Sensitivity analysis of data-related factors controlling AVA simultaneous inversion of partially stacked seismic amplitude data: Application to deepwater hydrocarbon reservoirs in the central Gulf of Mexico

Geophysics ◽  
2007 ◽  
Vol 72 (1) ◽  
pp. C19-C29 ◽  
Author(s):  
Arturo Contreras ◽  
Carlos Torres-Verdín ◽  
Tim Fasnacht
Keyword(s):  
Sensitivity Analysis ◽  
Gulf Of Mexico ◽  
P Wave ◽  
Spatial Distributions ◽  
Vertical Resolution ◽  
Elastic Parameters ◽  
Related Factors ◽  
Seismic Amplitude ◽  
Amplitude Data ◽  
Seismic Amplitudes

We consider the inversion of synthetic and recorded seismic amplitude variation with angle AVA data to appraise the influence of several data-related factors that control the vertical resolution and accuracy of the estimated spatial distributions of elastic properties. We use measurements acquired in deepwater hydrocarbon reservoirs in the central Gulf of Mexico to generate synthetic seismic amplitude data and evaluate inversion results with both synthetic and recorded seismic amplitudes. Detailed sensitivity analysis of synthetic amplitude data indicates that, even in the most ideal scenario (perfectly migrated data, isotropic media, noise-free seismic amplitude data, sufficient far-angle coverage, and accurate estimates of angle-dependent wavelets and low-frequency components), input elastic models are not reconstructedaccurately by the inversion of synthetic seismic amplitudes. We attribute this result to the relatively low vertical resolution of the seismic amplitude data. P-wave impedance is the most accurate of the inverted properties, followed by S-impedance and bulk density. Additionally, sufficient far-angle coverage is crucial for the accurate estimation of 1D distributions of S-impedance and bulk density. We show that time alignment of partial-angle stacks for correcting residual NMO effects improves the vertical resolution of the estimated spatial distributions of elastic parameters and consistently decreases the data misfit. Finally, we found that the accuracy of the inverted distributions of elastic parameters is improved substantially by (1) increasing the preserved AVA information via multiple single-angle stacks, (2) correcting the P-wave velocity field used for calculating angles in partial-angle stacking, and (3) excluding far-angle data with low signal-to-noise ratios.

Detection and Spatial Delineation of Thin-Sand Sedimentary Sequences With Joint Stochastic Inversion of Well Logs and 3D Prestack Seismic Amplitude Data

2009 ◽  
Vol 13 (02) ◽  
pp. 246-264 ◽  
Author(s):  
Germán D. Merletti ◽  
Carlos Torres-Verdín
Keyword(s):  
Selection Criterion ◽  
Spatial Distributions ◽  
Vertical Resolution ◽  
Inversion Algorithm ◽  
S Wave ◽  
Stochastic Inversion ◽  
Seismic Amplitude ◽  
Wave Velocities ◽  
Sedimentary Sequences ◽  
Amplitude Data

Summary We describe the successful application of a new prestack stochastic inversion algorithm to the spatial delineation of thin reservoir units otherwise poorly defined with deterministic inversion procedures. The inversion algorithm effectively combines the high vertical resolution of wireline logs with the relatively dense horizontal coverage of 3D prestack seismic amplitude data. Multiple partial-angle stacks of seismic amplitude data provide the degrees of freedom necessary to estimate spatial distributions of lithotype and compressional-wave (P-wave) and shear-wave (S-wave) velocities in a high-resolution stratigraphic/sedimentary grid. In turn, the estimated volumes of P- and S-wave velocity permit the statistical cosimulation of lithotype-dependent spatial distributions of porosity and permeability. The new stochastic inversion algorithm maximizes a Bayesian selection criterion to populate values of lithotype and P- and S-wave velocities in the 3D simulation grid between wells. Property values are accepted by the Bayesian selection criterion only when they increase the statistical correlation between the simulated and recorded seismic amplitudes of all partial-angle stacks. Furthermore, inversion results are conditioned by the predefined measures of spatial correlation (variograms) of the unknown properties, their statistical cross correlation, and the assumed global lithotype proportions. Using field data acquired in a fluvial-deltaic sedimentary-rock sequence, we show that deterministic prestack seismic-inversion techniques fail to delineate thin reservoir units (10-15 m) penetrated by wells because of insufficient vertical resolution and low contrast of elastic properties. By comparison, the new stochastic inversion yields spatial distributions of lithotype and elastic properties with a vertical resolution between 10-15 m that accurately describe spatial trends of clinoform sedimentary sequences and their associated reservoir units. Blind-well tests and cross validation of inversion results confirm the reliability of the estimated distributions of lithotype and P- and S-wave velocities. Inversion results provide new insight to the spatial and petrophysical character of existing flow units and enable the efficient planning of primary and secondary hydrocarbon recovery operations.

