Production from oil fields requires monitoring of hydrocarbon saturation in the reservoir. In the Bockstedt oilfield there exists a substantial difference in resistivity between oil-filled (approx. 100-16 Ohmm) and brine-filled (0.6 Ohmm) reservoir. Electromagnetic method is chosen to test whether sufficient resistivity differences can be observed via surface measurements. The target is a Lower Cretaceous clastic interval located at an approximate depth of 1200 metres.
Forward modelling demonstrates that the expected resistivity changes at reservoir level cannot be resolved with a survey setup of only surface electrical sources and sensors. Therefore a borehole-to-surface technique has been developed, whereby the metal casing of an abandoned production well serves as input electrode. CSEM surveys were acquired in 2014 and 2015 as timelapse baseline and monitor for both Ex and Ey components. Forward modelling indicates that induction effects from metal objects like casings of production wells cannot be ignored in the EM modelling. A shallow observation well was drilled in 2015 to make collection of Ez datasets possible. A new downhole sensor was developed for this purpose. Numerical simulations suggests Ez data is more sensitive to the anticipated resistivity changes. Since Ez is two orders of magnitude smaller than the horizontal components, verticality is of great importance to avoid masking the Ez signal by interference from unwanted horizontal components.
Similar acquisition parameters are adopted for 2014 baseline and 2015 monitor surveys to facilitate the comparisons. The repeatability is good, generating comparable response functions. The earth model, retained so far by the inversion algorithm, confirms the main resistivity units seen by the resistivity logs in the calibration well. Incorporation of the metal casings in the EM modelling scheme increases the lateral continuity of inverted resisitivity bodies.
This study is a follow-up of work presented in the 2015 ADIPEC conference. In November 2016 a new acquisition campaign will be undertaken to collect a second Ex – Ey monitor and the first monitor survey for the Ez component. A limited time-lapse test has been performed in spring 2016 to monitor Ez with only 1 source station incorporating the borehole electrode Bo-23.
How to establish a fine reservoir model is always the reservoir research's core issue on which reservoir engineers focus. The well-constrained geologic modeling method is widely applied in reservoir modeling. For an offshore oilfield, the reservoir model established with the well-constrained geologic modeling method has very high uncertainty because of skewed distribution of drilled wells. In this case, the use of seismic data of high transverse resolution is particularly important. The geostatistical inversion method fully combines the drilling data of high Vertical resolution with the seismic data of high transverse resolution and can be used to establish a high precision reservoir model taking geostatistics parameters as control and the inversion technology as the core, thus reducing the uncertainty of the reservoir model.
Presentation Date: Monday, October 17, 2016
Start Time: 4:10:00 PM
Presentation Type: ORAL
The Southwest Partnership on Carbon Sequestration (SWP) is one of seven large-scale demonstration projects sponsored by the U.S. Department of Energy. The SWP has a goal of permanently sequestering more than 1,000,000 metric tonnes of CO2 in an active EOR project in a mature waterflood in the Anadarko basin. The CO2 for this project is anthropogenically sourced from a fertilizer and an ethanol plant. As of the end of 2015 the field has 13 CO2 injectors and has sequestered 386,695 metric tonnes of CO2 between October 2013 and July of 2015. Goals of the project include optimizing the EOR/storage balance, ensuring storage permanence, and developing best practices for carbon storage utilizing man-made CO2.
The field site provides an excellent laboratory for testing a range of monitoring technologies in an operating CO2 flood since field development is sequential and allows for multiple opportunities to record zero CO2 baseline data, midflood data, and data from fully flooded patterns. The project has acquired data at a number of scales including a 42 mi2 3D seismic survey, baseline and repeat 3D VSP surveys centered on three injection wells, cross-well tomography baseline and repeat surveys between injector/producer pairs, a borehole passive seismic array to monitor for induced seismicity, a distributed temperature system, and bottomhole pressure and temperature sensors. The project has drilled three wells in the field, and has acquired over 750 ft of core in the Morrow B reservoir interval and associated caprock units. Additional monitoring focuses on CO2 soil flux, groundwater chemistry, reservoir fluids chemistry, and aqueous and gas-phase tracer studies.
All acquired data have contributed to detailed geologic models used for fluid flow and risk assessment simulations. 3D VSP and cross-well data with repeat surveys have allowed for direct comparisons of the reservoir prior to CO2 injection and at eight months into injection, with a goal of imaging the CO2 as it moves away from injection wells. Additional repeat surveys at regular intervals will continue to refine direct CO2 imaging as production and injection data are integrated with newly acquired and interpreted data. All models are regularly updated. In this paper the project goals will be outlined, progress towards goals enumerated, and current geologic and simulation models will be introduced. In addition, initial results from time-lapse monitoring of movement of CO2 in the reservoir will be discussed.
In this study, we use an example in a Barnett Shale play to demonstrate how supervised and unsupervised machine learning techniques provide the right leverages for seismic interpreters. By analyzing seismic facies map generated by unsupervised self-organizing map, gamma ray estimated by artificial neural network, and brittleness index estimated by supervised proximal support vector machine, we arrive at frackability and lithofacies interpretation of the Lower Barnett Shale. We find strong agreement between interpreted frackability in the Lower Barnett Shale with microseismic events.
