We carry out the inversion of marine controlled-source electromagnetic data using real coded genetic algorithm to estimate the isotropic resistivity. Unlike linearized inversion methods, genetic algorithms belonging to class of stochastic methods are not limited by the requirement of the good starting models. The objective function to be optimized contains data misfit and model roughness. The regularization weight is used as a temperature like annealing parameter. This inversion is cast into a Bayesian framework where the prior distribution of the model parameters is combined with the physics of the forward problem to estimate the aposteriori probability density function in the model space. The probability distribution derived with this approach can be used to quantify the uncertainty in the estimation of vertical resistivity profile. We apply our inversion scheme on three synthetic data sets generated from horizontally stratified earth models. For all cases, our inversion estimated the resistivity to a reasonable accuracy. The results obtained from this inversion can serve as starting models for linearized/higher dimensional inversion.
Presentation Date: Monday, October 15, 2018
Start Time: 1:50:00 PM
Location: Poster Station 13
Presentation Type: Poster
Alvarez, Pedro (Rock Solid Images) | Marin, William (Rock Solid Images) | Berrizbeitia, Juan (Rock Solid Images) | Newton, Paola (Rock Solid Images) | Bolivar, Francisco (Rock Solid Images) | Barrett, Michael (African Petroleum Corporation) | Wood, Harry (African Petroleum Corporation)
We have evaluated a case study, in which a class-1 amplitude variation with offset (AVO) turbiditic system located offshore Cote d’Ivoire, West Africa, is characterized in terms of rock properties (lithology, porosity, and fluid content) and stratigraphic elements using well-log and prestack seismic data. The methodology applied involves (1) the conditioning and modeling of well-log data to several plausible geologic scenarios at the prospect location, (2) the conditioning and inversion of prestack seismic data for P- and S-wave impedance estimation, and (3) the quantitative estimation of rock property volumes and their geologic interpretation. The approaches used for the quantitative interpretation of these rock properties were the multiattribute rotation scheme for lithology and porosity characterization and a Bayesian litho-fluid facies classification (statistical rock physics) for a probabilistic evaluation of fluid content. The result indicates how the application and integration of these different AVO- and rock-physics-based reservoir characterization workflows help us to understand key geologic stratigraphic elements of the architecture of the turbidite system and its static petrophysical characteristics (e.g., lithology, porosity, and net sand thickness). Furthermore, we found out how to quantify and interpret the risk related to the probability of finding hydrocarbon in a class-1 AVO setting using seismically derived elastic attributes, which are characterized by having a small level of sensitivity to changes in fluid saturation
Presentation Date: Tuesday, October 16, 2018
Start Time: 1:50:00 PM
Location: 209A (Anaheim Convention Center)
Presentation Type: Oral
Summary This paper presents a workflow to estimate brittleness, porosity, and total organic carbon from elastic attributes. The estimation is carried out through the application of the multi-attribute rotation scheme. This method is a hybrid rock-physics/statistical approach that uses a global search algorithm to estimate a customized transform for each geologic setting in order to predict petrophysical properties from elastic attributes. After the application of this technique, customized transforms were derived for the analyzed geological setting, to estimate porosity, brittleness, total organic carbon and litho facies logs from elastic attributes. The final goal of this workflow is to apply these transforms over seismically-derived attributes to generate volumes of these petrophysical properties that can be used for reservoir characterization and production optimization.
We carry out inversion of the marine controlled-source electromagnetic data using genetic algorithm to estimate the subsurface vertical resistivity. This inversion is cast into a Bayesian framework where the prior distribution of the model parameters is combined with the physics of the forward problem to estimate the a-posteriori probability density function in the model space. The probability distribution derived with this approach can be used to quantify the uncertainty in the estimation of vertical resistivity profile. We apply our inversion scheme on two synthetic data sets generated from two different horizontally stratified earth models. The first model had one thin resistive hydrocarbon layer between the low-resistive sediments, whereas the second model had multiple thin resistive layers. For both cases, our inversion estimated the resistivity to a reasonable accuracy. Additionally, we tested our method to invert the multi-frequency data which further improved the quality of the inverted results. The results obtained from this inversion can form a basis for higher dimensional modelling and inversions. Also, this method can be easily extended to implement the joint inversion using seismic data.
