- Home
- Conferences
- Conference Proceedings
- Conferences
82nd EAGE Annual Conference & Exhibition - Workshop Programme
- Conference date: October 18-21, 2021
- Location: Amsterdam, Netherlands / Online
- Published: 18 October 2021
1 - 20 of 45 results
-
-
Geological Uncertainty Quantification: A Futile Attempt in the Absence of a Decision Context
More LessSummaryIn most extractive industries, understanding geological uncertainty is key to making good decisions. In addition, the decisions should serve our goal of extracting economical amounts of resources. With this description, building the most realistic geological model should not be the goal. Instead, our goal should be to build models that support good extraction decisions. Yet, we rarely discuss the link between uncertainty models and their usefulness. We devote substantial efforts to gathering information and building sophisticated subsurface models but significantly less to exploring their economics and usefulness. If, for example, we make the same petroleum drilling decision irrespective of the granularity of the uncertainty model, then the time and effort spent on developing such sophisticated models have been in vain.
-
-
-
Acoustic versus Elastic 3-D FWI: a Case Study at the East Pacific Rise 9ºN
Authors M. Marjanovic, R. Plessix and A. StopinSummaryFull Waveform Inversion (FWI) has become a standard tool for imaging subsurface. Although the acoustic formulation of the wave equation in FWI has been commonly used, excluding the elastic effect could have a significant impact on the inversion results. To quantitatively evaluate the contribution of the elastic approach, we compare acoustic and elastic 3-D FWI applied to a 3-D seismic data set from the East Pacific Rise (EPR) 9°50’ N, collected in deep-marine setting. After conducting a number of tests, we suggest a simultaneous, multi-parameter inversion using frequencies below 7 Hz for both acoustic and elastic approach. The resulting residual from the elastic case is 10–15% lower than that for the acoustic case, suggesting that the elastic approach explains the observed data better. Furthermore, the final compressional velocity model of the subsurface using the two approaches differ significantly, not only in velocity magnitude but also suggest different geological interpretation. We argue that the results obtained from the elastic modelling are geologically more plausible and more reliable image of the subsurface at the EPR.
-
-
-
Incorporating Probabilistic Petrophysical Information into Elastic Full Waveform Inversion
More LessSummaryElastic full waveform inversion (EFWI) augmented with petrophysical information defines a high standard for velocity model building, as it delivers high-resolution, accurate and lithologically feasible subsurface models. The technique enhances the benefits of using an elastic wave equation over the acoustic implementation while constraining the inverted models to geologically plausible solutions. We derive elastic models of subsurface properties using EFWI and explicitly incorporate petrophysical penalties to guide models toward realistic lithology, i.e., to models consistent with the seismic data as well as with the petrophysical context in the area of study. This methodology mitigates several issues related to EFWI, as it reduces the high non-linearity of the inverse problem, mitigates the artifacts created by interparameter crosstalk, and prevents geologically implausible earth models. We define this penalty using multiple probability density functions (PDFs) derived from petrophysical information, such as well logs, where each PDF represents a different lithology. We demonstrate that the combined FWI objective function establishes a more robust foundation for EFWI by explicitly guiding models toward plausible solutions in the specific geological context of the exploration problem, while at the same time reducing the misfit between the observed and modelled data.
-
-
-
Subset 3D Geological Modelling for Petroleum Reservoir Estimation
Authors E. Pakyuz-Charrier and J. KeetleySummaryObtaining a reliable understanding of the geological structures at depth is essential in offshore Petroleum exploration scenarios. 3D geological modelling software such as GeoModeller may be used to answer this need. GeoModeller uses an implicit fully 3D modelling engine, the structural data is used directly to construct the volumes and surfaces with a cokriging interpolator. Geological formations are then topologically sorted using a binary (Erode/Onlap) stratigraphic pile. As both the input data and the modelling engine – through measurement errors and simplifications, respectively – are inherently imperfect, the end result of the modelling process is necessarily uncertain. A range of Monte Carlo Uncertainty Propagation (MCUP) methods have been developed over the last decade to provide the community with means to estimate uncertainty. In this paper, Intrepid Geophysics explores MCUP methods and their application to Oil & Gas reservoir estimation with the unique constraint of using MCUP methods to reduce uncertainties and propose alternative scenarios in a hybrid driven environment.
