WO2021002761A1 - Inversions améliorées de données géophysiques - Google Patents

Inversions améliorées de données géophysiques Download PDF

Info

Publication number
WO2021002761A1
WO2021002761A1 PCT/NO2020/050187 NO2020050187W WO2021002761A1 WO 2021002761 A1 WO2021002761 A1 WO 2021002761A1 NO 2020050187 W NO2020050187 W NO 2020050187W WO 2021002761 A1 WO2021002761 A1 WO 2021002761A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
models
vectors
sub
upgoing
Prior art date
Application number
PCT/NO2020/050187
Other languages
English (en)
Inventor
Emmanuel Causse
Original Assignee
Equinor Energy As
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Equinor Energy As filed Critical Equinor Energy As
Publication of WO2021002761A1 publication Critical patent/WO2021002761A1/fr

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V20/00—Geomodelling in general
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/30—Analysis
    • G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
    • G01V3/08—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with magnetic or electric fields produced or modified by objects or geological structures or by detecting devices
    • G01V3/083—Controlled source electromagnetic [CSEM] surveying
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
    • G01V3/38—Processing data, e.g. for analysis, for interpretation, for correction
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
    • G01V3/08—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with magnetic or electric fields produced or modified by objects or geological structures or by detecting devices
    • G01V3/083—Controlled source electromagnetic [CSEM] surveying
    • G01V2003/086—Processing
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/614—Synthetically generated data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/616—Data from specific type of measurement
    • G01V2210/6161—Seismic or acoustic, e.g. land or sea measurements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/616—Data from specific type of measurement
    • G01V2210/6163—Electromagnetic
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/62—Physical property of subsurface
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/66—Subsurface modeling

