Grover, Anurag (Abu Dhabi Company for Onshore Petroleum Operations Ltd.) | Al Mesmari, Abrar (Abu Dhabi Company for Onshore Petroleum Operations Ltd.) | Al Shamsi, Saif (Abu Dhabi Company for Onshore Petroleum Operations Ltd.) | Al Shabibi, Tariq (Abu Dhabi Company for Onshore Petroleum Operations Ltd.)
An uncertainty analysis involves the study of different scenarios with multiple realizations of those scenarios. Structural uncertainty analysis deals with the estimation of uncertainty only due to structural components. In any model, there are a tremendous amount of uncertain parameters. It is important to establish key parameters to be investigated for the better understanding of the model. Most important parameters influencing structural uncertainty using 3D seismic data are Horizon Picking, Velocity Modeling, Well Correlation / Marker Picking & Seismic Data Processing.
The motivation for the present study was the limited understanding of structure, especially, in the flanks of the field. Prognoses made in this area continue to both overestimate and underestimate formation depth and thickness. Given the average depth of the reservoir formation, several meters can have a significant impact on saturation and therefore productivity and reserves estimates. Additionally, this area has been marked as a viable location for future development projects. Capturing structural uncertainty ranges helped to understand the risks involved during well placement and provided development team, enough time, to investigate alternative locations. Structural uncertainty study using 3D seismic data was attempted to capture the GRV ranges in the reservoir with the ultimate aim to build factual static & dynamic models for optimum field development.
Time-Depth (TD) conversion often accounts for more than 50% of the GRV uncertainty. This case study presents systematic methodology applied for capturing structural uncertainties appearing due to variability in horizon picking & velocity modeling. These were two key parameters used for TD conversion. The first input to the study was the sets of alternative horizon interpretations predominantly done on seismic data. Inputs from various seismic attributes were used to guide the interpretations in the areas of poor seismic data quality. The other input was the sets of alternative velocity models for the field. These models were generated using the combination of well velocities (VSP/Checkshots/Sonic) and Pre-Stack Depth Migration velocity cube. A layer cake modelling approach was used for velocity modelling. Multiple depth surfaces have been generated using different sets of horizon interpretations and velocity models. Generated surfaces were analyzed with well markers at well locations.
A statistical approach has been taken to estimate structural uncertainty as 1st standard deviation (SD) depth errors in the reservoir. Multiple realizations on the 3D structural model were completed using Sequential Gaussian Simulation (SGS) for uncertainty calculations. SGS was used to generate an error surface. This error surface was applied on Base Case structural model to perform ‘n’ stochastic realizations in order to generate ‘n’ depth scenarios. Multiple depth scenarios were later used to calculate GRV ranges in the reservoir.
To overcome the lack of information on the most superficial part of the near surface obtained by the use of the First Arrivals (refracted waves), we have implemented an innovative combined workflow that use the information from the Surface Waves (Rayleigh waves) to complement the first-break measurements. As the Rayleigh waves propagates along the free surface interface, they carry significantly more detailed information of the near-surface characteristics which can be used to better constrain the first-break inversion.
The first step of the workflow starts with the surface wave dispersion curve picking. As reliability of the results directly depends on the quality of that picking, data regularization are used to improve picks accuracy on both low and high frequency of the phase velocity/frequency spectra. A surface wave tomography process is then applied to convert the spatially irregular frequency-dependent picks into a regularized (x, y, frequency) Rayleigh wave’s velocity volume. Lastly a laterally constrained depth inversion is performed, delivering a 3D shear wave’s near-surface depth velocity model.