AVA simultaneous inversion of partially stacked seismic amplitude data for the spatial delineation of lithology and fluid units of deepwater hydrocarbon reservoirs in the central Gulf of Mexico

Geophysics ◽  
2006 ◽  
Vol 71 (4) ◽  
pp. E41-E48 ◽  
Author(s):  
Arturo Contreras ◽  
Carlos Torres-Verdín ◽  
Tim Fasnacht
Keyword(s):  
Sensitivity Analysis ◽  
Gulf Of Mexico ◽  
Seismic Inversion ◽  
Hydrocarbon Reservoirs ◽  
Acoustic Properties ◽  
Fluid Substitution ◽  
Class Iii ◽  
Simultaneous Inversion ◽  
Seismic Amplitude ◽  
Amplitude Data

This paper describes the successful application of amplitude-versus-angle (AVA) inversion of prestack-seismic amplitude data to detect and delineate deepwater hydrocarbon reservoirs in the central Gulf of Mexico. Detailed AVA fluid/lithology sensitivity analysis was conducted to assess the nature of AVA effects in the study area based on well-log data. Standard techniques such as crossplot analysis, Biot-Gassmann fluid substitution, AVA reflectivity modeling, and numerical simulation of synthetic gathers were part of the AVA sensitivity analysis. Crossplot and Biot-Gassmann analyses indicate significant sensitivity of acoustic properties to fluid substitution. AVA reflectivity and angle-gather modeling indicate that the shale/sand interfaces represented by the top and base of the M-10 reservoir are associated with typical Class III AVA responses caused by relatively low-impedance gas-bearing sands. Consequently, prestack seismic inversion provided accurate and reliable quantitative information about the spatial distribution of lithology and fluid units within the turbidite reservoirs based on the interpretation of fluid/lithology-sensitive modulus attributes. From the integration of inversion results with analogous depositional models, the M-series reservoirs were interpreted as stacked terminal turbidite lobes within an overall fan complex. This interpretation is consistent with previous regional stratigraphic/depositional studies.

Offset‐amplitude variation and controlled‐amplitude processing

Geophysics ◽  
1985 ◽  
Vol 50 (12) ◽  
pp. 2697-2708 ◽  
Author(s):  
Gary Yu
Keyword(s):  
Seismic Waves ◽  
Gulf Coast ◽  
Seismic Profile ◽  
Angle Of Incidence ◽  
Elastic Parameters ◽  
Amplitude Analysis ◽  
Coherent Noise ◽  
Seismic Amplitude ◽  
Amplitude Data ◽  
Amplitude Anomaly

The partition of plane seismic waves at plane interfaces introduces changes in seismic amplitude which vary with angle of incidence. These amplitude variations are a function of the elastic parameters of rocks on either side of the interface. Controlled‐amplitude processing is designed to obtain the true amplitude information which is geologic in origin. The offset‐amplitude information may be successfully used to predict the fluid type in reservoir sands. Various tests were carried out on a seismic profile from the Gulf Coast. The processing comparison emphasized the effects and pitfalls of trace equalization, coherent noise, offset, and surface‐related problems. Two wells drilled at amplitude anomaly locations confirmed the prediction of hydrocarbons from offset‐amplitude analysis. Furthermore, controlled‐amplitude processing provided clues in evaluating reservoir quality, which was not evident on the conventional relative amplitude data.

Anisotropic correction of sonic logs in wells with large relative dip

Geophysics ◽  
2008 ◽  
Vol 73 (1) ◽  
pp. E1-E5 ◽  
Author(s):  
Lev Vernik
Keyword(s):  
Reservoir Characterization ◽  
Low Frequency ◽  
P Wave ◽  
Amplitude Variation With Offset ◽  
Seismic Amplitude ◽  
Avo Inversion ◽  
Pore Pressure Prediction ◽  
Siliciclastic Rocks ◽  
Dip Angle ◽  
Sonic Logs

Seismic reservoir characterization and pore-pressure prediction projects rely heavily on the accuracy and consistency of sonic logs. Sonic data acquisition in wells with large relative dip is known to suffer from anisotropic effects related to microanisotropy of shales and thin-bed laminations of sand, silt, and shale. Nonetheless, if anisotropy parameters can be related to shale content [Formula: see text] in siliciclastic rocks, then I show that it is straightforward to compute the anisotropy correction to both compressional and shear logs using [Formula: see text] and the formation relative dip angle. The resulting rotated P-wave sonic logs can be used to enhance time-depth ties, velocity to effective stress transforms, and low-frequency models necessary for prestack seismic amplitude variation with offset (AVO) inversion.