Definitions

  • the present invention relates to a method of estimating properties of a geological structure and more specifically to non-linear inversions of geophysical data including electromagnetic and seismic data.
  • Seismic imaging has long been used to determine the geological structure of a sub surface region, for example as an aid to the identification of hydrocarbon bearing regions.
  • the basis of all seismic imaging techniques is the generation at the surface, or in a body of water above the surface, of acoustic sound waves which propagate through the surface and the detection of upgoing waves at the surface (or again within the body of water) which result from reflection of the downgoing waves from sub surface reflectors.
  • upgoing wave refers generally to a wave which has been reflected at a sub-surface location, but should not be interpreted narrowly as travelling only in vertical direction because the wave will propagate in all directions.
  • the process of generating an image of the sub-surface from the seismic data is known as “inversion”.
  • EM electromagnetic
  • CSEM Controlled Source EM
  • Imaging is often enormously computationally expensive, meaning that it is both financially expensive and time consuming to perform. If an assumption turns out to be incorrect, it might be several days before that assumption can be corrected and an analysis re-run. Taking shortcuts can lead to poor imaging quality leading, in a worst- case scenario, to drilling wells in a poor location or making incorrect assumptions about existing hydrocarbon reserves.
  • the model consists of a 3-dimensional representation of the sub-surface made up of a large number of cells. Each cell is assigned a resistivity (or its inverse, conductivity) value. At each iterative step, the simulated results are compared with the recorded data to generate some difference measure and the model adjusted to reduce the difference.
  • the process of selecting the adjustment to make can be very computationally expensive as it typically requires adjusting the value of each cell in turn to determine which adjustment generates the best reduction in the determined difference.
  • mk and mk+1 are vectors defining the current model and the subsequent improved model (in terms of resistivity values within each cell of the model), mo is the initial model, Ad is the difference between the simulated data in model mk and the recorded data, and is the jacobian matrix of data derivatives with respect to the model parameters and * denotes a conjugate transpose.
  • J The information contained by J is important: it shows how the data difference is changed by changing the resistivity independently in each cell of the model. is the inverse of the covariance matrix of the data errors and is the inverse of the covariance matrix of model parameters describing a priori information on the model. H r will be explained later.
  • the inversion has to find one million unknowns.
  • a first problem is the very large numerical cost this generates.
  • a second problem is that the recorded CSEM data are not sufficient to properly constrain the problem: basically there are too few equations for the number of unknowns, so that an infinity of solutions exists.
  • the inversion problem is ill-posed and this generates mathematical instability that can result in errors and artefacts in the obtained resistivity models.
  • a mathematical approach called“regularization”.
  • a computer implemented method of obtaining a model of a sub-surface region comprising a three- dimensional matrix of sub-surface material property values, and comprising: a) providing a set of geologically realistic models, constructing a reference matrix using said models, and applying an analysis to that reference matrix to describe the reference matrix using vectors; b) selecting the most important vectors from said set of vectors to obtain a subset of vectors that describe said reference matrix to an approximation; c) obtaining upgoing energy wave data recorded at a plurality of spaced apart locations within, at or above the surface, the upgoing energy waves depending on downgoing energy waves and the sub-surface material properties; d) defining as a first approximation an initial model; e) determining a change to said initial model to provide an improved model, said change being one that seeks to minimise a penalty defined as a function of a difference between upgoing energy wave data simulated at said locations using the improved model and the recorded up
  • the initial model may comprise a combination of said most important vectors.
  • the analysis is a principal component analysis and said vectors are singular vectors.
  • the singular vectors may comprise one hundred or fewer singular vectors.
  • the number of vectors in said subset of vectors may be less than the number of models in said set of geologically realistic models.
  • the downgoing and upgoing energy waves may be electromagnetic energy waves, and said sub-surface material property values may be electrical resistivity or conductivity values.
  • said downgoing and upgoing energy waves may be seismic energy waves, and said sub-surface material property values may be seismic velocity and density values.
  • the set of geologically realistic models may comprise five hundred or fewer geologically realistic models.
  • Step d) may comprise operating one or more energy sources at, above, or within a surface of said sub-surface region to produce said downgoing energy waves and operating detectors at said spaced apart locations to detect said upgoing energy waves, and recording the detected upgoing waves.
  • step d) may comprise operating one or more energy sources within a body of water above a surface of said sub-surface region to produce said downgoing energy waves and operating detectors at said spaced apart locations to detect said upgoing energy waves, and recording the detected upgoing waves.
  • a method of selecting a region in which to drill in order to produce hydrocarbons comprising applying the method according to the first aspect to obtain a model of a sub-surface region, and examining the model to identify locations within the sub-surface likely to produce hydrocarbons.
  • a method of producing hydrocarbons comprising selecting a region in which to drill using the method of the second aspect, drilling one or more wells in that region, and producing hydrocarbons from the well or wells.
  • a computer program product comprising program instructions operative to be executed by a processor to cause the processor to perform a method according to the first aspect.
  • Figure 1 illustrates a set or reference models in a two dimensional vector space
  • Figure 2 also illustrates a set or reference models in a two dimensional vector space
  • Figure 3 is a flow diagram illustrating a method of obtaining a model of a sub-surface region
  • Figure 4 is a flow diagram further illustrating a method of obtaining a model of a sub surface region
  • Figure 5 is a flow diagram illustrating a method estimating an uncertainty of a model of a sub-surface region
  • Figure 6 is a flow diagram illustrating a method estimating an uncertainty of a model of a sub-surface region and utilising a plurality of different data types.