Frackability, which can be measured by the brittleness index (BI) of reservoir rocks, is a key parameter of recovering unconventional shale reservoirs. In this study we use both unsupervised learning technique (self-organizing map or SOM) and supervised learning technique (proximal support vector machine or PSVM) trying to estimate BI using five petrophysical attributes. From unsupervised SOM, we cannot directly get BI from the inputs, but can have clustered lithofacies, which need to be further calibrated with other petrophysical and engineering data. From supervised PSVM, we can directly calculate BI on seismic data based on the relation obtained from a training well. Because gamma ray is a good indicator of clay minerals as well as total organic carbon (TOC) which generally make a rock ductile, we also compare the estimated BI with a gamma ray volume estimated using artificial neural network (ANN). Good virtual correlations are identified among SOM facies, BI and gamma ray volumes. The estimated BI volume is further validated by microseismic data.
The target formation in this study is the Lower Barnett Shale (Figure 1), which was deposited in the Mississippian period and dominated by silica-rich mudstones. The shale formation is bonded by Forestburg Limestone and Viola Limestone, which are considered as fracture barriers when doing hydraulic fracturing in the shale formations.
Perez and Marfurt (2014) present a workflow to estimate BI from crossplotting Lambda-Rho and Mu-Rho, which can be derived from seismic prestack simultaneous inversion. Though easy to implement, such crossplot may not sufficiently recover the highly nonlinear relation between mineralogy derived BI and seismic derived elastic properties. This motivates the authors to deploy nonlinear machine learning techniques which provides the ability to analyze data in a higher dimensional space (i.e. analyzing multiple types of data simultaneously), to discover the relation between rock frackability and seismic measurements.
Quantifying localized deformation in the target reservoir formations is of importance for the drilling and production of hydrocarbons. Our recent efforts have been primarily focused on generating seismic geometric attributes (discontinuity, curvature, and flexure) and applying them for qualitative description of fractured reservoirs. This study presents a new method for quantitative strain analysis based on reflection geometry from 3D seismic data, and the generated tensor could help quantify both normal strain and shear strain of reservoir formations. We apply the method to one of the major fractured reservoirs at Teapot Dome (Wyoming) that is known to be caused by bending and shearing of the reservoir formation. The results not only help differentiate shear deformation from contractional and extensional ones, but also demonstrate a good correlation between producing wells and high-strain zones. The example indicates the potential of the technology for more robust and quantitative characterization of fractured reservoirs.
Characterizing subsurface structural deformation from three-dimensional (3D) seismic data is one of the most significant tasks in hydrocarbon exploration and production. Various seismic discontinuity attributes have been well developed and widely used for detecting faults/fractures in the subsurface (e.g., Bahorich and Farmer, 1995; Luo et al., 1996; Marfurt et al., 1998; Di and Gao, 2014a); however, the discontinuity attributes are basically qualitative and cannot differentiate shear deformation from extensional and contractional ones, which is a major limitation for evaluating fluid migration potential, leakage risk, and reservoir mechanical properties vital to successful hydrocarbon exploration and production. Many authors proposed using curvature and flexure to predict deformational intensity and to differentiating fracture mode (e.g., Lisle, 1994;; Roberts, 2001; Al- Dossary and Marfurt, 2006; Gao, 2013; Di and Gao, 2014b, 2014c, 2014d; Gao and Di, 2015); however, no quantitative relationship has been developed between geometric attributes and finite strain. Starr (2014) proposed using a modified curvature analysis for quantifying deformation intensity with the Marcellus shale; however, the method is a two-dimensional averaging method that provides little information for instantaneous strain as a 3D tensor.
Bahman, Hussasin Ali (Kuwait Oil Company) | Hajeyya, Abdullah Khalid (Kuwait Oil Company) | Al-Zankawi, Omran (Kuwait Oil Company) | Mukherjee, Pradip Kumar (Kuwait Oil Company) | Al-Sabea, Salem Hamad (Kuwait Oil Company) | Mohammed Ali, Farida (Kuwait Oil Company)
Geo-steering is a very critical part of today's field development economics, our production targets are getting more complex, thinner oil columns, which need more complex geo-steering, continual improvement needed in People, technology and processes. Drilling a well at an angle other than vertical can obtain more information by hitting the production targets and stimulate reservoirs in ways that cannot be achieved with a simple vertical well which became a valuable ability in oil business. To augment this aspect Kuwait Oil Company has established Geo-steering Center (
The establishment of Geo-steering control Room in FD (S&EK) is an outcome because of constant supervision and direct guidance by manager of Field Development South and East Kuwait, which added a new dimension to drilling the modern horizontal wells in the Greater Burgan Field. The team of Geologists of FD (S&EK) in this collaboration center ensures that horizontal wells are steered correctly and safely to their final targets.
