However, other technologies can often be employed to investigate properties of the earth that correlate better with the properties of interest. If the images from these technologies can be provided at appropriate resolution, and if the knowledge required for interpretation and wise application of these technologies is available within the industry, they should be used. For example, electrical methods are extremely sensitive to variations in saturation, yet surface-based methods provide very poor resolution. Reservoir compaction can be directly observed from surface deformation, and pore-volume or gas-saturation changes can be detected from changes in the gravitational field. Dramatic examples of surface deformation induced by reservoir compaction have been provided by releveling studies (involving repeated high-accuracy surveying) and satellite-based interferometry.
Differential compaction is an inherent process in carbonate systems that is thought to produce early natural fractures prior to any significant burial. Such fractures can persist and can be major permeability pathways, including areas of minor tectonic overprint. We forward model differential compaction fracturing in a carbonate reservoir in effort to predict the location of fractures in the subsurface.
3D finite-element geomechanical models are created to simulate differential compaction fracturing at a carbonate platform scale (kilometers) and the smaller carbonate build-up scale (10s of meters) commonly present within carbonate platforms. Interpreted seismic surfaces of key reservoir horizons are used as an input for the platform-scale model. Geometry of carbonate build-up from an outcrop analog is used for the build-up scale models. In both type of models layers identified to be compaction prone are restored to their expected pre-compaction state. A simplified mechanical stratigraphy scheme is adopted to distribute mechanical properties within the models consistent with their expected pre-burial properties.
Geomechanical modeling in this study was applied to a field which includes two carbonate platforms at different stratigraphic levels. Modeling results predict increased fracture intensity at the windward margin of the carbonate platform. This coincides with increased windward-leeward asymmetry of an underlying older platform. Increased fracture intensity is predicted at the center of the platform where the underlying older platform displays significantly less asymmetry. Predicted fracture locations over the platform top also correspond with the location of carbonate build-ups identified from seismic data. Fracture observations from image logs and indirectly from mud loss data within the upper platform are consistent with our modeling results. Predicted areas of greatest fracture intensity correspond with the location of wells with the highest fracture intensity observed from image logs.
Build-up scale models suggest that the build-up shape exerts a major control on the resulting differential compaction fracture pattern. Elongate build-ups tend to produce fractures oriented parallel to their axes. Circular build-ups tends to produce radial fracture patterns. Fracture orientation from image logs along with build-up shape observed using the coherence seismic attribute are consistent with these findings.
This study offers a process-based fracture modeling approach that can enhance the predictability of the location and orientations of natural fractures in carbonate reservoirs.
Mahajan, Sandeep (Petroleum Development Oman) | Stammeijer, Jan (Petroleum Development Oman) | Mukhaini, Hamed (Petroleum Development Oman) | Azri, Saif (Petroleum Development Oman) | Rahmoune, Rachid (Petroleum Development Oman) | Aamri, Mohammed (Petroleum Development Oman) | Tarmizi, Ikhsan (Petroleum Development Oman)
One of the PDO's largest producing fields in Oman consists of three stacked reservoir formations, two of which are currently producing while deeper reservoirs are being considered for development. The shallowest reservoir (~ 900 m depth) is a highly compacting carbonate gas reservoir under depletion, whereas the intermediate reservoir Shuaiba is an oil-bearing reservoir under water flood. The deeper reservoirs are oil and gas bearing located in the Sudair and Khuff formations.
Interpretation of 3D seismic data shows a major NE/SW and NW/SE fault system in all 3 reservoirs. Depletion in the shallow gas reservoir, which exhibits pore collapsing response on depletion, has induced surface subsidence which is active and expected to reach about 2.4 m at the end of field life. Subsurface deformations and induced stress changes have resulted in subset of the faults (NE/SW) to reactivate, causing seismic tremors, occasionally felt at surface.
Ongoing surface subsidence has resulted in some damage to surface facilities and subsurface well integrity issues. Furthermore, fault reactivation and/or loss of well integrity may induce leakage pathways for reservoir fluids to cross flow between reservoirs or to shallow aquifers. PDO has implemented an extensive monitoring program supported by parallel 3D geomechanical modeling studies, to manage ongoing field development whist mitigating the risks.