  • Non-linear inversion of CSEM data is a process by which an initial model of the subsurface is iteratively improved by reducing the misfit between data simulated in that model and the data recorded at seabed during an acquisition campaign.
  • the model is typically represented as a coloured 3D image with small cubic pixels, where the colour represents the value of the resistivity (or its inverse, conductivity).
  • the value of resistivity (or conductivity) in each pixel is updated to provide a better model.
  • What we mean here by“better” is a model that will result in simulated data that are more like the recorded data than at previous iteration, i.e. a model that reduces the data misfit.
  • the model update provided by the formula can be very unstable: it may for instance provide a physically unrealistic model (i.e. the colours are wrong) or a model with a lot of artefacts (the image is very noisy).
  • the instability can be treated in a mathematical way. For instance, adding a small constant to the diagonal of the system matrix will stabilize the system. This is a particular form of what is generally called Tikhonov regularization. Basically, regularization enforces a form of smoothness in the solution of the system. Here, it means that two neighbouring pixels will tend to have similar values rather than very different values. This reduces artefacts in the image.
  • CSEM inversion methods typically use Tikhonov regularization or some variant of it.
  • regularization the inversion tries not to reduce the data misfit, but rather tries to reduce a cost function which is the sum of the data misfit and a second term which can be called“model penalty”, and which is large for models that present a lot of contrasts.
  • model penalty a second term which can be called“model penalty”, and which is large for models that present a lot of contrasts.
  • the user has to adjust many parameters to tune regularization, and those can have a large impact on the final inversion result.
  • For CSEM one tries for instance to choose parameters that favour vertical contrasts rather than horizontal contrasts to increase the chance to detect hydrocarbon reservoirs, which are typically horizontally extending resistive objects.
  • Regularization may be introduced as a way to stabilize inversion, which otherwise provides unstable results. Regularization introduces a second term representing a model penalty, in addition to the data misfit that inversion tries to reduce. This second term penalizes variations in the 3D image.
  • the user chooses regularization parameters, runs the inversion (this can take days), checks the image, and tunes the regularization parameters to improve the image. This can be a difficult and time- consuming process.
  • the inventors have realised that the numerical cost can be reduced significantly by forcing the inversion to provide realistic models and at the same time to drastically reduce the number of unknowns, from one million to typically less than 100. When doing this, the number of numerical operations involved for calculating the model update is dramatically reduced.
  • Inversion seeks to find a model that minimizes a cost function.
  • the model update done at each iteration depends on how this cost function is defined.
  • the simplest approach is to define the cost function as the data misfit only:
  • 3D pixels is a co
  • Equation (1) is a least- squares sum over the different data points contained in
  • a typical value for is a constant times the identity matrix. If the constant is sufficiently large, the system becomes stable enough.
  • a set of geologically realistic models (the“reference set”) is used to represent the prior information available about the subsurface properties. This is discussed below in the context of electromagnetic (specifically, CSEM) based methods where the model consists of a matrix of resistivity values. However, the approach is equally applicable to seismic approaches in which case the model may consist of a matrix of acoustic impedance values.
  • Non-linear inversion algorithm leads typically to one single solution but the interpreter must be aware that other solutions may be valid. In practice, one will normally run several inversions with slightly different stabilization parameters leading to different 3D images. The best images can be picked for more detailed interpretation. Running many inversions is however costly.
  • the stabilization often is based on simplistic assumptions that do not reflect the complexity of the earth (e.g. in reality spatial resistivity variations show abrupt variations at several locations in the subsurface) the solutions found by inversion can be quite unrealistic, i.e. can badly fit the geological understanding of the subsurface obtained from other type of knowledge, e.g. based on seismic data, on measurements of rock properties in wells or general knowledge about the geological setting. For instance, high resistivity may be expected at the location of a hydrocarbon-filled reservoir. The location of this potential reservoir may be known quite accurately from interpretation of seismic data and measurements in neighboring wells.
  • the inversion may provide a solution where high resistivity in the 3D cube is erroneously imaged too shallow or too deep, at a depth where it is highly unlikely to find any hydrocarbons.
  • the 3D image is, geologically speaking, unrealistic.
  • the interpreter must consider the possibility of such errors, which complicates the interpretation process and reduces the value of the inversion result.
  • New inversion runs with new sets of stabilization parameters may be required to cope with that problem, adding more computer costs.
  • the proposed solution is to describe the space of possible, realistic geological models beforehand to reduce the possibility of such errors.
  • the inversion will be forced to look for models that by construction respect the available information available on the subsurface’s geology.
  • Such information is referred to as“a-priori” information, in the sense that it is available prior to any use of CSEM data.
  • the prior information is translated into a large number N of possible 3D resistivity models, for example 250 resistivity models, that we may call“prior models”.
  • N realistic 3D prior resistivity models describe a linear space that we can call the prior space.
  • N c singular vectors where N c is smaller than N, i.e. each of the N 3D models can be represented almost exactly by a linear combination of the N c singular vectors.
  • N c represents the number of degrees of freedom in the space of possible 3D resistivity models built from the available prior information.
  • N c may typically be lower than 50.
  • each reference model comprises a set of cells, e.g. 1 to N m
  • the reference set can be represented by a matrix M where each column of the matrix corresponds to a given reference model, or to the difference between a given reference model and the average of all reference models (the most probable model). For example, if each model comprises 1000000 cells and there are 250 models, then M will be a 1000000 X 250 matrix.
  • the singular vectors (F 0 ) are ordered so that the first vectors are relatively more organised whilst the later ones are less organised, i.e. more chaotic.
