In order to debottleneck very large three phase production separators in a Middle Eastern oil field, new high efficiency internals were installed. A Computational Fluid Dynamics (CFD) study was conducted of these three-phase separators. The study focused on the evaluation of internal modifications and the influence of the inflow characteristics. A combination of the Eulerian-Eulerian multiphase and k-e turbulence models was used to simulate the complex behavior of the flow inside the separator. A series of parametric case studies was conducted. The findings of the CFD study were compared with the results of field performance test.
This paper describes the approach adopted in the CFD study. The geometry, mesh, numerical models, simulation strategy as well as the results and findings are discussed in detail.
This paper will describe a new method to facilitate optimization of hydraulically fractured horizontal Bakken completions. A key component of this method is a predictive data-driven model developed from a detailed evaluation of over 50 horizontal Bakken completions. Using this model, production predictions and economic optimization can be quickly performed using only data obtained during lateral drilling operations. This type of modeling and optimization of new horizontal completions fits well with a factory mode (rapid pace of well completion and hydraulic fracturing) of resource development.
The need to create large amounts of fracture area and the associated costs create an opportunity for economy of scale. Unfortunately, the time and expense required performing logging runs to gather data, and estimate reservoir and rock properties to facilitate completion design and optimization reduces operational efficiency and increases well cost. A data-driven approach to optimizing hydraulically fractured horizontal completions does not reduce operational efficiencies since it can be developed from common data and information obtained during drilling operations.
The data-driven model shows that reservoir quality, as one would expect, has the dominate effect on Bakken production. In addition hydraulic fracture design/methodology, fracture spacing, proppant selection, frac staging methods, frac compartment isolation, drillout/cleanup procedures, wellbore length and orientation have a significant effect on well production and economics. Several case histories will be included in which data-driven modeling was used to evaluate various completion/frac scenarios, predict production and facilitate completion/frac optimization.
The challenge of optimizing horizontal Bakken completion and frac design (30 or more hydraulic fracturing treatments per well) in a fast paced factory mode of resource development is problematic. The methodology described in this paper has been helpful to identify production drivers and facilitate well planning while also supporting a factory mode of resource development.
Elsherif, Tamer Ahmed (Schlumberger Middle East SA.) | Balto, Abdulelah Adel (Saudi Aramco) | Baez, Fernando (Schlumberger Middle East SA.) | Al-Qahtani, Hassan B. (Saudi Aramco) | El-Kilany, Khaled Ahmed (Saudi Aramco)
Fiber optic enabled coiled tubing (FOECT) has been commonly used in qualitatively evaluating reservoir matrix chemical treatment in real time during the past couple of years. During this period, attempts of transforming qualitative evaluations to quantitative ones were made. The quantitative evaluation is based on two simultaneous criterions. The first one is a downhole pressure diagnostic plot (pressure transient analysis) created instantinuously using real-time acquired data by the downhole gauges. The second is an estimate of the zonal coverage based on the resulting temperature profile plot before, during and after a pumping treatment. Pressure transient analysis gives the skin as a direct output, while the cooling down/warming up DTS profiles identifies where the treatment fluids went in the formation, hence identifying the damaged zones.
It is strongly recommended to combine well testing analysis techniques with zone coverage evaluation in highly deviated and horizontal completed wells in both clastic and non-clastic rocks. Basically, deriving the skin from the injectivity test (pretreatment) and the skin from the post flush (post-treatment) provides an evaluation matrix treatment effectiveness.
A comparison between formation damage "skin?? before and after the treatment was performed on the spot, revealing positive results of nearly uniform distribution of treatment fluids, and skin value reduction across the 3400 ft horizontal section.
Following the innovative procedures executed in well-A, different techniques were proposed, providing time and cost savings; raising the operational excellence expectations levels higher than expected for an offshore environment. The application of FOECT technology helped to minimize uncertainties during treatment evaluation, and enhanced treatment distribution and placement. In addition to establishing more accurate and reliable Nodal Analysis and production forecast models.