Extensive monitoring efforts using a variety of techniques are in place since 1999. Frequent InSAR satellite data measures surface subsidence with such high accuracy and resolution that local zones of higher deformation can be reliably identified and flagged. Continuous GPS data acquisition in a few places throughout the field allows for detailed temporal assessment of subsidence and forms the basis for predictions of total subsidence at end of field life. Periodic in-well compaction monitoring data provides insights in elastic and non-elastic deformation at reservoir layer scale, which is compared against core compressibility data. Continuous microseismic monitoring in a dozen or more observation wells highlights geomechanically active faults in the main reservoir, overburden and underburden, thereby identifying potential risk zones on a near-24/7 basis.
All of this data is used both for well and facilities management, and for providing calibration data for geomechanical models. Results provide clarity on future surface subsidence and differential settlement, which helps to identify facilities with potential risk. The project teams are provided with reliable predictions of surface subsidence throughout the field to ensure the current design tolerance is adequate for integrity of the facilities until the end of field life. This paper presents modeling workflow and calibration with monitoring data related to the geomechanical assessment.
Tuzla is well known for salt mining. Previous research found that the massive exploitation activities of salt mining have caused up to 12 meters of ground subsidence. These results were obtained by an analysis of the time series topographical data from 1956 to 2003. Abandonment of salt exploitation in center of Tuzla was lasting from 2001 and finally finished in 2007. Subsidence causes damage to buildings and infrastructures. Other research using a GPS survey has revealed that subsidence was ongoing from 2004 to 2007. The GPS results showed that the subsidence was decreasing. However, the subsidence is still ongoing in center of Tuzla City, especially in an area near new salt-water lakes for swimming.
In this research, the Differential Interferometry Synthetic Aperture Radar (DInSAR) method was applied to measure the present subsidence in Tuzla. The main purpose of this research is to enhance the ground subsidence information on the spatial distribution and temporal transition. The Small Baseline Subset (SBAS) time series approach is employed. Sentinel-1 data from October 2014 to November 2017 is used to generate the time series of the ground subsidence. The DInSAR results show that the ground subsidence is still ongoing in some areas. The maximum subsidence velocity is about 40 mm/year. This means that continuous subsidence monitoring is very important. In addition, a comparison of the subsidence obtained by DInSAR and GPS is analyzed and discussed in this work.
Ground subsidence has been a major man-made hazard in Tuzla since 1950. The main factor in this ground subsidence is the salt mining activities (Mancini et al., 2009a). Ground subsidence of up to 12 meters was reported during the period of 1956 to 2003 (Mancini et al., 2009a). Research on ground subsidence was continued by means of static relative GPS positioning conducted four times between 2004 and 2007 (Stecchi, 2008). The GPS results from 2004 to 2005 show that the ground subsidence velocity was about 100-200 mm/year. From 2005 to 2006, the ground subsidence decreased to 0-50 mm/year. In a limited area near the salt-water swimming lakes, however, the subsidence velocity was still close to 200 mm/year. The subsidence gradually decreased from 2006 to 2007, except for the area near the salt-water swimming lakes where the subsidence velocity was about 100 mm/year.
Currently, three GNSS stations are installed and in operation to monitor the ground surface movement. The three stations are located in Tuzla (reference point), Tušanj, and Pannonica (Čeliković, 2016). These GNSS stations provide real-time ground subsidence monitoring. However, the subsidence monitoring results of GNSS lack spatial coverage. Spatial coverage is important to understanding the behavior of subsidence itself.
Andersen, Pål Østebø (Dept of Energy Resources, University of Stavanger and The National IOR Centre of Norway, University of Stavanger) | Berawala, Dhruvit Satishchandra (Dept of Energy and Petroleum Technology, University of Stavanger and The National IOR Centre of Norway, University of Stavanger)
Chemically reactive flow is of significant importance for EOR due to possible wettability alteration (low salinity and smart water brines), scaling and chemically enhanced compaction, which all can affect hydrocarbon transportation. In particular, chalks (Ekofisk, Valhall) are highly sensitive to the composition of the injected brine (typically modified seawater) as demonstrated on lab and field scale. We present numerical and analytical solutions to interpret the link between geochemical alterations and creep compaction in chalk cores.
