Forward modeling of fracture-induced sonic anisotropy using a combination of borehole image and sonic logs

Geophysics ◽  
2007 ◽  
Vol 72 (4) ◽  
pp. E135-E147 ◽  
Author(s):  
Romain Prioul ◽  
Adam Donald ◽  
Randy Koepsell ◽  
Zakariae El Marzouki ◽  
Tom Bratton
Keyword(s):  
Order Approximation ◽  
Data Interpretation ◽  
Forward Modeling ◽  
Horizontal Stress ◽  
Depth Interval ◽  
Geologic Fractures ◽  
Induced Fractures ◽  
The Individual ◽  
Borehole Image ◽  
Sonic Logs

We develop a methodology to model and interpret borehole dipole sonic anisotropy related to the effect of geologic fractures, using a forward-modeling approach. We use a classical excess-compliance fracture model that relies on the orientation of the individual fractures, the elastic properties of the host rock, and the normal and tangential fracture-compliance parameters. Orientations of individual fractures are extracted from borehole-image log analysis. The model is validated using borehole-resistivity image and sonic logs in a gas-sand reservoir over a [Formula: see text] (50 m) vertical interval of a well. Significant amounts of sonic anisotropy are observed at three zones, with a fast-shear azimuth (FSA) exhibiting 60° of variation and slowness difference between 2% and 16%. Numerous quasivertical fractures with varying dip azimuths are identified on the image log at the locations of strong sonic anisotropy. The maximum horizontal-stress direction, given by breakouts and drilling-induced fractures, is shown not to be aligned with the strike of natural fractures. We show that using just two adjustable fracture-compliance parameters, one fornatural fractures and one for drilling-induced fractures, is an excel-lent first-order approximation to explain the fracture-induced anisotropy response over a depth interval of [Formula: see text]. Given the presence of gas and the absence of clay filling within the fractures, we assumed equal normal and tangential compliances. The two inverted normal compliances are [Formula: see text] and [Formula: see text]. Predicted FSA matches measured FSA over [Formula: see text] (40 m) of the [Formula: see text] (50 m) studied interval. Predicted slowness anisotropy matches the overall variation and measured values of anisotropy for two of the three strong anisotropy zones. Analysis of the symmetries of the modeled anisotropic response shows that the medium is mostly a horizontal transverse isotropic medium, with small azimuthal variation of the symmetry axis. Analysis of each independent fracture type shows that the anisotropy is mainly driven by open or partially healed fractures, but also consistent with stress-related, drilling-induced fractures. Therefore, the measured sonic anisotropy is caused by the combination of stress and fracture effects where the predominance of one mechanism over the other is depth-dependent. This method provides a consistent approach to data interpretation by integrating borehole image and sonic logs that probe the formation at different depths of investigation around the borehole.

Forward modeling of fracture‐induced sonic anisotropy using a combination of borehole image and sonic logs

2007 ◽  
Author(s):  
Romain Prioul ◽  
Adam Donald ◽  
Randy Koepsell ◽  
Tom Bratton ◽  
Claude Signer ◽  
...  
Keyword(s):  
Forward Modeling ◽  
Borehole Image ◽  
Sonic Logs

Periodically Unsteady Flow in an Imbedded Stage of a Multistage, Axial-Flow Turbomachine

1979 ◽  
Vol 101 (1) ◽  
pp. 42-51 ◽  
Author(s):  
J. H. Wagner ◽  
T. H. Okiishi ◽  
G. J. Holbrook
Keyword(s):  
Order Approximation ◽  
Axial Flow ◽  
Data Interpretation ◽  
Rotor Blades ◽  
Wake Flow ◽  
Field Variation ◽  
Data Sampling ◽  
Action Sequence ◽  
Average Flow ◽  
Sampling Position

A periodic-average flow measurement technique involving a hot-wire sensor was used to measure the periodically unsteady velocity field in the first stage of a low-speed, multistage, axial-flow research compressor. In portions of the compressor annulus, the periodic-average velocity patterns for imbedded rotor and stator exit flows showed appreciable sequential variation with the systematically changed data sampling position of the rotor blades. Representative examples of periodic-average flow field variation with rotor blade sampling position in stop-action sequence are shown for various locations in the compressor. A simple, first-order approximation physical description of blade wake flow transport and interaction based on experimental data interpretation is proposed to organize and thus help understand the data obtained.