"Sandstone matrix chemical treatment is the process of attempting to restore the original permeability of the rock that has been altered by the effect of formation damage. Eliminating formation damage leads to reduce the pressure loss in the wellbore sand face and in the critical matrix that will have a direct impact on increasing well's productivity??. Reservoir Stimulation, Second Edition, M.J.Economides, K.G.Nolte, 1989. The near wellbore matrix is where most of the permeability damage occurs from mud filtrate, adverse oil or water saturations, and clay migration. Sandstone matrix treatment is faced by multiple challenges such as; uncertainty in the lithology of the rock, uncertainty in damage identification, secondary reactions that may cause blocking rather than stimulating.
Dong, Chengli (Shell International E&P (Rijswijk)) | Petro, David Robert (Marathon Oil) | Hayden, Ron S. (Schlumberger) | Zuo, Julian Youxiang (Schlumberger) | Pomerantz, Andrew (Schlumberger) | Mullins, Oliver C. (Schlumberger)
Characterization of complicated reservoir architecture with multiple compartments, baffles and tortuous connectivity is critical; additionally, reservoir fluids undergo dynamic processes (multiple charging, biodegradation and water/gas washes) that lead to complex fluid columns with significant property variation. Accurate understanding of both reservoir and fluids is critical for reserve assessment, field management and production planning. In this paper, a methodology is presented for reservoir connectivity analysis, which integrates reservoir fluid property distributions with an asphaltene Equation of State (EoS) model developed recently. The implications of reservoir fluid equilibrium are treated within laboratory experimentation and equation of state modeling. In addition to cubic EoS modeling for light end gradients, the industry's first asphaltene EoS the Flory-Huggins-Zuo EoS is successfully utilized for asphaltene gradients. This new EoS has been enabled by the resolution of asphaltene nanoscience embodied in the Yen-Mullins model. Specific reservoir fluid gradients, such as gas-oil ratio (GOR), composition and asphaltene content, can be measured in real time and under downhole conditions with downhole fluid analysis (DFA) conveyed by formation tester tools. Integration of the DFA methods with the asphaltene EoS model provides an effective method to analyze connectivity at the field scale, for both volatile oil/condensate gas reservoirs with large GOR variation, and black oil/mobile heavy oil fields with asphaltene variation in dominant.
A field case study is presented that involves multiple stacked sands in five wells in a complicated offshore field. Formation pressure analysis is inconclusive in determining formation connectivity due to measurement uncertainties; furthermore, conventional PVT laboratory analysis does not indicate significant fluid property variation. In this highly under-saturated black oil field, measurement of asphaltene content using DFA shows significant variation and is critical for understanding the reservoir fluid distribution. When integrated with the asphaltene EoS model, connectivity across multiple sands and wells is determined with high confidence, and the results are confirmed by actual production data. Advanced laboratory fluid analysis, such as two-dimensional gas chromatography, is also conducted on fluid samples, which further confirms the result of the DFA and asphaltene EoS model.
In recent years, there has been a growing recognition that complex reservoir architectures and complex fluid distributions are often the norm. Reservoir fluids undergo many dynamic processes over time, which may lead to highly complex oil columns. Factors that give rise to fluid complexities include current/multiple reservoir charging, biodegradation, water/gas washes, and leaky seals. Reservoir architecture is often complex with multiple compartments, baffles and tortuous connectivity. Because the reservoir fluids often exhibit large variations, reservoir compartmentalization can appear as stair-step discontinuous fluid properties. In contrast, well connected reservoirs exhibit smooth distributions of reservoir fluid properties. Analysis of reservoir fluid property distribution often functions as a proxy for reservoir connectivity in field scale, which has been supported by many field case studies.