The Geo-steering center can be operated 24 hours a day if require. Each geologist may be responsible for as many as 3 wells in different fields (BG, MG & AH) and different reservoirs. Like driving, geo-steering requires constant attention and dedication all the time. The center recently moved into a new and expanded facility that is equipped with the latest in visualization, communication and computer technology in order to properly place and geologically navigate us with many complex horizontal wells path in Greater Burgan field. Geo-steering horizontal wells can be done remotely from the center, with data coming into the center from more than one well at any given time. For every well, Logging-While-Drilling (LWD) sensors near the drill bit send information about the Lithology and directional survey of the well to the control unit at the rig from where data is then transmitted by satellite to the geo-steering center. The team developed software instantly can load the data so geologists can see on their workstations the LWD and trajectory data to determine where the drill bit is in relation to the drilling plan and the reservoir target.
Massey, C. (GNS Science) | Richards, L. (Rock Engineering Consultant) | Pasqua, F. Della (GNS Science) | McSaveney, M. (GNS Science) | Holden, C. (GNS Science) | Kaiser, A. (GNS Science) | Archibald, G. (GNS Science) | Wartman, J. (University of Washington) | Yetton, M. (Geotech Consulting (NZ) Ltd.)
The 2010/11 Canterbury earthquakes triggered many mass movements in the Port Hills including rockfalls, rock and debris avalanches, slides and slumps and associated cliff-top cracking. The most abundant mass movements with the highest risk to people and buildings were rockfalls and rock/debris avalanches. Over 100 residential homes were impacted by landslides, leading to the evacuation of several hundred residents.
Volumes of rock leaving several of the larger cliffs during the earthquake sequence were determined from terrestrial laser scan change models. There were no seismically instrumented cliff sites and there was some distance between the cliffs and nearest strong-motion sites. Therefore, we synthesised free-field rock-outcrop seismograms by employing a stochastic approach controlled by source models and regional parameters derived using spectral inversion of the extensive strong motion data set.
Relationships between volumes leaving cliffs during the earthquakes and site peak ground acceleration (PGA), peak ground velocity (PGV) and Arias intensity were compared for different sites. Multiple linear regression was used to analyse the variables that best predict the volumes of debris that fell from the slopes during the main earthquakes. The best correlation between the volume of debris falling per square metre of slope face and the seismic forcing parameters was for vertical PGV. The results from the multiple linear regression incorporating slope height, inclination, PGV horizontal and PGV vertical, improved the statistical relationship.
Field data and results from the 2D seismic site response assessment indicate that the following factors affect dynamic performance of the modelled cliffs: 1) cliff geology—mainly material modulus, shear strength and shear wave velocity; 2) slope geometry—ridge-scale versus site-scale effects; and 3) the temporal aspects of the earthquake shaking (i.e., single acceleration peaks of large amplitude versus multiple peaks of smaller amplitude). Model results show that amplification of shaking does not increase linearly with increasing height, but instead reflects changes in the cliff geology where material strength, modulus and shear-wave velocity contrasts lead to acceleration contrasts.
These factors show that the use of well documented case histories provide the basis for more certainty in seismic landslide assessments compared to those that are only empirically based.
Geophysical inversions play an important role in today’s mining exploration. Accompanying the availability of multiple types of geophysical data collected over the same study area is the need of jointly inverting them all and generating a common earth model, as opposed to the common practice of separately and independently inverting each data set. We investigate in this paper the use of multi-domain joint clustering inversion of magnetic and IP data in the context of sulfide deposit exploration. We compare the recovered models from joint inversion with those from L2-norm inversion, L1-norm inversion and separate clustering inversion of single data type, and demonstrate that only the jointly inverted models can reliably represent the complex geological structures in such a setting.
Xia, Qinglong (China National Offshore Oil Corporation Limited) | Tian, Lixin (China National Offshore Oil Corporation Limited) | Zhou, Donghong (China National Offshore Oil Corporation Limited) | Xiong, Xiaojun (Chengdu University of Technology)
Seismic interval velocity analysis is the most useful tool for predrill geopressure prediction. In this paper, we have combined two kinds of seismic interval velocity computational methods with geopressure prediction. First is a method of 3-D Dix constraint inversion, which can effectively overcome the defects of traditional Dix’s interval velocity equation, obtain high-precision interval velocity section with clear reflection of structural details, and effectively identify the anomalous region of interval velocity. Second is a method of wave impedance inversion, which can obtain high-precision seismic interval velocity of target formation with high detail resolution. Then, we combine two kinds of seismic interval velocities by the empirical approach proposed by Fillippone (Fillippone equation) to predict geopressure and obtain accurate results of geopressure prediction in the area of Bohai Bay, China.
It is well known that rough seas cause higher levels of noise in marine seismic data, and that the noise level is higher for shallower streamer tows. It is also understood, although less well known, that the roughness of the sea surface induces time and space dependent variations in both receiver-side and source-side ghost reflections. Since broadband processing aims at the removal of ghost effects, it is important to assess the impact of these factors on broadband data quality. This paper reviews previous work related to these subjects and discusses two adjacent 3D seismic lines acquired in the central North Sea (Quad. 29/30), in calm and rough conditions respectively. The objective of this analysis is to draw qualitative and quantitative conclusions on broadband seismic data quality in rough and calm conditions.