-
-
-
The Role of Minimum Phase in Internal Multiple Removal
By M. DukalskiSummaryGiven true-amplitude pre-processed data, Marchenko equation based methods could remove all overburden-borne internal multiples without the adaptive subtraction. The method hinges on calculating an inverse transmission response, however in many practical cases to find a solution, one is required to provide a part of it on input. This requirement can be lifted by invoking minimum phase - a mathematical property familiar to many geophysicists, yet normally not associated with a de-multiple workflow. Here we discuss the state of the art, challenges and road ahead for minimum phase enriched internal de-multiple. In particular we focus on the differences in minimum phase reconstruction between single input single output (1.5-D single mode) vs multiple input multiple output systems (everything else).
-
-
-
Realistic Uncertainty Estimation in an Integrated Geostatistical Seismic Property Modeling
More LessSummaryGeostatistical Seismic property modelling is about taking all available prior information and measurements into account. Evidently, one very important and valuable piece of information is the seismic data, which has extensive lateral coverage compared to the sparse information provided at well locations. Using Bayesian inference, we can incorporate a variety of data and expert knowledge to obtain predictions that are more realistic.
At the same time, using multiple realizations we are taking into account overall uncertainties including both variance and bias ( Moradi Tehrani 2016 ) that give us a handle to estimate and mitigate risk and make informed decision.
As there is no single right answer, the answer depends on the question. If we know our objective, then we can get the answer. In this case, we can also estimate how reliable the answer is.
By analyzing the realizations through ranking, we can determine the range of possible answers. Ranking relies on defining an objective criterion that captures the key characteristic of interest.
This approach is powerful for realistic uncertainty estimation and therefore provides a sturdier foundation for making informed decisions.
-
-
-
Elastodynamic Marchenko Method: Advances and Remaining Challenges
Authors C. Reinicke, M. Dukalski and K. WapenaarSummaryMarchenko methods aim to remove all overburden-related internal multiples. The acoustic and elastodynamic formulations observe identical equations, but different physics. The elastodynamic case highlights that the Marchenko method only handles overburden-generated reflections, i.e. forward-scattered transmitted waves (and so-called fast multiples) remain in the data. Moreover, to constrain an underdetermined problem, the Marchenko method makes two assumptions that are reasonable for acoustic, but not for elastodynamic waves. Firstly, the scheme requires an initial guess that can be realistically estimated for sufficiently-simple acoustic cases, but remains unpredictable for elastic media without detailed overburden knowledge. Secondly, the scheme assumes temporal separability of upgoing focusing and Green’s functions, which holds for many acoustic media but easily fails in presence of elastic effects. The latter limitation is nearly-identical to the monotonicity requirement of the inverse scattering series, indicating that this limitation may be due to the underlying physics and not algorithm dependent. Provided that monotonicity holds, the aforementioned initial estimate can be retrieved by augmenting the Marchenko method with energy conservation and a minimum-phase condition. However, the augmentation relies on the availability of an elastic minimum-phase reconstruction method, which is currently under investigation. Finally, we discuss a geological setting where an acoustic approximation suffices.
-
-
-
When Do We Need Elastic Waveform Inversion for Velocity Model Building? Marine and Land Examples
Authors C. Pérez Solano, G. Chang, R. Plessix, K. Bao, A. Stopin and X. WangSummaryFull waveform inversion is a widespread data-fitting technique commonly used for velocity model building. In this context, the target input data are low-frequency transmissions and post-critical reflections in a long offset range. These seismic data, associated to diving waves, are highly sensitive and therefore can be used to retrieve the P-wave velocities of the subsurface. However, according to the diffraction theory, the transmission data can be influenced by shear parameter variations in the subsurface which create tuning and interference effects at the low frequencies, notably when the variations are large and occur inside the first Fresnel zone. The elastic interference challenges the applicability of an acoustic waveform inversion approach in a global manner, and despite its relatively low computational costs and practicality, its results have proven mixed. We discuss marine and land seismic data examples, from the Gulf of Mexico and the Middle-East, where elastic waveform inversion provides compelling results superseding the acoustic waveform inversion results.
-
-
-
3D Elastic FWI for Land Seismic Data: A Graph Space OT Approach
Authors W. He, R. Brossier and L. MetivierSummaryIntegrating surface wave information is challenging for land seismic full waveform inversion. Cycle-skipping of surface waves can easily occur due to their highly dispersive and oscillating properties. While this issue can be mitigated using wider basin objective functions, a more severe difficulty is related to the unbalanced amplitude distribution between surface waves and body waves. The energetic surface waves dominate the objective function and drive the inversion to update only the shallow structure. The contribution from body waves is masked and the deep structures are not recovered. In a recent study, we have shown how an optimal-transport based function can help mitigating this issue, providing naturally a better balance between events (KR-OT). Here, we apply a newly introduced OT based misfit function, relying on a graph space approach (GS-OT), in this framework of elastic FWI for land data. GS-OT better handles cycle skipping than KR-OT. We show here that it also helps to balance the amplitude of seismic events. We design a practical workflow based on the GS-OT misfit function, coupled with an on-the-fly source estimation wavelet and a Gaussian-time window strategy. The method is applied to a synthetic case study from the SEAM II Foothill model.
