Acoustic Velocity Model Research Articles

Spectral recomposition feature for optimizing seismic velocity model prediction with a neural network

An approach incorporating spectral recomposition results for training a neural network is proposed. The initial procedure is based on an inversion scheme for reconstructing the seismic spectrum of wavelets associated with a reflection event. This allows one to estimate each wavelet's temporal position within a seismogram. Since each calculated wavelet's temporal position is identified, this information is integrated into the seismogram as an additional feature for a trained neural network. A Fully Convolutional Network is trained with models that have features commonly found in offshore hydrocarbon reservoirs. The training step is based on composing a set of an acoustic velocity model set with its seismograms and their position matrices at the time of each wavelet for each seismogram. Next, the network training with a specific number of sets is performed. This approach was able to predict acoustic velocity models more accurately, especially for predicting the shape of salt bodies and the interface of inclined layers, which makes the proposed approach an interesting mean for estimating specific features in offshore hydrocarbon reservoir structures.

Full waveform inversion based on inversion network reparameterized velocity

AbstractSeismic velocity plays an important role in imaging and identifying underground geology. Conventional seismic velocity inversion methods, like full waveform inversion, directly update the velocity model based on the misfit between the observed and synthetic data. However, seismic velocity inversion is a highly nonlinear process, and the inversion effect greatly relies on the initial inversion model. In this paper, we propose a novel network‐domain full waveform inversion method. Different from the existing network‐domain full waveform inversion methods, which use random or fixed numbers as network input, we reparameterize the low‐dimensional acoustic velocity model in a high‐dimensional inversion network parameter domain with seismic observed data as the network input. In this way, the physical information within the observed data can be directly encoded into the inversion parameters, leading to a better inversion effect than the current network‐domain full waveform inversion method. Moreover, comparison experiments on the Society of Exploration Geophysicists and the European Association of Geoscientists and Engineers Overthrust model and the Marmousi model show the advantages of the proposed method over conventional full waveform inversion from the aspects of inversion accuracy, robustness to noisy data, and more complex geological structures. These advantages may benefit from the fact that reparameterization within the inversion network domain can empower the inversion process with the regularization ability of denoising and mitigating the cycle‐skipping issue. In the end, the potential of the proposed method in terms of network initialization is further discussed.

SEISMIC-DERIVED ROCK TRUE RESISTIVITY (Rt ) REVISITED. PART II: REFORMULATION USING WYLLIE’S TIME-AVERAGE MODEL

This paper is the second of a three-part presentation. As highlighted in the previous paper (Part I, Widarsono Mendrofa, 2006), the main objective of the study is to re-evaluate the potential of acoustic impedance as a source of resistivity data. This essentially came from the very idea of extracting information of resistivity (Rt ), data that plays a very important role in the determination of water saturation in reservoir, from seismic-derived acoustic impedance (AI).As observed in the past view years, there have been a lot of efforts devoted to the extraction of water saturation information from seismic. However, as Widarsono Mendrofa (2006) put it, most of the efforts were mainly based on pattern recognition activities with little attention was given to the theoretical aspects of relationships between seismic signals and water saturation. The work reported in this threepart presentation is concentrated more as re-establishing (a reformulation of works reported in Widarsono Saptono, 2003; 2004) the theoretical relationship between resistivity and acoustic impedance.In the Part I (Widarsono Mendrofa, 2006), a reformulation between the classical Gassmann acoustic velocity model and shally sand models of Modified Simandoux and Hossin is presented. In the reformulation, a new resistivity function of acoustic impedance has been established. In principle, whenever acoustic impedance data from seismic has been made available resistivity data for the determination of fluid saturation can be estimated.Despite the theoretical correctness of the resistivity function presented in the Part I, practicallity is not the function’s best aspect. In other words, the resistivity function is not an easy one to be used practically. Various parameters (e.g. matrix moduli) have to be assumed, since the data cannot easily measured even in the laboratory. This is indeed the main reason why gassmann model, and others such as Biot, has not been used much in day-to-day practices such as log interpretation for porosity determination.Being aware of such difficulties, in 1954 M.R.J. Wyllie et al proposed their “time average” model (named after its proportional averaging of pore fluid, rock matrix, and shale transit time values to represent transit time of a fluid-filled porous medium) for any practical uses related to P-wave velocity in porous media. Due to its simplicity, the model, as well as its subsequent modifications, has been used extensively since then in some areas especially in log analysis for porosity determination. Considering this simplicity aspect, this three-part study also adopted Wyllie “time average” model into its reformulation works. This Part II paper presents the formulation using Wyllie and the two shally sand models following the same manner that was adopted and presented in the Part I paper.Summarily, the objectives of the works presented in this paper are:- To establish a model/method to obtain formation rock true resistivity (Rt ) from seismic-derived acoustic impedance (AI),- To provide correction/modification onto previous works reported in Widarsono Saptono (2003, 2004), and- To provide a simpler alternative to the resistivity function yielded from the reformulation works presented in Part I paper (Widarsono Mendrofa, 2006)