Dense surface seismic data confirm non-double-couple source mechanisms induced by hydraulic fracturing

Geophysics ◽  
2016 ◽  
Vol 81 (6) ◽  
pp. KS207-KS217 ◽  
Author(s):  
Jeremy D. Pesicek ◽  
Konrad Cieślik ◽  
Marc-André Lambert ◽  
Pedro Carrillo ◽  
Brad Birkelo
Keyword(s):  
Hydraulic Fracturing ◽  
Seismic Data ◽  
P Wave ◽  
Source Modeling ◽  
Moment Tensors ◽  
Amplitude Data ◽  
Double Couple ◽  
Source Mechanisms ◽  
Source Models ◽  
Favorable Conditions

We have determined source mechanisms for nine high-quality microseismic events induced during hydraulic fracturing of the Montney Shale in Canada. Seismic data were recorded using a dense regularly spaced grid of sensors at the surface. The design and geometry of the survey are such that the recorded P-wave amplitudes essentially map the upper focal hemisphere, allowing the source mechanism to be interpreted directly from the data. Given the inherent difficulties of computing reliable moment tensors (MTs) from high-frequency microseismic data, the surface amplitude and polarity maps provide important additional confirmation of the source mechanisms. This is especially critical when interpreting non-shear source processes, which are notoriously susceptible to artifacts due to incomplete or inaccurate source modeling. We have found that most of the nine events contain significant non-double-couple (DC) components, as evident in the surface amplitude data and the resulting MT models. Furthermore, we found that source models that are constrained to be purely shear do not explain the data for most events. Thus, even though non-DC components of MTs can often be attributed to modeling artifacts, we argue that they are required by the data in some cases, and can be reliably computed and confidently interpreted under favorable conditions.

Using relative geologic time to constrain CNN-based seismic interpretation and property estimation

Geophysics ◽  
2021 ◽  
pp. 1-44
Author(s):  
Aria Abubakar ◽  
Haibin Di ◽  
Zhun Li
Keyword(s):  
Large Scale ◽  
Three Dimensional ◽  
Seismic Interpretation ◽  
Seismic Survey ◽  
Geologic Time ◽  
Seismic Amplitude ◽  
Property Estimation ◽  
Constrained Learning ◽  
Seismic Amplitudes ◽  
Multiplicative Regularization

Three-dimensional seismic interpretation and property estimation is essential to subsurface mapping and characterization, in which machine learning, particularly supervised convolutional neural network (CNN) has been extensively implemented for improved efficiency and accuracy in the past years. In most seismic applications, however, the amount of available expert annotations is often limited, which raises the risk of overfitting a CNN particularly when only seismic amplitudes are used for learning. In such a case, the trained CNN would have poor generalization capability, causing the interpretation and property results of obvious artifacts, limited lateral consistency and thus restricted application to following interpretation/modeling procedures. This study proposes addressing such an issue by using relative geologic time (RGT), which explicitly preserves the large-scale continuity of seismic patterns, to constrain a seismic interpretation and/or property estimation CNN. Such constrained learning is enforced in twofold: (1) from the perspective of input, the RGT is used as an additional feature channel besides seismic amplitude; and more innovatively (2) the CNN has two output branches, with one for matching the target interpretation or properties and the other for reconstructing the RGT. In addition is the use of multiplicative regularization to facilitate the simultaneous minimization of the target-matching loss and the RGT-reconstruction loss. The performance of such an RGT-constrained CNN is validated by two examples, including facies identification in the Parihaka dataset and property estimation in the F3 Netherlands dataset. Compared to those purely from seismic amplitudes, both the facies and property predictions with using the proposed RGT constraint demonstrate significantly reduced artifacts and improved lateral consistency throughout a seismic survey.

scholarly journals Life-Cycle Break-Even Analysis of Electric Carsharing: A Comparative Study in China

2020 ◽  
Vol 12 (16) ◽  
pp. 6584
Author(s):  
Jingjing Jia ◽  
Shujie Ma ◽  
Yixi Xue ◽  
Deyang Kong
Keyword(s):  
Sensitivity Analysis ◽  
Life Cycle ◽  
Life Cycle Cost ◽  
Net Present Value ◽  
Market Price ◽  
Critical Factor ◽  
Field Investigation ◽  
Related Factors ◽  
Business To Consumer ◽  
Influential Variable

Electric carsharing (ECS) is a potential option to address the problem of unsustainability in the transportation sector. The business-to-consumer model of ECS, which is one of several different electric carsharing models, has gained much popularity in recent years. Generating sufficient revenue to cover costs is a critical factor for ECS companies to maintain healthy development. This study makes an economic analysis, on the basis of life-cycle cost and monetary revenue associated with the operation of ECS, of two Chinese ECS companies: EVCARD and LCCS. Based on data gathered by field investigation, this study aims to determine the break-even moment for each company’s main vehicle models by means of the net present value method. The results show that EVCARD achieved an earlier break-even moment than LCCS. The break-even moment of Chery eQ of EVCARD was the shortest of all the vehicle models, at only 181.3 min. Moreover, a sensitivity analysis was conducted to portray how different cost-related and revenue-related factors influence the break-even moment. Our findings indicate that a wide difference exists in terms of the influence of different factors on the break-even moment. Among these, the manufacturer’s suggested retail price is the most influential variable, followed by the unit rental price. The reaction of the break-even moment to the market price of a charging pile and the non-rental revenue per vehicle—especially the latter—was found to be negligible in the sensitivity analysis.

Sign in / Sign up

Export Citation Format

Share Document