  • the singular values (S 0 ) are sorted in decreasing order and only the first few are retained. Whilst the matrix M now no longer exactly describes the reference models, the approximation can still be very good, i.e. M » FSK, where F consists of the N c first columns of F 0 , S corresponds to the N c first rows and columns of S 0 and K corresponds to the N c first rows of K 0 .
  • the singular values S and coefficients K can be combined into a single matrix W with components which represents the weights w i for each of the reference models.
  • the known inversion approaches explained above for m can be applied to w or c instead. Since the number of unknowns is strongly reduced, the computational burden is now much smaller. As soon as the unknown values in vectors w or c are found by inversion, the values in vector m are easily obtained by the formulae above.
  • the confidence ellipses correspond to one and two standard deviations and indicate the general spread of the reference models, i.e. the prior uncertainty on model properties. We see that weight w t is of greater significance than weight w 2 and has a larger prior uncertainty.
  • Figure 2 shows how the same set of reference models is represented by the coefficients c 1 and c 2 of vector c instead.
  • the ellipses are now circles, indicating that the prior covariance matrix (or“the covariance matrix of model parameters”) for vector c is the identity matrix.
  • matrix H reg is in this case equal to the identity matrix.
  • Figure 3 illustrates a method comprising the steps of:
  • an initial model for example the average of all of the set of geologically realistic models.
  • step e) iteratively repeating step e) for each improved model until an appropriately accurate model is obtained.
  • a technical effect of the claimed method is a great reduction in computational cost.
  • the computer runtime for a typical inversion process may be in the order of weeks or even months, with a corresponding amount of electrical energy consumed by the computer.
  • the number of unknowns may be reduced from one million to less than 100, and the computational runtime and electricity consumption will be reduced correspondingly.
  • a further effect is an improved estimation accuracy because the models are restricted to geologically realistic models, without allowing the algorithm to arrive at solutions which produce a good match to the upgoing energy wave data while providing a model of the subsurface which geologically cannot be true.
  • This improved estimation may avoid, for example, drilling wells in a poor location or making incorrect assumptions about existing hydrocarbon reserves.
  • hydrocarbons may be located in a trap formed by a suitably shaped medium which inhibits upward migration of hydrocarbons through permeable rock formation. If the shape of the medium is incorrectly determined, the location of the trapped hydrocarbons is also incorrectly determined and an incorrect drilling operation may be carried out.
  • the embodiments described above may be incorporated into an exploration process to identify regions likely to produce hydrocarbons. As part of that process data may be obtained by performing EM, seismic or other surveys in the area of interest. Of course, historical data may be reanalysed to provide new insight into a region. The result of the process may be the drilling of one or more wells within the location.
  • the workflow of Figure 5 illustrates a use of the workflow of Figure 4 to produce a set of output models for estimating uncertainty.
  • the model generation process is repeated multiple times. Whilst the same set of realistic models / geological scenarios are used for cycle, a different initial model is used. For example, a different one of the set of realistic models may be used.
  • Experiments demonstrate that a slightly different output model will be returned for each cycle. It will be appreciated that, if all of the output models are very similar, there is likely to be little uncertainty on the estimated earth properties (i.e. on the resistivity in the CSEM case). However, if the output models differ significantly, the uncertainty is likely to be much greater.
  • model space covered by the final models will be much smaller than the space covered by the initial models. This is because most of the input models are not compatible with the observed data. It may happen however that the model space covered by the output models is not contained within the space covered by the input models. This can happen, for instance, if the original assumptions (the prior information) are incorrect (e.g. it was assumed that the resistivity in a given part of the model is between 1 and 4 but the result is resistivities around 5).
  • the workflow of Figure 5 can provide extremely useful additional information: for example, by counting the number of models in each of the clusters, one may obtain an indication of the likelihood of finding hydrocarbons versus the likelihood of not finding hydrocarbons.
  • an efficient approach is to use the set of realistic models or a subset of that set, as alternative initial models: these models cover the whole“prior space” evenly (this is related to the mathematical properties of the singular value decomposition. By starting with evenly distributed prior models the chance to cover all potential clusters (if there are several) is increased.
  • the approaches of Figures 4 and 5 assume that we are concerned only with a single type of geophysical data. In many cases however, a number of different types of data are available.
  • the workflow of Figure 6 illustrates one possibility for using different types of geophysical data to improve the model generation process. Considering again the workflow of Figure 5, this assumes that, for a given a set of input models (a cloud of points), the data will“shrink” the set of models by removing those that are incompatible with the data, retaining only those that are compatible with the data, thereby providing a“narrower” set of models.
  • the final cloud(s) obtained can be even more focused: some ambiguities may be removed and the uncertainties typically be reduced, giving fewer and narrower clouds, and a corresponding population of models with more precisely estimated properties.
  • each new type of geophysical data may be used to shrink the space of possible models even more.
  • CSEM data indirectly shrinks the input cloud representing the initial velocity models, compared to a case where no CSEM is be used.
  • this shrinking process may not be very efficient since general correlations between velocities and resistivities are not very strong.
  • the models may contain the properties that each type of data relates to.
  • the vectors representing the set of input realistic models may also contain seismic velocities, densities, anisotropies and all quantities the used type of geophysical data is sensitive to.
  • This approach is the most complete but may be challenging to use due to issues related to the scaling between the different data types, and because the different types of data are often not available at the same time. For instance, seismic data are typically available long before CSEM data are acquired.