A 1D core scale model is proposed for interpreting geochemical compaction during reactive brine injection into chalk cores loaded uniaxially in creep state (compaction under constant applied effective stresses). An analytical solution is derived to describe the steady state ion and dissolution rate distributions. An analytical model for creep compaction is proposed based on the applied affective stress and the rocks ability to carry that stress as function of porosity. The two models are coupled as follows: The compaction rate is assumed enhanced by the dissolution rate. Further, the solid volume changes by mineral dissolution and precipitation, also affecting the compaction rate. Brine-dependent and non-uniform compaction is therefore built into the model via the dissolution rate distribution.
The model is validated against data from ~ 25 core samples where simple Mg-Ca-Na-Cl brines were injected at Ekofisk reservoir conditions (130 °C), in particular experimentally measured effluent concentrations, distributions in mineralogy after flooding and creep compaction behavior. The model captures the effect of varying key parameters such as brine composition, injection rate and initial porosity and can predict ionic and mineralogical profiles along the core, axial and radial deformation profiles locally and with time. This model is a highly useful tool for interpreting experimental data, predicting in-situ mineralogical distributions where measurements have not been made, and for predicting compaction behavior at changes in brine composition, injection rate or effective stress.
The model is intended for giving a prediction of qualitative and quantitative trends during flooding-compaction tests in chalks. The model and its methodology are translatable to other systems but is validated for lab measurements on chalk samples. Current modeling approaches do not consider the complex interplay between brine and rock compositions, reaction and compaction. This work aims to contribute to the current understanding of this topic.
Sun, Zhuang (The University of Texas at Austin) | Tang, Hewei (Texas A&M University) | Espinoza, D. Nicolas (The University of Texas at Austin) | Balhoff, Matthew T. (The University of Texas at Austin) | Killough, John E. (Texas A&M University)
The reduction of pore pressure caused by depletion can induce significant reservoir compaction, especially in unconsolidated reservoirs. Experiments using unconsolidated core samples are often sparse and costly. We develop a numerical approach based on computer-based simulations of rock samples and mechanical tests. The numerical sample consists of crushable grains simulated with the discrete element method (DEM) and the bonded-particle model (BPM). Model parameters are calibrated through numerical single-grain-crushing tests which reproduce the experimentally-measured sand strength. Grain crushing induced by the uniaxial strain stress path results in a pronounced reduction of porosity and permeability, which manifests more readily for samples with large grain size. The change of particle size distribution indicates that the high effective stress causes grain crushing and produces a significant amount of fines. We perform numerical uniaxial strain tests on numerical samples comprising stiff and soft mineral grains. Simulation results indicate that the presence of soft grains and inclusions (e.g. shale fragments) facilitates the grain crushing. Reservoir simulations, incorporating the change of porosity and permeability as a compaction table, show that the upscaled compaction can enhance production due to compaction drive but also reduces production rate by impairing the reservoir permeability. This multiscale numerical workflow bridges particle-scale compaction behavior and field-scale reservoir production. In this paper, (a) DEM simulations provide a useful tool to investigate compaction effects and complement laboratory experiments; (b) the multi-scale numerical approach can predict the depletion-induced evolution of reservoir production.
This study presents a novel method to estimated pore volume compressibility of shale samples based on mercury injection test data. We revisit our previous study (SPE-185059-PA) for more realistic estimation of pore volume compressibility for shale samples. We present a mathematical model to determine accessible porosity and pore compressibility as a function of pressure using Mercury Injection Capillary Pressure (MICP) data.
During MICP testing in a typical shale sample, the rock sample experiences conformance, compression, and intrusion stages as effective pressure increases. By evaluating compression stage, we calculate bulk compressibility. Further by introducing a system of equations, bulk compressibility is decomposed to estimate accessible pore and grain compressibility separately. Different from our previous model, in this study grain compressibility is calculated based on weight average of mineralogy determined from Fourier-transform infrared spectroscopy (FTIR) experiments. Moreover, bulk compressibility obtained from MICP data is compared with the values calculated from ultrasonic velocity measurements. Samples from both Haynesville shale plays are used to perform our study and validate the hypothesis.