-
-
-
Reconstruction of Quantitative Variable Density Acoustic Reflectivity in the Context of Velocity Model Building
Authors M. Farshad and H. ChaurisSummaryThe quality of focusing panels (Common Image Gathers) plays a fundamental role in the construction of the macro-model via image domain techniques. Recent works demonstrated that iterative least-squares migration is recommended for obtaining reliable focusing panels: this ensures relevant tomographic macro-velocity updates. In practice, iterative least-squares migration needs to be accelerated through suitable pre-conditioners such as pseudo-inverses of the forward modelling operator. The pseudo-inverses are currently limited to the constant density acoustic case. In this paper, we first discuss the impact of density variations on focusing panels, and then propose an approach to quantitatively reconstruct two acoustic parameters. The main ingredient is the Radon transform. From an extended reflectivity (single iteration), we apply the Radon transform to reconstruct the inverse of the bulk modulus and the density perturbations in the physical domain, while preserving the data fit. We validate our approach on the Marmousi-II dataset, demonstrating that the proposed approach is an efficient alternative to the more expensive least-squares migration. As expected, there is a leakage between the inverted two parameters.
-
-
-
Uncertainty Quantification Enhancement by Combining Data of Varying Accuracy and Precision
Authors J. Chautru, H. Binet, P. Masoudi, S. Rodriguez and M. PapouinSummaryA common issue in Depth conversion and Volumetrics calculation is the estimation of the range of variation of important reservoir parameters, like structural closure depth or reservoir volumes, which are often poorly estimated. For example, the GRV and hydrocarbon volumes in most of the developed reservoirs are found to be closer to the P90 than to the P50 determined at the exploration phase, which means that the range of variation is underestimated.
Combining data sources of varying precision and accuracy such as seismic data, CCAL, SCAL or logs, which correspond to different scales, is a difficult issue. It requires appropriate mathematical tools, like the ones provided by Geostatistics. This paper details the geostatistical techniques than can be used for calculating geological models and quantifying the associated uncertainty. Focus is put on the role played by input data accuracy and precision in the final estimation of GRV and hydrocarbon volumes range of variation, accounting for the varying accuracy and precision that are manageable in practice. This allows decision makers to take the most appropriate decisions concerning the field development. The methodologies and results are illustrated with a simplified modelling based on a real dataset, emphasis being put on structural modelling.
-
-
-
Assessing and Communicating Geological Model Uncertainty
Authors A.K. Turner and M. BianchiSummaryAn overview of recent alternative procedures for assessing and communicating the multiple sources of uncertainty in geological models
-
-
-
Extending the Potential of Acoustic Optimal Transport FWI in the South of Oman
Authors S. Shutova, D. Carotti and O. HermantSummaryBroadband land WAZ acquisitions provide the ideal data for land FWI thanks to low frequencies and long offsets. While acoustic FWI has been successful in the north of Oman, the complex near-surface geology and the presence of strong velocity contrasts in the south of Oman better fits with the use of elastic propagation to obtain an accurate velocity model. In this paper we discuss the challenge of improving acoustic land FWI results in the south of Oman, thus avoiding the use of elastic FWI, which is still prohibitively expensive for deriving high-resolution velocity models. The key success factors for this acoustic land FWI study were proper pre-processing of the input seismic data and the use of an Optimal Transport objective function to mitigate cycle skipping issues.
-
-
-
The Use of Multi-Scattered Energy in Imaging and Velocity Model Building
Authors T. Alkhalifah, Q. Guo, G. Wang and Z. WuSummaryMultiples and multi-scattered seismic recorded energy are often ignored or suppressed in imaging and inversion applications. This happens because they tend to appear weak in our recorded data and they also violate the linearity clause, complicating their treatment. On the other hand, they carry valuable information that allows for better illumination of the Earth and better velocity model building. We will share those features through our generalized internal multiple imaging strategy, as well as incorporating multi scattered energy in Full waveform inversion. The results demonstrate the ability of multi-scattered energy to better image the subsurface, as well as yield a better velocity models.