Rui, Zhenhua (Independent Project Analysis) | Metz, Paul A. (University of Alaska Fairbanks) | Wang, Xiaoqing (University of Alaska Fairbanks) | Chen, Gang (University of Alaska Fairbanks) | Zhou, Xiyu (University of Alaska Fairbanks) | Reynolds, Douglas B. (University of Alaska Fairbanks)
The aim of this paper is to investigate pipeline compressor station project cost overruns. A total of 220 pipeline compressor station projects constructed between 1992 and 2008 have been collected, including material, labor, miscellaneous, land, and total costs, compressor station capacity, location, and year of completion. Statistical methods are applied to identify the distribution of cost overruns and overrun causes. Overall average overrun rates of pipeline compressor station material, labor, miscellaneous, land and total costs are 0.03, 0.60, 0.02, -0.14, and 0.11, respectively. Cost estimations of compressor station construction components are biased except for the material cost. In addition, the cost estimation errors of underestimated compressor station construction components are generally larger than those of overestimated components. Results of the analysis show that compressor station project size, capacity, location, and year of completion have different impacts on individual construction cost component cost overruns.
The Ompoyi and Orindi fields are located 5 km offshore Gabon in a water depth of 20 m. The high oil density (23o API) and high produced water salinity (150 kppm) combined with reservoir pressure depletion meant that primary production of the Ozouri reservoir required artificial lift. Initially, progressing cavity pumps (PCPs) were selected, however, following failures of the elastomers, electric submersible pumps (ESPs) were deployed to produce at higher rates with high gas/oil ratios (GORs). The production instability associated with the dual-porosity reservoir behaviour and high free-gas content in both the inflow and outflow of the wells presented the main challenge to ESP design and operation. Additionally, most of the wells were remote from the main production facility and were produced via multiphase subsea flowlines, which made well testing difficult due to the phase segregation in the flowlines.
To achieve economically viable production utilizing ESPs, innovative use was made of a range of existing technology. Operationally, however, real-time monitoring was essential to setting wellhead pressures and pump speed to maximize drawdowns. Key technology elements were monopods to minimize the offshore structure cost, fit-for-purpose subsea power cable and helicoaxial downhole pumps for high gas void fraction operation. Key lessons were learned following the trial of several completion architectures to find the optimal combination of gas venting, reservoir access, and a dual barrier mechanism.
The first well was put on production in 2002. Since then a further 10 wells have been drilled to reach a total liquid production of 8,000 B/D with a 60% water cut. Production has been economical and thus confirmed ESPs as being the right solution for this reservoir. The lessons learned prove that the application of ESPs is not limited to traditional waterfloods and that it is feasible to produce challenging reservoirs with ESPs.
Pettitt, William (Itasca Consulting Group Inc) | Damjanac, Branko (Itasca Consulting Group) | Hazzard, Jim (Itasca Consulting Group) | Han, Yanhui (Itasca Houston Inc.) | Sanchez-Nagel, Marisela (Itasca Houston Inc.) | Nagel, Neal Borden (Applied Seismology Consultants) | Reyes-Montes, Juan Miguel (University of Toronto) | Young, R. Paul
Observations from field monitoring indicate the presence of existing fracture networks significantly affect, and may control, the history of hydraulic treatments. Existing fractures allow pathways for treatment fluids to migrate efficiently through the formation, stimulating connectivity to larger reservoir volumes. This paper examines the relationship between hydraulic treatments and fracture networks, and investigates the ability to engineer the hydraulic conductivity of the stimulated volume. Fracture network engineering (FNE) involves the engineering design of rock mass disturbance through the use of advanced techniques to model fractured rock masses numerically, and then correlate field observations with simulated fractures generated within the models. Modeling algorithms accurately address the hydromechanical physics of hydraulic treatments developed across a range of applications in rock engineering.
A Synthetic Rock Mass (SRM) model is constructed by explicitly defining a discrete fracture network within a modeled rock matrix. Hydraulic treatment into the SRM allows fracture dilatancy, propagation and shearing to be realized in the emergent behavior of the bonded particle model in three dimensions and at the scale of the treated volume. Simulated microseismicity is generated when the fractures are disturbed or propagated providing a method for correlation with field observations. Illustrative models show how hydraulic treatments stimulate fracture networks and generate new hydraulic fractures in volumes not expected in conventional design analysis connecting existing fractures with preferential alignment. By combining SRM models with field observations it is possible to investigate the relation and sensitivity of hydraulic treatments to existing fracture networks, and therefore, engineer the most optimal use of these fractures in the performance of a given project. Developmental challenges are discussed, including the sensitivity of the techniques to fracture network uncertainties and the accurate representation of microseismic source data within the models.