Our results indicate that pore compressibility values are higher than anticipated, where calculated values are in the range of 1E-5 1/psi for shale samples at lower pressure. When pressure reaches to 8000 psi, pore compressibility reduces to the range of 1E-6 1/psi in most of the cases. Moreover, when compared with ultrasonic velocity measurements, results indicate that bulk compressibility obtained from MICP is overestimated at lower pressures and slightly underestimated at higher pressures.
The outcome of the paper changes the industry's take on prediction of the reservoir performance, especially the rock compaction mechanism. This study suggests that production owing to rock compaction can be much greater than what has often regarded, which can change the performance evaluation on a great number of reservoirs in terms of economic feasibility.
ABSTRACT: Hydraulic stimulation and production-induced permeability alteration in stress-sensitive Monterey Formation rocks require an in-depth understanding of the deformation behavior of different rock types. Here, we evaluate and compare deformation behavior of two types of clastic rocks—shale and sandstone—across nano-to-micro scales using experiments and simulations. Using core samples of shale from the Monterey Formation and tight sandstone from the Dominguez Hills in California, we conduct nanoindentation, scanning electron microscopy and particle-based simulation to understand the difference in deformation behavior of shale and sandstone under different loading-unloading conditions. We quantify Young’s modulus, hardness modulus, and stiffness of the rocks using nanoindentation data. Our work provides important insights into grain-scale deformation behavior of two different rock types commonly found in petroleum reservoirs. Understanding of grain-scale failure mechanisms can inform development of new upscaled constitutive models for usage in continuum-scale field simulations, which cannot afford to resolve the grain-scale processes due to computational cost.
Hydraulic fracturing in conjunction with directional drilling has been a game-changing technology for the development of oil and gas resources. It has unlocked vast oil and gas resources in shales and low permeability sandstones which were once considered unfit for commercial production. However, the success of hydraulic fracturing has not been uniform across different types of shales and sandstones. For example, the potential for unlocking millions of barrels of oil from Monterey Formation, California by hydraulic fracturing has been a topic of much debate (EIA, 2011; USC, 2013; EIA, 2014; USGS, 2015). Belridge diatomite, a part of Monterey Formation, has been produced successfully by increasing its low matrix permeability with hydraulic fracturing in 1970-80s. However, the Belridge Field is also well-known for production-induced subsidence and widespread casing failures (Fredrich et al., 1996). Diatomite rock is highly stress-sensitive and can experience pore collapse, reservoir compaction, and induced fracturing due to pressure depletion during production. Our lack of understanding of the geomechanical processes and properties in a geologically complex rock such as Monterey prevents us from designing successful hydraulic stimulation and pressure depletion strategies.
Schutjens, P. M. T. M. (Shell Global Solutions International B.V.) | Fokker, P. (Shell Global Solutions International B.V.) | Rai, B. B. (Shell Global Solutions International B.V.) | Kandpal, J. (Shell Global Solutions International B.V.) | Cid Alfaro, M. V. (Shell Global Solutions International B.V.) | Hummel, N. D. (Shell Global Solutions International B.V.) | Yuan, R. (Shell Global Solutions International B.V.) | Klever, F. (Shell Global Solutions International B.V.) | De Gennaro, S. (Shell Global Solutions International B.V.) | Vaibav, J. (Shell Global Solutions International B.V.) | Bourgeois, F. (Mærsk Oil) | Calvert, M. (Mærsk Oil) | Ditlevsen, F. (Mærsk Oil) | Hendriksen, P. (Mærsk Oil) | Derer, C. (Mærsk Oil) | Richards, G. (Rockfield Software) | Price, J. (Rockfield Software) | Bere, A. (Rockfield Software) | Cain, J. (Rockfield Software)
ABSTRACT: Multi-scale numerical geomechanical models for reservoir and overburden deformation in the Tyra chalk field (Denmark) were made, and calibrated by laboratory deformation tests and field data. The mechanical interaction between the compacting and deforming formation, cement and casing was 1) modeled as a function of well orientation, cement distribution, and mechanical properties, 2) followed by probabilistic analysis of the model results in well-failure risking models to gain insight in the effects of rock deformation on well failure, both in space and time, and then 3) used as input in fluid-flow models to forecast the impact of well-failure on production. The risk analysis revealed that, whilst further Tyra compaction will probably lead to more well failure, its impact on production is probably low. Our geomechanical modeling helped to reduce uncertainty in the high-cost multi-year Tyra Future field upgrade planned for the next years to support Tyra production over the next decades.