Towards a more robust input for stereotomography

ABSTRACTStereotomography is a slope‐based tomographic method that uses seismic reflection data, consisting of traveltimes and slopes, to estimate the subsurface velocity macro‐model. Here, we interpret the macro‐model as the large‐scale components of the acoustic velocity model of the subsurface. Tomographic methods are often solved with local optimization techniques, which are dependent on the initial input data. In that sense, selecting the correct input data by estimating and picking slopes is a crucial part of the stereotomography problem. In this paper, we address the problem of slope estimation and picking in stereotomography. Our central proposal uses the attributes obtained from the common‐offset common‐reflection‐surface method as input data for stereotomography. We use the global optimization method known as differential evolution to estimate these attributes, resulting in gathers of attributes estimated for the complete seismic data. This strategy presents reliable estimates when the seismic data are highly corrupted by random noise or present geological features that may lead to the failure of classical methods. We also propose an automatic picking strategy to extract the traveltimes and slopes from the common‐offset common‐reflection‐surface attributes gathers. We illustrate with synthetic examples the benefits of using the proposed framework for obtaining input data for stereotomography.

Construction of the Central Arctic Sea Ice Structure and Acoustic Velocity Model at the Short-Term Ice Station During N11 CHINARE

Construction of the Central Arctic Sea Ice Structure and Acoustic Velocity Model at the Short-Term Ice Station During N11 CHINARE

Wave-equation migration velocity analysis in the common-offset domain to avoid cycle-skipping for FWI

Full waveform inversion (FWI) has been used successfully to build high-resolution acoustic velocity models for imaging but can fail when the starting model is far from accurate because the absence of low frequency information can lead to conversion to local minima. Wave-equation based migration velocity analysis (WEMVA), is an iterative method that has successfully been used to provide low-frequency starting models for full waveform inversion (FWI) and avoid cycle-skipping. In this work, we extend WEMVA to the common-offset domain by minimizing an image residual in the common-offset domain, which can be formed after either reverse-time migration (RTM) or least-squares migration (LSM). Although very expensive in comparison to RTM, LSM has an increased amplitude fidelity that helps to ensure that the image residual used in WEMVA is largely a function of moveout instead of errors in amplitude, thus increasing the likelihood of converging to a global minimum. We demonstrate the method, theory and results on three synthetic models of increasing complexity.

Full-waveform inversion based on Kaniadakis statistics.

Full-waveform inversion (FWI) is a wave-equation-based methodology to estimate the subsurface physical parameters that honor the geologic structures. Classically, FWI is formulated as a local optimization problem, in which the misfit function, to be minimized, is based on the least-squares distance between the observed data and the modeled data (residuals or errors). From a probabilistic maximum-likelihood viewpoint, the minimization of the least-squares distance assumes a Gaussian distribution for the residuals, which obeys Gauss's error law. However, in real situations, the error is seldom Gaussian and therefore it is necessary to explore alternative misfit functions based on non-Gaussian error laws. In this way, starting from the κ-generalized exponential function, we propose a misfit function based on the κ-generalized Gaussian probability distribution, associated with the Kaniadakis statistics (or κ-statistics), which we call κ-FWI. In this study, we perform numerical simulations on a realistic acoustic velocity model, considering two noisy data scenarios. In the first one, we considered Gaussian noisy data, while in the second one, we considered realistic noisy data with outliers. The results show that the κ-FWI outperforms the least-squares FWI, providing better parameter estimation of the subsurface, especially in situations where the seismic data are very noisy and with outliers, independently of the κ-parameter. Although the κ-parameter does not affect the quality of the results, it is important for the fast convergence of FWI.

Laboratory validation of effective acoustic velocity models for samples bearing hydrates of different type

This paper reports the results of a series of laboratory experiments on forming synthetic methane and tetrahydrofuran (THF) hydrates in sand samples and measuring their acoustic properties, i.e. P- and S-wave velocities. Experimental data were used for validating different effective acoustic velocity models connecting P- and S-wave velocities to the hydrate saturation. The envelope-cementing effective model gives the best fit for the methane gas-hydrate formation using the “excess-gas” method. These results may represent the case of onshore hydrate deposits formed from gas accumulations, e.g. Messoyakha Field in Siberia. The load-bearing effective model gives the best fit for THF hydrate formation that avoids gas phase in the system. These results should be typical to the most widespread hydrate deposits forms in marine sediments from the dissolved methane. We compare our experimental data with a compilation of similar published results to confirm their reliability and support our conclusions. Calibrated effective models can be further used for developing seismic-exploration methods for characterizing as well as confirming the origination history of natural gas-hydrate accumulations of different type.

