Produção Científica

**Apresentação**

Analysis of Eshelby-Cheng's and Hudson on synthetic cracked modelsPhysical modeling of cracked/fractured media through downscaled laboratory experiments has worked as a great alternative for understanding the effect of anisotropy in the hydrocarbon reservoir, and in the crustal and mantle seismology. The main goal of this work was to experimentally verify the predictions of effective elastic parameters in anisotropic cracked media predicted by Hudson and Eshelby-Cheng's effective cracked models. For this proposal, we carried out ultrasonic measurements on synthetic anisotropic samples with low crack density and different aspect ratios. Six samples were prepared with 5% of crack density. These samples, with three-different aspect ratio cracks (0.13, 0.17 and 0.26), were simulated by penny-shape rubber inclusions in a homogeneous isotropic matrix made with epoxy resin. Moreover, an isotropic sample for reference was constructed with epoxy resin only. Among all samples, three presented only one aspect ratio type (samples with single crack aspect ratio), while other three showed three types of different aspect ratio (mixed crack aspect ratio samples). Regarding predictions performed by the theoretical models, Eshelby-Cheng shows a better fit when compared to the experimental results for samples with single and mix crack aspect ratio. Our comparisons were also performed in terms of γ parameter (Thomsen parameters). |

**Apresentação**

Seismic data inversion with curvelet denoising preconditioningSeismic inversion methods are highly sensitive to noise present in the data set. The need to enhance the signal-to-noise ratio (SNR) motivates many researchers do develop increasingly sophisticated denoising methods and combine them into other techniques. While some methodologies operate on a single scale, the curvelet transform established itself as multi-scale transform useful to decompose the seismic signals into multi-resolution elements. In this study, we evaluate the benefits of curvelet denoising as a preconditioning method for poststack seismic data in a 2D acoustic inversion process using a Bayesian framework. Our tests on a synthetic data set simulated in the Marmousi model, and a real data set from the Campos Basin offshore Brazil have shown that the curvelet thresholding method can be successfully applied for random noise elimination. Even the use of a hard global threshold might allow improvements in the deepest parts. Future work will have to show whether alternatives that ensure a more robust way of selecting the coefficients can take into account the wavelength change with depth. |

**Apresentação**

Mechanical Hysteresis of Carbonate Outcrop SamplesReservoir rocks characterization consists of a critical stage on petroleum reserves evaluation in which determining the relationships between petrophysical and mechanical rock properties is required to guarantee the accuracy of data. Mechanical properties affect both drilling programs and hydrocarbon exploitation. Hydrostatic compression tests on rocks, if conducted at stresses below failure, typically exhibit both non-linearity and hysteresis in the stress–strain curve. This behavior can be associated with the presence of cracks and pores. In this study, porosity and bulk volume deformation were analyzed and the different behavior under loading/ unloading conditions reveals the grain/pore interaction influence on these properties. Experiments were conducted increasing hydrostatic pressure from 400 psi to 1000 psi, registering loading and unloading data, on carbonate samples of American outcrops. The dissipated energy related to frictional sliding and adhesion effect, during confinement cycle, was estimated by the area of hysteresis loop in the effective pressure versus bulk volume crossplot. |

**Apresentação**

A comparison of iterative methods performance to exact and pseudo adjoint operators in least-squares migrationLeast-squares migration is often used to attenuate migration artifacts that arise in conventional migration due to, for instance, data incompleteness and limited recording aperture. It uses iterative methods to obtain a model that best fits the data, and it requires a forward/adjoint operator pair to do so. These methods use the gradient of the cost function to estimate search directions and update the current model. The gradient is obtained by migrating the residuals between estimated and observed data at each iteration. Migration operators are regarded as adjoint of forward operators but only those which pass the dot-product test are exact adjoints. These operators estimate more accurate search directions, enhancing the convergence rates of iterative methods. We test the performance of adjoint and pseudo-adjoint operators in LSM based on three different iterative methods (steepest descent, conjugate gradients and limited-memory BFGS) in order to assess their sensitivity to the adjointness of the migration operator. We then compare the performance of each method. |

**Apresentação**

Seismic trace analysis using minimum phase and singular value decomposition methods. Application to ground-roll attenuationThe spiking deconvolution aim is to correct the effect of the wavelet in seismic trace by apply an inverse filter, assumed to be minimum delay, and to get an estimate of reflectivity. In order to compensate propagation and absorption effects one way is to use a time-variant deconvolution where a different inverse filter is computed and applied for each output sample position. We modify this procedure by estimating a minimum-delay wavelet for each time-sample position of the seismic trace. This gives a decomposition of the seismic trace as a sum of minimum-delay wavelets, each multiplied by a reflectivity coefficient. We done the SVD decomposition through moving windows in the matrix which contains the minimum phase wavelets in each column. The seismic trace can be represented as a sum of the eigenimages of the wavelet matrix multiplied by the reflective function. In this way we can obtain an estimate of the reflectivity function through the inversion of this system. This seismic signal decomposition method has a range of applications in the data processing and interpretation of the seismic signal. We use this method to ground-roll attenuation present in land seismic data. Improvements in data quality are evident in prestack data panels and velocity analysis. |