Landscapes

  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Geophysics (AREA)
  • Environmental & Geological Engineering (AREA)
  • Geology (AREA)
  • Electromagnetism (AREA)
  • Acoustics & Sound (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

L'invention concerne un procédé informatique permettant d'obtenir un modèle d'une région de sous-surface, le modèle comprenant une matrice tridimensionnelle de valeurs de propriété de matériau de sous-surface, et comprenant : a) la fourniture d'un ensemble de modèles géologiquement réalistes, la construction d'une matrice de référence à l'aide desdits modèles, et l'application d'une analyse à cette matrice de référence pour décrire la matrice de référence à l'aide de vecteurs ; b) la sélection des vecteurs les plus importants à partir dudit ensemble de vecteurs pour obtenir un sous-ensemble de vecteurs qui décrivent ladite matrice de référence en une approximation ; c) l'obtention de données sur les ondes d'énergie montante enregistrées à plusieurs endroits espacés les uns des autres, à la surface ou au-dessus de la surface, les ondes d'énergie montante dépendant des ondes d'énergie descendante et des propriétés des matériaux de sous-surface ; d) la définition sous la forme d'une première approximation d'un modèle initial ; e) la détermination d'un changement dans ledit modèle initial pour fournir un modèle amélioré, ledit changement en étant un qui cherche à minimiser une pénalité définie en fonction d'une différence entre des données d'ondes d'énergie montante simulées auxdits emplacements à l'aide du modèle amélioré et des données d'ondes d'énergie montante enregistrées, le modèle amélioré comprenant une combinaison desdits vecteurs les plus importants ; et f) la répétition de manière itérative de l'étape e) pour chaque modèle amélioré jusqu'à ce qu'un modèle de précision appropriée soit obtenu.
PCT/NO2020/050187 2019-07-02 2020-06-30 Inversions améliorées de données géophysiques WO2021002761A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB1909529.8A GB2585216B (en) 2019-07-02 2019-07-02 Improved inversions of geophysical data
GB1909529.8 2019-07-02

Publications (1)

Publication Number Publication Date
WO2021002761A1 true WO2021002761A1 (fr) 2021-01-07

Family

ID=67540081

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/NO2020/050187 WO2021002761A1 (fr) 2019-07-02 2020-06-30 Inversions améliorées de données géophysiques

Country Status (2)

Country Link
GB (1) GB2585216B (fr)
WO (1) WO2021002761A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114063177A (zh) * 2021-11-16 2022-02-18 核工业北京地质研究院 一种大地电磁数据去噪方法及系统
CN115267927A (zh) * 2022-09-28 2022-11-01 中石化经纬有限公司 一种基于蚁群-梯度串联算法的多边界幕式地质导向方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6839658B2 (en) * 2000-07-17 2005-01-04 Sintef Petroleumsforskning Seismic processing with general non-hyperbolic travel-time corrections
US7333392B2 (en) * 2005-09-19 2008-02-19 Saudi Arabian Oil Company Method for estimating and reconstructing seismic reflection signals
US8688616B2 (en) * 2010-06-14 2014-04-01 Blue Prism Technologies Pte. Ltd. High-dimensional data analysis
WO2019143253A1 (fr) * 2018-01-22 2019-07-25 Equinor Energy As Régularisation d'inversions non linéaires de données géophysiques
WO2019209114A1 (fr) * 2018-04-26 2019-10-31 Equinor Energy As Prévision d'une incertitude dans des inversions non linéaires de données géophysiques