Presentation Date: Wednesday, September 27, 2017
Start Time: 2:40 PM
Location: Exhibit Hall C/D
Presentation Type: POSTER
The multi-attribute rotation scheme (MARS) is a methodology that uses a numerical solution to estimate a transform to estimate petrophysical properties from elastic attributes. This is achieved by estimating a new attribute in the direction of maximum change of a target property in an n-dimensional Euclidian space formed by n attributes, and subsequent scaling of this attribute to the target unit properties. This approach is performed using well log-derived elastic attributes and petrophysical properties, and posteriorly applied over seismically-derived elastic attributes. In this study MARS was applied to predict a transform to estimate water saturation and total porosity from elastic attributes, using a two- and three-dimensional approach, respectively. The final goal of this workflow is to apply these transforms over seismically-derived attributes to generate volumes of these properties, which can be used in exploration and production settings for reservoir characterization and delineation, as well as soft variables in geostatistical workflows for static model generation and reserve estimation.
A common way to understand the relationship between seismic attributes and a petrophysical property is by the use of rock physics templates or simply by cross-plotting well log derived elastic attributes color-coded by a petrophysical property. Both ways graphically illustrate the relationship between the elastic and petrophysical domains, which can be used to estimate reservoir properties from seismic inversion attributes. MARS is a methodology that uses a numerical solution to estimate a mathematical expression that reproduces the aforementioned phenomena. This methodology uses, as input, measured and/or rock physics-modelled well log information, to estimate a well log-derived transform between several elastic attributes and target petrophysical properties. The objective of this workflow is to apply the resultant transform over seismically-derived elastic attributes to predict the spatial distribution of petrophysical reservoir properties.
Theory & Method
MARS estimates a new attribute (τ) in the direction of maximum change of a target property in a n-dimensional Euclidian space formed by n attributes. We search for the maximum correlation between the target property and all the possible attributes that can be estimated via axis rotation of the basis that forms the aforementioned space.
Summary An integrated study of the well Zhao-104 and surrounding seismic volume within the shale gas reservoir in South China has been conducted with the objective of generating shale formation properties related to fracture orientation and intensity in the area and deriving such reservoir rock properties as data quality allows. Seismic attribute analysis of anisotropy from elliptical velocity inversion indicates that anisotropy varies horizontally and vertically, and that it is dominantly controlled by stress azimuth, which conforms to the current day stress field as independently determined from borehole breakouts. For the reservoir, it appears that the modern-day SH (N40E) orientation approximates the conjugate fracture orientation of a wrench-faulted tectonic regime; this map pattern suggests a clockwise net rotation of the stress field from time of deposition to the present-day by 40 . Very large strike-slip faults (cutting the survey) have low anisotropy. Intermediate strike-slip faults cutting the entire shale section may exhibit larger anisotropy.
Summary Changes in rock properties, fluid content and their combination can be related to multiple seismic and EM anomalies. Rock physics models are commonly used to better understand the underlying physics of observed log responses and how they are governed by local petrophysical properties (Smith, 2011). In this study from the Gulf of Mexico, we present a case that shows a strong EM anomaly that can be tied to the rock properties at the wellbore location by generating 1D CSEM models. This multi-physics approach addresses the importance of understanding of rock physics fundamentals and their integration with other technologies, such as seismic and EM. Introduction This feasibility study aims to understand the impact of elastic and electric reservoir properties using the discovery well as the calibration well for such anomaly response in a Miocene reservoir.
If a seismic trace can be represented by a sparse model of the underlying reflection series then adjacent reflections can be resolved to a greater extent than might otherwise be possible, and in particular to a greater extent than is possible after wavelet or deconvolution processing aimed at amplifying the high frequency content of the data. The enhanced resolution can be used to derive a frequency-extended version of the input trace that contains higher useable frequencies than those obtainable by wavelet or deconvolution processing.
We show examples of a form of blind deconvolution that uses an orthogonal matching pursuits algorithm to derive a sparse reflectivity estimate as the basis for a frequency-extended version of the trace. We discuss the dependence of the output resolution on the reflection sparseness and the input bandwidth and establish some approximate limits for the degree of resolution enhancement that is available in practice. These limits need to be acknowledged in the frequency extension approach if the result is to remain a valid representation of the underlying reflectivity.