Hydraulic fractures are traditionally modeled as planar features developed by the tensile failure of the rock. Laboratory nanoseismic and field mine-back studies show that most of the fractures are non-planar complex features. Fracture properties are strongly affected by the magnitudes and directions of the stresses in the formation. Low stresses are associated with a complex fracture development while high stresses create simpler, straighter and more planar fractures. We report the results of controlled laboratory triaxial hydraulic fracturing experiments instrumented with piezoelectric sensors. We performed tests on Lyons sandstone which was determined to have an initially isotropic velocity structure. The fracturing experiments have been performed under triaxial stress state to replicate the insitu stress reservoir conditions. The uncertainty in hypocenter locations, frequency analysis, source mechanisms and the effects of stress on fracture propagation will be discussed. Microscopic observations of the fractures are correlated with the mapped microseismic events. Fractures are observed to be non-planar visually and at the SEM scale. Shear failure recorded by focal mechanisms appears to dominate the failure mode. The deviation from planarity will surely affect proppant transport and dispersement.
Historically, and across the E&P industry, several projects were found to underperform in their actual performance as compared to the promises made at the time of project approval. Investment decisions on upstream projects rely, to a large extent, on the robustness of the predicted ultimate recovery and production forecast associated with the chosen development concept. A 2010 upstream industry benchmarking consortium reviewed global performance of E&P projects and concluded that the majority of production forecasts presented at project approval were not attained. In many cases, this underdelivery resulted in eroded project economics and reputation damage with stakeholders.
The challenge for the E&P industry is to ensure that project approvals are based on realistic forecasts (P50 case). Modern technology has offered a paradigm shift in the amount of detail which can be included in the subsurface static and dynamic models. However, a universal truth is that no single model is the perfect representation of reality, and the recovery processes based on modelling are, therefore, always approximations of real life situations. Simulation models are guiding tools rather than sources of perfect answers.
This work is intended to increase awareness amongst the forecasters and the decision makers about pitfalls associated with production forecasting (especially those generated by dynamic models). It provides insight into the need to condition model input data or alternatively the model output to ensure that forecasts and resource volumes generated by these models are realistic. Root causes of unrealistic forecasts are identified. Dos and don'ts are described that help instil realism in the forecasting process. Pragmatic techniques to adjust the output obtained from dynamic modelling are illustrated using real-life examples.
This work presents a pragmatic approach to create reliable long-term production forecasts which can form the basis for sound business decisions.
Gas decline curve analysis has traditionally relied on the use of liquid (oil) type curves along with the concepts of pseudo-pressure and pseudo-time and/or empirical curve fitting of rate-time production data using Fetkovich type curves for gas well performance and reserve estimations. In this work, we show that a single-line, universal type curve can instead be derived for the analysis of unsteady state of natural gas wells under boundary dominated flow (BDF). The proposed decline equation— presented in its analytical, closed functional form—models production from any gas well producing at constant bottom-hole pressure under BDF. On the basis of this decline model, a universal type-curve is formulated for type curve matching of natural gas well producing under BDF, for any draw-down condition, and regardless of intrinsic reservoir and fluid properties. New-generation analytical procedures for gas well performance analysis are presented, which does not necessitate the calculation of pseudo-pressure or pseudo-time. Explicit OGIP predictions are thus enabled from the proposed universal type curve matching. The importance of conducting unambiguous reserve estimations based on rate-time data and reliably predicting future well performance cannot be overstated. The proposed single-line type curve is demonstrated to do so by successfully matching rate-time production BDF data, predicting future performance, and reliably estimating fluid in place for a large number of numerical simulations and field cases. It is finally shown that the proposed formulation can be alternatively implemented in terms of straight-line analysis of 1/qsc b versus time plots that enable future well performance and explicit reserves estimations.