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3295220A1 (fr) * 2015-05-08 2018-03-21 Statoil Petroleum SA Compression de modèle

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6839658B2 (en) * 2000-07-17 2005-01-04 Sintef Petroleumsforskning Seismic processing with general non-hyperbolic travel-time corrections
US7333392B2 (en) * 2005-09-19 2008-02-19 Saudi Arabian Oil Company Method for estimating and reconstructing seismic reflection signals
US8688616B2 (en) * 2010-06-14 2014-04-01 Blue Prism Technologies Pte. Ltd. High-dimensional data analysis
WO2019143253A1 (fr) * 2018-01-22 2019-07-25 Equinor Energy As Régularisation d'inversions non linéaires de données géophysiques
WO2019209114A1 (fr) * 2018-04-26 2019-10-31 Equinor Energy As Prévision d'une incertitude dans des inversions non linéaires de données géophysiques

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
P. ELIASSON ET AL.: "Uncertainty quantification in waveform-based imaging methods - a Sleipner CO2 monitoring study", 13TH INTERNATIONAL CONFERENCE ON GREENHOUSE GAS CONTROL TECHNOLOGIES, GHGT-13, vol. 114, 2017, Lausanne, Switzerland, pages 3905 - 3915, XP085178009 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114063177A (zh) * 2021-11-16 2022-02-18 核工业北京地质研究院 一种大地电磁数据去噪方法及系统
CN114063177B (zh) * 2021-11-16 2023-09-29 核工业北京地质研究院 一种大地电磁数据去噪方法及系统
CN115267927A (zh) * 2022-09-28 2022-11-01 中石化经纬有限公司 一种基于蚁群-梯度串联算法的多边界幕式地质导向方法
CN115267927B (zh) * 2022-09-28 2022-12-30 中石化经纬有限公司 一种基于蚁群-梯度串联算法的多边界幕式地质导向方法

Also Published As

Publication number Publication date
GB2585216B (en) 2021-12-01
GB201909529D0 (en) 2019-08-14
GB2585216A (en) 2021-01-06

Similar Documents

Publication Publication Date Title
EP0750203B1 (fr) Modélisation de subsurface à partir de données sismiques et de mesures secondaires
EP2810101B1 (fr) Amélioration de l'efficacité d'algorithmes d'inversion à base de pixels
US8095345B2 (en) Stochastic inversion of geophysical data for estimating earth model parameters
US9195783B2 (en) Reducing the dimensionality of the joint inversion problem
US9207344B2 (en) Combining geomechanical velocity modeling and tomographic update for velocity model building
US20090083006A1 (en) Methods and apparatus for three-dimensional inversion of electromagnetic data
WO2010151354A1 (fr) Construction de modèles de résistivité à partir d'une inversion stochastique
US11555936B2 (en) Analytics and machine learning method for estimating petrophysical property values
CA3069322C (fr) Limites de materialite de reservoir a partir d'une inversion sismique
Xiao et al. Conditioning of deep-learning surrogate models to image data with application to reservoir characterization
CN107810432B (zh) 模型压缩
WO2021002761A1 (fr) Inversions améliorées de données géophysiques
Zhou et al. An intelligent MT data inversion method with seismic attribute enhancement
US20210405250A1 (en) Conditioning of Surface-Based Geologic Models
Colombo et al. Machine-learning inversion via adaptive learning and statistical sampling: Application to airborne micro-TEM for seismic sand corrections
WO2019143253A1 (fr) Régularisation d'inversions non linéaires de données géophysiques
Zhou et al. Stochastic structure-constrained image-guided inversion of geophysical data
Van Zon et al. Structural inversion of gravity data using linear programming
EP4222536A1 (fr) Procédé et système de traitement d'images sismiques permettant d'obtenir des surfaces d'horizon sismique pour une formation géologique
WO2019209114A1 (fr) Prévision d'une incertitude dans des inversions non linéaires de données géophysiques
Bensdorp et al. An approximate 3D inversion method for inversion of single-well induction-logging responses
US11630226B2 (en) Formation evaluation based on seismic horizon mapping with multi-scale optimization
US20240111067A1 (en) Faulted seismic horizon mapping
Taillandier et al. Refraction traveltime tomography based on adjoint state techniques
US20240103192A1 (en) Graph based multi-survey horizon optimization

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20834883

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20834883

Country of ref document: EP

Kind code of ref document: A1