The second step of this workflow uses this S-wave velocity model, which contains the near-surface details captured by the Surface-Waves, to constrain the refracted P-wave first-break tomography. A regional scaling, here a 1D VP/VS ratio estimated from knowledge over the area or from fast-track refraction first-break analysis, is required to convert the S-wave velocity into a P-wave model. The constrained tomography aims to scale the trend from the high-resolution S-waves velocity model by fitting it with the trend of the P-wave field derived from the first arrivals. This corresponds in a sense in inverting the VP/VS ratio in such a way to preserve the high-resolution feature scaptured by the surface waves while remaining consistent with P-wave information. The resulting near-surface velocity model looks more geologic, better respects P-wave travel times and can be used with more confidence to compute the primary statics solution than the conventional P-wave field only obtained from the first-arrivals tomography…
Furthermore, this accurate update of VP/VS ratio can be used to estimate the Poison’s ratio. It can be used to better plan geotechnical survey in order to reduce shallow drilling hazards.
Understanding the created fracture geometry is key to the effectiveness of any stimulation program, as fracture surface area directly impacts production performance. Microseismic monitoring of hydraulic stimulations can provide in real-time extensive diagnostic information on fracture development and geometry. Thus, it can help with the immediate needs of optimizing the stimulation program for production performance and long-term concerns associated to field development. However, microseismic monitoring is often underutilized at the expense of productivity in the exploration and appraisal phases of a field.
Geology is a fundamental element in the design of a stimulation program and the interpretation of its results. Rock properties and geomechanics govern the achievable fracture geometry and influence the type of fluids to be injected in the formation and the pumping schedule. Rock layering controls the location of the monitoring device, guides the depth at which perforations should be located, and influences how hydrocarbons flow within the formation. Despite this importance, the impact geology may have on the stimulation results is often overlooked as it is all too common to see assumed laterally homogeneous formations, invariant stress field (both laterally and vertically), stimulated fractures having a symmetric planar geometry, etc.
As exploration and appraisal moves toward active tectonics areas (as opposed to relatively quiet passive margins and depositional basins), understanding the impact of complex geology and the stress field on fracture geometry is critical to optimizing stimulation treatments and establishing robust field development plans. Mapping of hypocenters detected using microseismic monitoring is an ideal tool to help understand near- and far-field fracture geometry. Additionally, moment tensor inversion performed on mapped hypocenters can contribute to understanding the rock failure mechanisms and help with evaluating asymmetric and complex fracture geometry. Understanding this fracture complexity helps address key uncertainties such as achievable fracture coverage of the reservoir.
We present the results of several hydraulic fracture stimulations in various geological environments that have been monitored using microseismic data. We illustrate with these case studies that in some rare cases, simple radial and planar fracture system (often mislabeled penny shape-like fracture) may be generated as predicted using simple modeling techniques. However, in most cases, the final fracture system geometry is complex and asymmetric, largely governed by stress, geologic discontinuities, rock fabric, etc. Understanding this impact and optimizing the well design to enhance productivity is key to evaluating reservoir potential and commercial viability during exploration and appraisal phases and for maximizing return on investment during development.
We introduce a generalized concept of the so-called blending and deblending, and establish the generalized-blending and -deblending models. Accordingly, we establish a method of deblending, or deblended-data reconstruction, using these models.
The generalized blending can handle real-life situations; this includes random encoding both in the space and time domain, both at the source and receiver side, thus all incoherent shooting, inhomogeneous shooting, non-uniform and under sampling. Similarly, the generalized deblending includes data reconstruction that works for all shot-generated-wavefields separation, spectrum recovery and balancing, regularization and interpolation, again both at the source and receiver side. However, we do face a challenging question: how to fully reconstruct deblended data from the fully generalized blended data. To address this challenge, we consider an iterative optimization scheme using a so-called closed-loop approach. We use the properties of blended signal specified by the blending code: the coherency of blended signal versus the incoherency of blending noise in the pseudo-deblended domain. This can be posed as an inverse problem with quantifying the coherency and its solutions by selecting optimal metrics of the coherency. We applied this method to synthetic datasets. The results show that our method succeeded to fully reconstruct deblended data from the fully generalized blended data.
We discuss its applicability to time-lapse seismic monitoring as it ensures high repeatability of the surveys. Our methodology should reduce the repeatability problem because reconstructing deblended data in monitor surveys is much more realistic and reliable than positioning sources and receivers exactly as the baseline survey.
