Produção Científica



Apresentação
07/11/2019

Full-waveform inversion using alternative objective functions in the presence of noise and uncertainties of source signature
Full-waveform inversion (FWI) is a method able to estimate model parameters in subsurface from seismic data. The algorithm consists on the minimization of an objective function that relates observed seismic data and synthetic data for the estimated model. Thus, the successful application of FWI depends on the accurate correspondence between modeled and field data.

Some problems that appear in the application of FWI to field data are dealt by using alternative functionals for objective function. The results, obtained by using synthetic data, show that the definition based on the least absolute value norm and the cross-correlation error measure have advantages over least squares norm, turning the method more robust when facing noisy data and uncertainties in source signature used for inversion.

The quasi-Newton formulation of L-BFGS method is applied to non-quadratic objective functions without affecting, in practice, the convergence of inversion scheme.
Apresentação
07/11/2019

Otimização automática do custo de processamento de programas SPITS na AWS
Plataformas de serviço de computação em nuvem oferecem uma ampla variedade de recursos computacionais que possuem características de desempenho com custos diferenciados. Neste trabalho, investigamos como as instancias Spot e as zonas de disponibilidade da Amazon Web Services (AWS) podem ser utilizadas para a redução do custo de processamento. Em adição propomos um algoritmo de gerenciamento automático de instâncias na AWS para otimizar custo na execução de programas implementados sobre o modelo de programação Scalable Partially Idempotent Task System (SPITS). Os resultados obtidos indicam que o método proposto é capaz de identificar para ajustar dinamicamente os tipos de máquinas virtuais que oferecem o melhor custo-benefício.
Apresentação
07/11/2019

Denoising of the groundroll using Wavelets Thresholding
In this work the method is based on the spectral decomposition of the seismic data using Wavelet Transforms (WT) to attenuate the groundroll. The WT is used to represent a signal at different resolutions and various time and frequency contents. The filtering method separates the signal by frequency range and the band of coefficients related to the noise, then signal-to-noise ratio of the seismic data is improved by attenuating the groundroll. The signal filtering is done by estimating a cutoff, threshold (l), for signal amplitudes which are related to noise. The determination of l takes into account the standard deviation and the size of the signal sample. In order to test the efficiency of the method in question, the land seismic line (RL5090) of the Tacutu basin was processed and good results were obtained.
Apresentação
07/11/2019

Inversão cinemática em meios TI usando uma nova aproximação do tempo de reflexão
One of the challenges of traveltime inversion in anisotropic media is the fact that the equations are nonlinear. We present a simple approach to estimate the model parameters in a medium with TI - transverse isotropy. The proposed approach uses the Levenberg-Marquardt (LM) inversion method and shows a better performance in the inversion and in the nmo correction of the seismic section. It was observed that the precision in the estimation of the anisotropy parameter ï¨ with the proposed approximation depends not only on the correct extraction of the transit times of the data, but also on the value of (ï¤ï¥ï€­) and the layer inclination angle of inclination.
Apresentação
07/11/2019

Poroelastic modeling in stratified media: Biot-JKD equations
We represent an exact mathematical procedure for the analysis of the elastic wave propagation in plane layered porous media taking into account the effect of high temporal frequencies. For the characterization of the effect we use the dynamic permeability expression proposed by Jonhson, Koplik and Dashen in 1987. The algorithm is based on a formalism introduced by Ursin in 1983.
Apresentação
07/11/2019

Microporosity analysis from petrographic thin-sections and elastic modulus estimative through DEM model in outcrop carbonate rocks
Carbonate reservoirs present a complex pore system characterized by their heterogeneity. It affects directly the rock's elastic properties. The external pressure's variation may cause pore volume reductions and affect the propagation velocity of elastics waves in sedimentary rocks. Thus, this paper aims to study the velocity variation of compressional and shear waves and the porosity in carbonate rock samples. By the data analysis from a number of samples, it intends to empirical models (Differential effective medium model – DEM) that relate variations in elastic velocity with changes in porosity.
Apresentação
07/11/2019

Synthesis and petrophysical characterization of an artificial mudstone analogous
Carbonate rocks are important for rock properties research because they are linked to significant oil and gas reserves around the world. The goal of synthesizing carbonate rocks in laboratory is to simulate the natural carbonate rock matrix regarding the main factors of lithification such as grain size and shape, concentration of cementing material and compaction pressure, and allowing the production of rock specimens that can be used for tests in extreme conditions in replacement of the high cost natural cores. The reproduction of rocks in laboratory also allows to access samples with predetermined characteristics enabling a better understanding of the relationship between their physical properties. In this work, samples of synthetic carbonate rocks were made by mixing a fixed amount of calcite and sand, and varying the amount of cement material (Portland cement and water). The behavior of the main petrophysical characteristics of these samples was investigated, such as porosity, permeability and P- and S-wave velocities versus the amount of cementing material. The results were compared with natural carbonate samples and showed a high similarity to the petrophysical behavior of natural mudstones rocks.
Apresentação
07/11/2019

Using the Benford's Law to analyze the sonic reflectivity
The caliper log can be sensitivity to borehole enlargement due drilling effect on different rock formations. The density of the lithology is affected by mud infiltration and the consequence is unrealistic values on the well-to-seismictie. In this work we use, in the first time, the statistical Benford's Law (BL) to analyze and check on the existence of errors or inconsistencies on the reflectivity log. The Benford's analysis is applied on the reflectivity series of two wells of the Viking Graben field. Effect of sonic and density logs despiking as well as the size of sample window to find the BL distribution were analyzed. The results on the real dataset analyzed shows that it is possible to find a optimum coefficient for the despiking process on the well logs according to the Benford's Law and that the segments of the reflectivity where the caliper log is stable the conformity with the Benford's Law is higher.
Apresentação
07/11/2019

Comparative Analysis of First Break Picking Methods in Noisy and Non-noisy P-wave Signal
The estimative of acoustic velocities is an important data that is extracted from data of exploration seismic, well logging (acoustic logging) and laboratory experiments. To realize this estimative a technique called first break picking is used, which consist in an analysis of time x Amplitude data to identify the wave arrival. The complexity of this analysis is in the fact of the data, generally is contaminated with noise and this one could be either inherent to the measuring equipment or other factors, that could make the simple visual analysis impossible. Considering this problem, the development of techniques which would be capable to identify the wave arrival automatic, accurate and faster, justify the study of advanced methods to accomplish this task.
Apresentação
07/11/2019

Differences between static and dynamic methods for pore compressibility in carbonates
The production of hydrocarbons from a reservoir occurs due to the variation of pore volume through a pressure differential. These variations are quantified by pore compressibility, an important parameter that interferes with the flow of fluids that saturate the porous medium, porosity and elastic properties of the rock. For many times pore compressibility was considered a constant, but the oil industry with the prospect and need to potentiate the reservoir's productive capacity has been attributed as an important factor in reserve calculations and decision-making related to well design in order to avoid unexpected interventions and loss of the reservoir by collapse. The compressibility of reservoir rock pores is determined through well profiles and seismic data, however for more accurate results, laboratory measurements are crucial. In this context, the objective of this work is to obtain pore compressibility through static and dynamic tests in samples of carbonates highlighting the influence of each method on the difference of the results. The static method was performed from uniaxial compression tests based on the technique of Unalmiser & Swalwell (1993). Dynamic tests were based on pore-elasticity theory through the propagation velocity of compression waves (P) and shear (S) and elastic constants of the rock when subjected to hydrostatic pressure variation. The correlation between the two methods is economically important for the industry as well as serving as the basis for the calibration process of the logging tools.
<<  <   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65  66  67   >  >>