1.1 Problem definition
About five meters of maximum subsidence has occurred so far in the Tyra chalk field (Denmark), significantly reducing the gap between wave crest and platform base (see Figures 1a, b). In addition, well deformations in reservoir and overburden have been measured and inferred by caliper data and hold-up-depth incidents (HUD) during logging and work-overs (Figures 1c, d). In the overburden, these have been interpreted mainly from the Upper Lark to Sele-Lista formation, 120 m to 400 m above the top reservoir at about 1950 m TVDss.
With the remaining 35% of the total depletion planned for the next decades, some 8 meters of total subsidence may occur by the end of Tyra field life (about 2042). Also, in view of the increasing reservoir compaction strains, there is concern that wells could catastrophically fail (and thus stop producing, i.e. terminally fail) as a result of the accumulated deformation and/or a possible acceleration of reservoir/overburden deformation. In the framework of the multi-year “Tyra Future” project (in which the Tyra production facilities are adapted to the subsiding sea-bed) field-wide and well-scale finite-element geomechanical models were built to 1) describe the compaction and subsidence, 2) study the mechanical effects of the compacting reservoir and its deforming overburden on cement and casing as function of well inclination, cement distribution, and mechanical properties of formation, cement and casing, and 3) provide input to reservoir fluid-flow simulations to assess the impact of Tyra well failure on production.
ABSTRACT: The City of Red Lodge, Montana, is underlain by up to five levels of room and pillar coal mines, ranging in depth between 55 meters to over 200 meters, that have been closed for nearly 75 years. Potential trough-type subsidence estimates ranged between 75 to 175 millimeters and rates were suspected to be near 3.5 millimeters/year. A leveling network was designed to measure subsidence rates of this order. Pre-survey precision was simulated to gauge a suitable network geometry and level of redundancy. Measured survey elevations were computed with a routine that used weighted least squares (L2-norm) to find the best linear unbiased estimates for subsidence rates. Gross measurement errors were identified with two methods: 1) comparing residuals from the L2-norm with those from the least absolute error method (L1-norm); and 2) examining L2-norm standardized residuals. Several elevation estimates showed small downward movements, many of which had a 95% confidence interval (2-sided) where the entire range was downward. Estimated movements ranged between no movement to 10 mm (downward) with an average of 4 mm (downward). This paper presents a case history of the methods used to evaluate subsidence above historic, abandoned coal mines where surficial evidence is not definitive.
The City of Red Lodge is in south-central Montana, approximately 97 kilometers southwest of Billings. As many as 7 coal seams underlying part of Red Lodge were mined in the early 20th century. The seams vary in depth and condition, with depths below ground surface ranging from 15 meters to over 210 meters. The coal seams are located in sandstone and claystone bedrock, which is overlain by alluvial gravel and boulders.
Underground mining operations started in the 1880s and effectively ceased in the early 1930s. The resulting underground workings lie beneath an area of nearly 8.5 square kilometers. Of that, approximately 0.8 square kilometers lie beneath the city limits, much of which is residential. Figure 1 shows a plan view of the known mine workings in the area immediate to the city. According to historic documents, until around 1908, the method of working the seams was room and pillar, where pillars were left in place. The distance of rooms from center to center was 21 meters and the general distance between levels was 168 meters long (Rowe, 1908). This geometry yields a 41% extraction ratio. The segment of mine map shown in Figure 2 shows typical room and pillar geometries used for extraction in Red Lodge.