainty of the influencing input parameters on the output, numerical models were used and response surfaces as surrogate reservoir models were created, and potential waterflood candidates were ranked as an output. The inclusion of a wide range of influencing parameters while speeding up the screening process without jeopardizing the quality of the results is the major benefit of the approach introduced in this paper.

The last paper of this category, History Matching a Field Case Using the Ensemble Kalman Filter With Covariance Localization, deals with a crucial problem in reservoir simulation: Fast and efficient history matching. The authors applied the ensemble Kalman filter with covariance localization to history-match the production data from a real field case in order to generate multiple realizations of the permeability field, and compared the results with that of a single manually history-matched model. The results showed that one can reduce the computation time with the new approach called the "half-iteration ensemble Kalman filter" without compromising the quality of the results.

Reservoir Characterization

A data mining method to characterize the flow units between injectors and producers in a waterflood application is reported in An Active Method for Characterization of Flow Units Between Injection-Production Wells by Injection Rate Design. The method allows the computation of weight factors representing the influence of any of the injectors surrounding a producer and it is validated using streamline and capacitance models with real data fitting. After using a wavelet approach to design the perturbation in the injection rates and to analyze the observed variations in the production rates, the authors estimated the weight factors and used them to characterize the effective contribution of injection wells to the total gross production. The paper also includes a case study for a tight formation waterflood where the weight factors are intended to detect high permeability channels.

A methodology for formation evaluation using S_w and R_w from the logs of resistivity and sigma if water salinity is not available was proposed in the paper Combining Resistivity and Capture Sigma Logs for Formation Evaluation in Unknown Water Salinity--A Case Study in a Mature Carbonate Field. The proposed technique was tested using induction and sigma logs acquired from a few wells in a mature carbonate field. The results of the proposed technique combining resistivity-sigma logs were in a good agreement with those of carbon-oxygen (C/O) logging. For low-porosity formations, the answers were even more reliable than those of C/O logging.

Data obtained from modern well testing tools can be used for the assessment of reservoir heterogeneity and anisotropy. The paper A Novel Analysis Procedure for Estimating Thickness-Independent Horizontal and Vertical Permeabilities from Pressure Data at an Observation Probe Acquired by Packer-Probe Wireline Formation Testers introduced a new procedure to obtain horizontal and vertical permeabilities only using pressure transient data acquired at an observation probe of a dual-packer-probe wireline formation tester. The procedure uses a new spherical-flow cubic-analysis and applies for all inclination angles of the wellbore in a single-layer or anisotropic (3-D) reservoir without any knowledge of formation thickness or radial flow conditions. Although the method provides unique estimates of horizontal and vertical permeabilities for both vertical and horizontal wellbores, two possible solutions were obtained for the horizontal and vertical permeabilities in the case of the slanted wells. The suggestion made for this type of well is that additional information like core and pretest data is needed to determine the appropriate solution.

Reservoir Dynamics

Three papers under this category report interesting analyses of reservoir behavior using experimental and numerical methods. The first paper, Recovery Mechanisms and Oil Recovery From a Tight, Fractured Basement Reservoir, Yemen, deals with an oil reservoir in a non-sedimentary environment (fractured basement) in which significant and stable production rates (5,000 - 10,000 bbl/day) from eight wells were observed over a five-year period just by depletion. The essential part of the study was the determination and characterization of two sets of fractures (background fractures (BF) with a very low effective permeability of less than 0.001 md and fracture corridors (FC) with an effective permeability of up to several millidarcies) with negligible matrix contribution. Through a series of simulation studies and the available production history, a number of reservoir management strategies were investigated. The authors showed that the low permeability FC and much lower permeability BF yielded only a 14% recovery factor for gas injection. They also reported that the solution gas drive is much more efficient than in conventional reservoirs.

Shales are a very difficult environment for oil recovery and new ideas are needed to recover the tough oil in these types of reservoirs. Flow Rate Behavior and Imbibition in Shale introduces a new chemical imbibition idea using surfactant or brine formulations for oil recovery from shales. The paper tested this idea on an outcrop shale (Pierre Shale) sample from North Dakota. Oil imbibition recovery from dry shale cores could be as high as that of obtained by forced imbibition. The authors reported that an increase in permeability due to mineral dissolution and cracking caused by clay swelling were observed during the tests, which was contrary to earlier studies reporting permeability reduction due to swelling. This is highly encouraging for enhanced oil recovery applications from shales.

In the paper titled Dynamic Asphaltene Behavior for Gas-Injection Risk Analysis, the authors investigated static asphaltene behavior by the solid detection system using a near-infrared light scattering technique in a gas-injection pilot in an offshore carbonate field. They conducted experiments to calibrate the numerical model and evaluate the asphaltene risks in this pilot area, adjusting the target oil composition by considering the existing oil compositional gradient in the field. Upon obtaining inconsistent results with field observations, the authors questioned the applicability of static tests and suggested further studies to understand the dynamic asphaltene behavior for a realistic risk evaluation in this gas injection field. They also emphasized that the highest asphaltene risk is in the early injection stage and this might be reduced by sweeping the remaining oil by injection gas or by restricting the injection pressure below the asphaltene precipitation envelope determined experimentally and numerically.

Tayfun Babadagli
Co-Executive Editor of SPE Res Eval & Eng