OMV and ADNOC signed a study agreement in 2013 to explore for hydrocarbons in a large (10,000km2) under-explored onshore area, named East Abu Dhabi. The objective of the work programme was to evaluate the conventional and unconventional hydrocarbon potential within multiple play types and structural settings, via the analysis of existing vintage data, acquisition of new seismic followed by exploration drilling.
To date 1,800km2 3D (4S) and 700km 2D seismic have been acquired focused on two principal play types; namely, the ‘Pabdeh’ stratigraphic play and the ‘Thamama’ combined structural/stratigraphic play. Additional studies completed include fluid inclusion stratigraphy using data from nearby vintage wells, and the completion of an unconventional study covering the wider area of interest. The first OMV operated exploration well reached its TD in the Jurassic in March 2017. Two tests have been performed in Lower Cretaceous and Jurassic resulting in a dry sour gas discovery.
The main results of the well that have an impact on the understanding of the regional geology can be summarized as follows: 1) Source Rock, three potential source rock intervals have been penetrated (Middle and Lower Cretaceous & Jurassic). 2) Reservoir, The middle Cretaceous has been found in a back-shoal facies with its suggested corresponding platform margin being located in close proximity to the South-West. The Aptian is represented by the classical Lower Shuaiba fm. and overlain by the Bab shales. No isolated platform has been encountered. 3) Clear stratigraphic and structural evidence supporting structural deformation of the Thamama Group during the Lower Cretaceous. Several distinct fault trends are evidenced from both the well data and 3D seismic depth slices. Understanding these faults and related fracture systems will be fundamental in understanding the play potential in the wider area.
This is the first exploration well to be drilled in the area since the ‘80s. Multiple intervals of regional interest have been encountered spanning the massive loss circulation intervals of the Palaeocene, conventional and unconventional reservoir within the Middle Cretaceous, the entire Lower Cretaceous sequence and the Asab equivalents of the Upper Jurassic.
In 2012, the PDO Geophysical Operations team started to acquire the largest 3D seismic survey to be acquired in Oman, with a total area of 8000km2. Here we will present geophysical workflow for that survey, including 3,000 km2 of the Umm As Samim sabkha. Before the acquisition of this survey, the area had limited 3D seismic coverage due to its highly complex and logistically difficult environment. To ensure the survey would be acquired safely; extensive scouting was put in place, bulldozers were used to construct graded tracks across the buckled sabkha crust outside which the workforce was not allowed to walk. Furthermore there was a continuous measurement of the water level below the sabkha and there were contingency plans in place in case the sabkha got flooded.
The 8000km2 area was split into 4 blocks overlapping with an acquisition geometry of 25x250m receiver grid and 25x100m shot grid. The acquisition recorded up to 30km inline offset and 6km cross-line offset and sweeping from 1.5 to 86 Hz. To expedite the seismic for interpretation, the processing started a couple of weeks after the first shot, and it was done in 4 phases. The first phase consisted of signal processing and pre-stack time migration of the first two blocks. This was followed up by pre-stack depth migration of the first two blocks in the second phase. In phase 3, the phase 1 processing was reperated for blocks 3 and 4 and finally in phase 4 all blocks were merged and depth migrated to give seamless volume. The processing of the data was executed in phases to speed up the evaluation of the data for exploration purposes. The integrated team has allowed an optimum processing of the data. The survey took 29 months to acquire with a detailed pre acquisition planning to mitigate potential HSSE risks in the area such as Umm As Samim sabkha and hydrocarbon field facilities. In order to minimize the acquisition gaps in the wet sabkha, light vibrators were used. Due to the massive size of the survey, data quality over the area was different and this led to a change in the workflow from one block to the other. This has lead to re-work to ensure a seamless merge could be accomplished over the whole area. Superior data quality compared to legacy data led to a number of wells being planned or changed position due to the improved imaging. Phase three and four led to an increased confidence on evaluation of prospects (both volumetric and risking) and reducing HSSE risk. As part of the deliverables, acoustic impedance volumes have been generated; these volumes will help the interpreters to more accurately pick events as well as having a first view on the reservoir quality of identified prospects. This project is a result of close cooperation and synergy in delivery between Exploration and Production.