**Apresentação**

Weighted AB semblance using very fast simulated annealingWe present a modification applied to a work presentend by Ebrahimi et. al, 2017, that shows the weighted AB semblance method which deals with the AVO phenomenon in velocity spectra. This method is based on the use of two weighting functions applied to the AB semblance coherence measure. The first weighting function, uses the ratio between the first and the second singular value of the time window. The second weighting function is based on the position of the seismic wavelet in the time window. Each weighting function depends on two empirical coefficients, determined by the analysis of a matrix which contains the measure called ECM (Energy Concentration af Matrix). In this work, we made a modification in the calculus of these coefficients values. We use the method Very Fast Simulated Annealing to get the best coefficient values that generates a spectrum with the best resolution. We applied the method in both synthetic and real CMP gathers. Results show that the VFSA made easier to obtain the coefficient values. These values allow to generate velocity spectra with excelent resolution. |

**Apresentação**

Verificação experimental do modelo de efetivo de Hudson-Crampin para meios anisotrópicos fissurados cujo o background apresenta anisotropia VTIA modelagem física de meios fissurados/ fraturados utilizando experimentos laboratoriais em escala reduzida reduzida tem funcionado como uma excelente alternativa para a compreensão do efeito da anisotropia na caracterização do reservatório de hidrocarbonetos fissurado/fraturado. O objetivo principal deste trabalho foi verificar experimentalmente asprevisão do modelo de meio efetivo para meios fissurados Hudson e Crampim. Para isso, foram realizadas medições ultrassônicas em amostras anisotrópicas sintéticas com diferentes densidade de fissuras distribuídas em quatro razões de aspectos. Dezesseis amostras (quatro amostra para cada razão de aspecto) fissuradas e uma amostra de referência com ansiostropia VTI e sem fissuras. As razões de aspectos variam entre Isso resulta em amostras com valores de razão de aspecto variando de 0.08 a 0.52. Todas as amostras fissuradas foram simuladas por espaços vazios na forma de discos numa matriz poro elástica com baixa anisotropia VTI homogênea feita com areia e cimento. Em relação às previsões de velocidade realizadas pelo modelos teórico, Crampin apresenta um melhor ajuste para VP e VS se propagando perpendicular ao plano de fissuras para as condições secas e saturadas. Estas previsões se tornam mais evidentes no caso em que a densidade de fissuras e razões de aspectos são baixas. A partir dos valores de velocidade, nossa comparação também foi realizada em termos dos parâmetros ε, γ e δ. |

**Apresentação**

Estimativa automática de tempo de trânsiotode ondas compressionais: aplicação na determinação do parâmetro de ThomsenO trabalho tem como objetivo estimar o parâmetro elástico ε, que é conhecido como “anisotropia da onda P”. Para estimar este parâmetro, é utilizado um algoritmo que fornece o tempo de chegada da onda P a partir da teoria STA/LTA. As amostras utilizadas no trabalho foram feitas com uma mistura de cimento e areia, tem formato cilíndrico, com diâmetro de 38 milímetros alturas que variam entre 51 a 61 milímetros e inclusões cilíndricas, as amostras simulam um meio VTI de anisotropia fraca. |

**Apresentação**

Least Squares Imaging using l1 and Student residualsThis paper studies the Least Squares Migration (LSM) procedure as an optimization problem. Particularly, we study the behavior of this imaging procedure upon the use of robust loss functions. Our results demonstrate that LSM can work even under heavy noise if a suitable loss function is employed. Both the ℓ1 norm and the Student's t-norm showed themselves more stable with regard to outliers than the ℓ2 norm conventionally used in LSM. In the optimization step, we employ the hybrid deterministic-stochastic L-BFGS algorithm that exploits the structure of the objective function to reduce the computational burden of this imaging procedure. The quality of the achieved results show no deterioration over the conventional L-BFGS algorithm. |

**Apresentação**

AVO analysis using an alternative formulation for estimation of seismic properties of hydrocarbon fluids: A study on gas condensateThe equations proposed by Batzle and Wang (1992), referred as BW formulation, have been largely used by the Geoscience community since the early 90’s. These equations can be applied for the determination of fluid properties in a porous medium and the computation of the compressional and shear wave velocities within the fluid. This information is valuable for the application of Gasmmann fluid substitution and seismic reservoir characterization studies. Carvalho and Moraes (2018) compile an alternative formulation based on the equations proposed by Dranchunk and Abou-Kassem (1975) and Sutton and Hamman (2009) (referred as DASH formulation), which yields superior fit to experimental data of acoustic velocity in natural gas fluids collected from literature (Lemmon et al. 2018; Younglove et al. 1993; Plantier et al. 2005; Ahmadi et al. 2017). The present work expands this study, analyzing the differences in AVO analysis in a 2-layer model for fluid properties calculated with the use of BW and DASH formulations |