FACIES CLASSIFICATION OF A COMPLEX RESERVOIR USING MACHINE LEARNING: CASE STUDY FROM VOLCANIC FORMATION, THE YURIHARA OIL FIELD, JAPAN

2021 ◽  
Author(s):  
Yuki Maehara ◽  
◽  
Takeaki Otani ◽  
Tetsuya Yamamoto ◽  
◽  
...  
Keyword(s):  
Machine Learning ◽  
Pillow Lava ◽  
Gaussian Mixture ◽  
Oil Field ◽  
Well Logs ◽  
Depth Interval ◽  
Volcanic System ◽  
Facies Classification ◽  
Volcanic Formation ◽  
Borehole Image

Lithological facies classification using well logs is essential in the reservoir characterization. The facies are manually classified from characteristic log responses derived, which is challenging and time consuming for geologically complex reservoirs due to high variation of log responses for each facies. To overcome such a challenge, machine learning (ML) is helpful to determine characteristic log responses. In this study, we classified the lithofacies by applying ML to the conventional well logs for the volcanic formation, onshore, northeast Japan. The volcanic formation of the Yurihara oil field is petrologically classified into five lithofacies: mudstone, hyaloclastite, pillow lava, sheet lava, and dolerite, with pillow lava being predominant reservoir. The former four lithofacies are the members of the volcanic system in Miocene, and dolerite randomly intruded later into those. Understanding the distribution of omnidirectional tight dykes at the well location is important for the estimation of potential near-lateral seal distribution compartmentalizing the reservoir. The facies are best classified by core data, which are unfortunately available in a limited number of wells. The conventional logs, with the help of the borehole image log, have been used for the facies classification in most of the wells. However, distinguishing dolerite from sheet lava by manual classification is very ambiguous, as they appear similar in these logs. Therefore, automated clustering of well logs with ML was attempted for the facies classification. All the available log data was audited in the target well prior to applying ML. A total of 10 well logs are available in the reservoir depth interval. To prioritize the logs for the clustering, the information of each log was first analyzed by Principal Component Analysis (PCA). The dimension of variable space was reduced from 10 to 5 using PCA. Final set of 5 variables, gamma-ray, density, formation photoelectric factor, neutron porosity, and laterolog resistivity, were used for the next clustering process. ML was applied to the selected 5 logs for automated clustering. Cross-Entropy Clustering (CEC) was first initialized using k-means++ algorithm. Multiple initialization processes were randomly conducted to find the global minimum of cost function, which automatically derived the optimized number of classes. The resulting classes were further refined by the Gaussian Mixture Model (GMM) and subsequently by the Hidden Markov Model (HMM), which takes the serial dependency of the classes between successive depths into account. Resulting 14 classes were manually merged into 5 classes referring to the lithofacies defined by the borehole image log analysis. The difference of the log responses between basaltic sheet lava and dolerite was too subtle to be captured with confidence by the conventional manual workflow, while the ML technique could successfully capture it. The result was verified by the petrological analyses on sidewall cores (SWCs) and cuttings. In this study, the automated clustering with the combination of several ML algorithms was demonstrated more efficient and reasonable facies classification. The unsupervised learning approach would provide supportive information to reveal the regional facies distribution when it is applied in the other wells, and to comprehend the dynamic behavior of the fluids in the reservoir.

Borehole image electrofacies with a comparative carbonate petrography analysis: An outcrop well study associated with reservoir application in the Ordovician Tarim Basin

2018 ◽  
Vol 6 (3) ◽  
pp. T723-T737
Author(s):  
Tao Nian ◽  
Zaixing Jiang ◽  
Hongyu Song
Keyword(s):  
Diagnostic Criteria ◽  
Carbonate Platform ◽  
Image Interpretation ◽  
Image Data ◽  
Depth Interval ◽  
Petrographic Analysis ◽  
Thin Sections ◽  
Direct Interpretation ◽  
Borehole Image ◽  
Subsurface Reservoir

Electrical borehole image logs have the potential for direct interpretation of lithofacies characteristics. The challenge is to establish a set of reliable diagnostic criteria with which electrical images can be correlated to lithofacies features such as lithology, sedimentary structures, and bedding sequences. We used the “behind-outcrop” logging procedure that can link borehole images to actual rocks and also reduce errors that are associated with core-shift process. To better reveal the correlation between borehole images and carbonate lithofacies for subsurface reservoir applications, and also make a comparative petrographic analysis with the aim of establishing diagnostic criteria for borehole images, a 200 m well was drilled in the Tarim Ordovician outcrop. A full set of borehole image data and cores with approximately 100% coring recovery rate was acquired at the same depth interval, and more than 100 stained thin sections were prepared. Electrical borehole images in wells adjacent to the outcrop were further interpreted to validate the proposed criteria. Borehole image electrofacies were established according to the image elements, such as stacked mode, bed thickness, conglomerate diameter, rim characteristics, and internal structure of bed/conglomerate, to interpret depositional/diagenetic textures and platform-slope associations. Nine image electrofacies types, corresponding to mud/wacke/pack/grain/bindstone texture, were identified and interpreted in detail. Our method reveals a set of diagnostic criteria for borehole image interpretation in carbonate platform slope, and it finally provides a powerful tool for direct interpretation of electrical images in similar reservoir environment.