Shekhar, Ravi (Zakum Development Company) | Almazrouei, Khulood A. (Zakum Development Company) | Wendland, Corey (Zakum Development Company) | Obeta, Chukwudi (Zakum Development Company) | Al Zinati, Osama (Zakum Development Company) | Omura, Taihei (Zakum Development Company) | Bengharbia, Mourad (Zakum Development Company) | Al Neyadi, Abdulla (Zakum Development Company) | Khouri, Naeema (Zakum Development Company) | Al Shehhi, Budoor Hasan (Zakum Development Company)
The challenges to effectively manage a field with stacked reservoirs that are in known hydraulic communication are enormous. Conventional management of such fields can result in many undesired outcomes such as premature water production caused by cross-communication from different reservoirs, inefficient pressure support to some of the individual reservoirs, or large volumes of bypassed oil. The major challenge is to identify the reasons and magnitude of inter-reservoir communication, and accurately capture these in the models which are utilized for optimized well placement and efficient field development.
This paper discusses an integrated multi-disciplinary approach to identify and model inter-reservoir communication in a giant offshore carbonate field from UAE. Tracer study conducted for several years combined with detailed seismic interpretation suggest that the cross-flow of reservoirs due to fault juxtapositions is the major cause of inter-reservoir communication. This observation is further supported by pressure-transient analysis, logs, cores, drilling reports and regional structural studies. Hence, it becomes absolutely essential to build a robust static and dynamic model that can accurately capture the communication between different reservoirs. This paper proposes a novel approach to build a full-field integrated framework that allows coupling of multiple reservoirs that have been previously modeled independently. The methodology includes; detailed reinterpretation of faults and key chronostratigraphic surfaces, qualitative/ quantitative attribute analysis using reprocessed 3D seismic (post-stack time migrated) data, interpretation of pressure transit analysis, logs, surveillance data and regional structural history studies. The updated framework ensures accurate fault throw and fault extension in order to capture fault juxtapositions.
The ability of the new model to allow inter-reservoir communication has been tested and confirmed in dynamic simulation model. This was achieved through series of simulation sensitivities where tracers injected in wells targeting specific reservoirs were successfully sampled from different reservoirs due to inter-reservoir communication through fault juxtaposition. Based on the results of sensitivity test, it is expected that the new integrated framework will provide a much improved history match in faulted areas where cross-communication across reservoir is very prominent. The improved model will lead to a better understanding of field and possibly will be used as guidance for field development plan.
The Wasia Formation presents opportunities to explore for stratigraphic traps in the Saudi Arabian Rub’ Al-Khali Basin because it contains numerous interbedded reservoirs, sources, and sealing rocks. The mid-Cretaceous Wasia Formation includes a rudist carbonate platform with five, third-order sequences comprising, from oldest to youngest, Safaniya, Mauddud, Ahmadi, Rumaila, and Mishrif members. These members include proximal shallow-marine, highstand carbonate shoals at the platform margin in close proximity to fine-grained carbonate deposits in the Shilaf Basin. The resulting depositional cycles and stratigraphic architecture position muddy-tight seals, adjacent to porous shallow-marine carbonate-shoal bodies. Two members (Safaniya and lowermost Mishrif) have high organic-matter content situated in the oil window.
Core data, well logs, seismic signals, and modern analogs were analyzed to help understand the Wasia deposition. Detailed correlations were made of well logs and neural network training was used to generate electro-facies. Next, supervised waveform analysis was used, correlated to five well log facies, to create five waveform facies including (1) lagoon, (2) back-shoal, (3) shoal, (4) slope, and, (5) basin facies. Sources of potential uncertainties include data processing, seismic to well ties, position of stratigraphic tops and seismic horizon interpretation. To minimize these, care was taken in data processing and a blind test was performed to validate the final interpretation.