-
-
-
Quantification of Effective Permeability Uncertainty Estimation on Geothermal Aquifer Scale
Authors M. Verberne, R. Dalman, J. Breunese, B. Van Kempen and K. GeelSummaryThe economic extraction of energy from geothermal systems and hydrocarbon reservoirs rely largely on production rates, which in turn is largely reliant on the permeability. The effective permeability of a reservoir is often inconsistent with the local (borehole) permeability leading to high rate of economic failures. This mismatch is caused by heterogeneities related to e.g. permeability contrasts, faulting and other types of low permeable barriers such as lateral facies changes, these are notoriously hard to capture in permeability estimation.
This study empirically tests several analytical methods to estimate the effective permeability on a production system scale for a suite of geological facies and permeability distributions representative for the Netherlands deep subsurface. Ultimately we construct a toolbox of optimal techniques for determining the accuracy and precision distribution for permeability for a suite of genetically related reservoir types. Thereby allowing a significantly improved estimation of effective permeability and subsequently derisking pre- and postdrill evaluations of both geothermal and hydrocarbon projects.
-
-
-
An Overview of Marchenko Methods
Authors K. Wapenaar, M. Staring, J. Brackenhoff, L. Zhang, J. Thorbecke and E. SlobSummarySince the introduction of the Marchenko method in geophysics, many variants have been developed. Using a compact unified notation, we review redatuming by multidimensional deconvolution and by double focusing, virtual seismology, double dereverberation and transmission-compensated Marchenko multiple elimination, and discuss the underlying assumptions, merits and limitations of these methods.
-
-
-
Data-Driven Internal Multiple Elimination Applications Using Imperfectly Sampled Reflection Data
Authors J. Brackenhoff, J. Van IJsseldijk and K. WapenaarSummaryWe consider reflection data that have been subsampled by 70% and use Point-Spread-Functions to reconstruct the original data. The subsampled, original and reconstructed reflection data are used to image the medium of interest with the Marchenko method. The image obtained using the subsampled data shows artifacts caused by internal multiples, which are eliminated when the original and reconstructed data are used.
-
-
-
Including Internal Multiples in Imaging and Velocity Updating: Potential and Limitations
Authors E. Verschuur and Y. SunSummaryIn many areas around the world the impact of internal multiples on seismic data can be strong, due to large subsurface contrasts. There is a choice of removing these multiples in advance and apply primaries-only imaging. Such a method often requires densely sampled data, especially for Marchenko-based approaches. If data sampling is coarser, an image-driven approach can be adequate, where the Earth model is used as a constraint in the inversion. Such an approach is followed by full wavefield migration (FWM) and joint migration inversion (JMI), where the subsurface is parametrized by reflectivity and a velocity model. In this way, internal multiples, together with transmission effects, can be handled and an imprint-free image is ideally obtained. In this paper, the potential and limitations of this approach will be analyzed.
-
-
-
Predicting Internal Multiples by Inversion Based Imaging
Authors M. Davydenko and D.J. VerschuurSummaryTo handle internal multiples in imaging can be done via data-driven procedures (e.g. Marchenko-based multiple removal) or via an image-driven process that includes an option to model internal multiples from the reflectivities. The advantage of data-driven methods is that it can utilize the physics implicitly available in the data, while the image-driven method has to make this physical model more explicit via an image, which may create some leakage. However, a main advantage of the image-driven approach is that it is very insensitive to coarse source sampling, while data-driven methods requires dense data. The aspect of coarse sampling and its effect on the final multiple suppression effects of the image-driven approach is demonstrated.
-
-
-
Target-Oriented Elastic Parameter Models Estimation from Surface Seismic Data Uusing JMI-res
Authors A. Garg and E. VerschuurSummaryElastic full waveform inversion (FWI) has the potential to provide a high-resolution velocity model by inverting all wavelengths of the subsurface structures. However, due to the associated computational costs and non-linear coupling between different elastic parameters, it is only applied for low frequencies, especially in a full elastic 3D fashion. Mostly, high-resolution elastic parameter models are only required within the limited area of the earth subsurface. Therefore, high-resolution elastic-FWI for full bandwidth can also be restricted to a target area. With this concept in mind, we present reservoir-oriented joint migration inversion (JMI-res) using both a numerical and a field data example. It first reconstructs the localized elastic data at the reservoir depth (local impulse responses) and then use this as input for elastic-FWI for the area of interest. In addition, the redatuming step of JMI-res correctly account for the overburden-related multiple scattering. Thus, the localized elastic data is of high-resolution and free of overburden artefacts, while avoiding a complete elastic FWI for the whole medium.
-