Interpretation of borehole sonic measurements acquired in vertical transversely isotropic formations penetrated by vertical wells

Geophysics ◽  
2018 ◽  
Vol 83 (6) ◽  
pp. D187-D202 ◽  
Author(s):  
Elsa Maalouf ◽  
Carlos Torres-Verdín
Keyword(s):  
Transversely Isotropic ◽  
Depth Interval ◽  
Bias Error ◽  
Layer By Layer ◽  
Spatial Averaging ◽  
Vertical Wells ◽  
Frequency Dependent ◽  
Sensitivity Functions ◽  
Stiffness Coefficients ◽  
Sonic Logs

Detecting vertical transversely isotropic (VTI) formations and quantifying the magnitude of anisotropy are fundamental for describing organic mudrocks. Methods used to estimate stiffness coefficients of VTI formations often provide discontinuous or spatially averaged results over depth intervals where formation layers are thinner than the receiver aperture of acoustic tools. We have developed an inversion-based method to estimate stiffness coefficients of VTI formations that are continuous over the examined depth interval and that are mitigated for spatial averaging effects. To estimate the coefficients, we use logs of frequency-dependent compressional, Stoneley, and quadrupole/flexural modes measured with wireline or logging-while-drilling (LWD) instruments in vertical wells penetrating horizontal layers. First, we calculate the axial sensitivity functions of borehole sonic modes to stiffness coefficients; next, we use the sensitivity functions to estimate the stiffness coefficients of VTI layers sequentially from frequency-dependent borehole sonic logs. Because sonic logs exhibit spatial averaging effects, we deaverage the logs by calculating layer-by-layer slownesses of formations prior to estimating stiffness coefficients. The method is verified with synthetic models of homogeneous and thinly bedded formations constructed from field examples of organic mudrocks. Results consist of layer-by-layer estimates of [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text]. We observe three sources of error in the estimated coefficients: (1) bias error originating from deaveraging the sonic logs prior to the sequential inversion, (2) error propagated during the sequential inversion, and (3) error associated with noisy slowness logs. We found that the relative bias and uncertainty of the estimated coefficients are largest for [Formula: see text] and [Formula: see text] because borehole modes exhibit low sensitivity to these two coefficients. The main advantage of our method is that it mitigates spatial averaging effects of sonic logs, while at the same time it detects the presence of anisotropic layers and yields continuous estimations of stiffness coefficients along the depth interval of interest.

scholarly journals Combining spectral induced polarization with X-ray tomography to investigate the importance of DNAPL geometry in sand samples

Geophysics ◽  
2019 ◽  
Vol 84 (3) ◽  
pp. E173-E188 ◽  
Author(s):  
Sara Johansson ◽  
Matteo Rossi ◽  
Stephen A. Hall ◽  
Charlotte Sparrenbom ◽  
David Hagerberg ◽  
...  
Keyword(s):  
Image Analysis ◽  
Pore Space ◽  
Total Concentration ◽  
Data Interpretation ◽  
Induced Polarization ◽  
X Ray ◽  
Fine Grained ◽  
Surface Areas ◽  
Spectral Induced Polarization ◽  
The Individual

Although many studies have been performed to investigate the spectral induced polarization (SIP) response of nonaqueous phase liquid (NAPL)-contaminated soil samples, there are still many uncertainties in the interpretation of the data. A key issue is that altered pore space geometries due to the presence of a NAPL phase will change the measured IP spectra. However, without any information on the NAPL distribution in the pore space, assumptions are necessary for the SIP data interpretation. Therefore, experimental data of SIP signals directly associated with different NAPL distributions are needed. We used high-resolution X-ray tomography and 3D image processing to quantitatively assess NAPL distributions in samples of fine-grained sand containing different concentrations of tetrachloroethylene and link this to SIP measurements on the same samples. The total concentration of the sample constituents as well as the volumes of the individual NAPL blobs were calculated and used for the interpretation of the associated SIP responses. The X-ray tomography and image analysis showed that the real sample properties (porosity and NAPL distributions) differed from the targeted ones. Both contaminated samples contained less NAPL than expected from the manual sample preparation. The SIP results showed higher real conductivity and lower imaginary conductivity in the contaminated samples compared to a clean sample. This is interpreted as an effect of increased surface conductivity along interconnected NAPL blobs and decreased surface areas in the samples due to NAPL blobs larger than and enclosing grains. We conclude that the combination of SIP, X-ray tomography, and image analysis is a very promising approach to achieve a better understanding of the measured SIP responses of NAPL-contaminated samples.

scholarly journals A Functional Approach to Approximations for the Individual Risk Model

2004 ◽  
Vol 34 (2) ◽  
pp. 379-397 ◽  
Author(s):  
Susan M. Pitts
Keyword(s):  
Negative Binomial ◽  
Risk Model ◽  
Order Approximation ◽  
Functional Approach ◽  
Individual Risk ◽  
Claim Amount ◽  
Total Claim Amount ◽  
Total Claim ◽  
Binomial Approximation ◽  
The Individual

A functional approach is taken for the total claim amount distribution for the individual risk model. Various commonly used approximations for this distribution are considered, including the compound Poisson approximation, the compound binomial approximation, the compound negative binomial approximation and the normal approximation. These are shown to arise as zeroth order approximations in the functional set-up. By taking the derivative of the functional that maps the individual claim distributions onto the total claim amount distribution, new first order approximation formulae are obtained as refinements to the existing approximations. For particular choices of input, these new approximations are simple to calculate. Numerical examples, including the well-known Gerber portfolio, are considered. Corresponding approximations for stop-loss premiums are given.

Sign in / Sign up

Export Citation Format

Share Document