On the basis of integrating the aforementioned data with our waveform facies, a reference geological model was built demonstrating that potential stratigraphic traps are porous, shallow-marine carbonate shoals intercalated with muddy-tight slope deposits resulting in isolated, porous carbonate reservoir bodies sealed by tight rocks. For example, the Ahmadi Memberseal was deemed to be too thin to seal the oil in the underlying Mauddud. In addition, muddy-tight lowermost Mishrif Member strata are also too thin to seal oil in the underlying Rumaila. In the worst case, laterally extensive upper Mishrif reservoirs are not sealed by interbedded lateral seals even though the Aruma shale seals their tops. The two best trap configurations include (1) the first highstand lower sequence of the Mishrif reservoir sealed by interbedded extensive transgressive muddy Mishrif carbonates and (2) thick Ahmadi and lowermost Mishrif fine-grained carbonates sealing Mauddud and Rumaila highstand system tracts.
Most fracture characterization techniques using seismic data are based on seismic anisotropy analysis. Providing a more reliable dataset for such analysis should be considered during the entire seismic processing sequence. This case study demonstrates that the new techniques in seismic regularization of high dense, single-source, single-receiver (S4) seismic data along discrete azimuth directions with regular source-detector distance intervals (radial domain interpolation) can provide enhanced imaging through azimuthal velocity analysis, and deliver datasets for seismic inversion with improved noise attenuation. A comprehensive workflow is presented to improve the final image quality, and to preserve the azimuthal amplitude variation with offset and azimuth (AVOAz) for its analysis, which in turn can lead to the derivation of intrinsic rock property attributes for better reservoir management decisions & drilling plans. To compensate for the overburden effect and improves resolution at the zone of interest, a multi azimuth prestack depth migration approach resolved most of the effects of heterogeneities in the velocity field, which can be misinterpreted as azimuthal anisotropy.
The processing sequence for this dense seismic (acquired with 5×5 meter bin size, 1600 traces per bin), broadband survey in Abu Dhabi supports the understanding that azimuth-rich single-source, single-receiver acquisition geometries can provide advantages over conventional techniques (such as the use of receiver arrays or narrow-azimuth surveys) for the imaging, fracture characterization, and reservoir management of very subtle low-relief structures, in particular shedding light on the complex geological environments of carbonate reservoirs in the Middle East.
Nakatsukasa, M. (Japan Oil, Gas and Metals National Corporation) | Ban, H. (Japan Oil, Gas and Metals National Corporation) | Kato, A. (Japan Oil, Gas and Metals National Corporation) | Worth, K. (Petroleum Technical Research Centre) | White, D. J. (Geological Survey of Canada)
Time-lapse seismic is often used for reservoir monitoring. Although some case studies have been reported, especially for offshore oil fields, similar onshore Abu Dhabi studies have been limited. Onshore monitoring challenges are mainly due to hard carbonate reservoir rock, in which fluid displacement invokes only minor changes in rock properties. Moreover, complex near-surface structures create strong surface-related noises such as surface waves and elastic scattering. These noises significantly degrade the repeatability if repeat data are recorded with positioning error of 1 meter. To address the problems above, permanent reservoir monitoring has been spotlighted recently. Although permanent receivers are widely used for seismology, few studies of permanent seismic sources have been reported. In this study, we demonstrate a permanent source, Accurately Controlled Rotational Operated Signal System (ACROSS), which has a fixed position and excites highly repeatable seismic waves by rotating an accurately controlled eccentric mass. We deployed ACROSS at the Aquistore CCS test field onshore in Canada and recorded the wavefield with permanent receivers buried underground. The ACROSS system was operated several times throughout one year. The data acquired in a continuous 45 day operation were analyzed to calculate variation of amplitudes. The result indicated that we had obtained very stable waveforms with less than 3% variation if the data is stacked sufficiently. We also compared the two datasets acquired at different times of the year by calculating the normalized root mean square (NRMS) as a repeatability index. The histogram showed that the peak of NRMS distribution approached ~15% with only front and back mute processing. These results support our conclusion that an ACROSS system has potential to be used for reservoir monitoring in onshore fields by providing highly repeatable data for an extended time period.