Improved Anderson–Hampton acoustic velocity model for marine sandy gas-bearing sediments

Anderson & Hampton model is an important theoretical foundation for detection of marine shallow gas and identification of gas-bearing sediments. In the acoustic test of sandy gas-bearing sediments, it was found that if the excitation frequency is lower than the bubble resonance frequency, this model overestimates the attenuation of compressional wave velocity of the sediments, and it is difficult to accurately determine the original occurrence status of gas in the sediments by back analysis. The main reason is that the Gassmann equation in Anderson-Hampton model does not accurately describe the modulus parameters of sandy gassy sand sediments. Based on the occurrence characteristics of marine sandy gas-bearing sediments, an improved Anderson-Hampton acoustic velocity model was established by introducing saturation Sr and an attenuation factor reflecting the decay speed of wave speed. The advantage of the new model was pointed out and the sensitivity of parameter was analyzed simultaneously.

Simulating natural hydrate formation and accumulation in sediments from dissolved methane using a large three-dimensional simulator

The slow hydrate formation and accumulation process from dissolved gas in nature was simulated at large-scale in sandy sediments using a three-dimensional hydrate simulator. A low hydrate formation rate (0.37% pore volume per day) was achieved by maintaining the supersaturation of dissolved methane and migration velocity of the pore fluid at very low levels. Electrical and acoustic measurements were performed to track the hydrate formation in the sediments. A dynamic hydrate evolution process was observed: Crystallization – migration – accumulation – recrystallization. This evolution process has a critical effect on the connectivity of the pores and the strength of the sediment frame. After the dynamic evolution process, stiff frame-supporting hydrates and patchy hydrates were observed as the final morphologies in the pores, although the variation in the acoustic velocity indicated that frame supporting was the dominating morphology in the pores. Based on this, hydrate saturation values were inversed from resistivity and acoustic velocity models for frame-supporting hydrates. The calculated results showed that the order of hydrate saturation at different sites from the resistivity model was consistent with that from the acoustic velocity model by overestimating the hydrate saturation obtained from the resistivity inversion. Analysis of the spatial hydrate distribution showed that hydrate distribution in the sediments was discontinuous. The concentration, temperature, and fluid velocity profiles play a key role on this discontinuity, which we propose should be considered in detail in future experimental and numerical studies.

Numerical simulation of bubble plumes and an analysis of their seismic attributes

To study the bubble plume’s seismic response characteristics, the model of a plume water body has been built in this article using the bubble-contained medium acoustic velocity model and the stochastic medium theory based on an analysis of both the acoustic characteristics of a bubble-contained water body and the actual features of a plume. The finite difference method is used for forward modelling, and the single-shot seismic record exhibits the characteristics of a scattered wave field generated by a plume. A meaningful conclusion is obtained by extracting seismic attributes from the pre-stack shot gather record of a plume. The values of the amplitude-related seismic attributes increase greatly as the bubble content goes up, and changes in bubble radius will not cause seismic attributes to change, which is primarily observed because the bubble content has a strong impact on the plume’s acoustic velocity, while the bubble radius has a weak impact on the acoustic velocity. The above conclusion provides a theoretical reference for identifying hydrate plumes using seismic methods and contributes to further study on hydrate decomposition and migration, as well as on distribution of the methane bubble in seawater.

Rock physics modelling of acoustic velocities for heavy oil sand

The rock physics models for the acoustic velocities of the heavy oil sand with different oil distributions are proposed in this paper. The heavy oil distributions are classified into three primary types, based on which the corresponding acoustic velocity models are given. For the first type, the heavy oil is detached from the sand grains, for which the Hertz-Mindlin- Hashin-Shtrikman (HMHS) model can be used. In the second type, a continuous matrix is formed by the heavy oil and the sand grains float inside the matrix. The Hashin- Shtrikman (HS) lower bound gives good estimation results under this condition. For the third type, the heavy oil cements the sand grains at the grain contacts. The modified Contact Cement Theory (CCT) can describe this cementation effect well. To validate the proposed models, we analyzed the heavy oil sand data from Xinjiang Oil Field of China. The heavy oil distribution is obtained and the corresponding model is selected to estimate the acoustic velocities. The results show that it can predict the measured data well. Furthermore, we also compare the responses of the acoustic velocities for the heavy oil sand with different oil distributions. It reveals that different heavy oil distribution results in different acoustic velocities responses. The heavy oil distribution can thus be obtained by matching the measured data with the results estimated by the models. The model that match the data best implies the primary oil distribution. Many properties of the sand can then be estimated from the heavy oil distribution, such as the strength and permeability of the sand.

One dimensional acoustic direct nonlinear inversion using the Volterra inverse scattering series

Direct inversion of acoustic scattering problems is nonlinear. One way to treat the inverse scattering problem is based on the reversion of the Born–Neumann series solution of the Lippmann–Schwinger equation. An important issue for this approach is the radius of convergence of the Born–Neumann series for the forward problem. However, this issue can be tackled by employing a renormalization technique to transform the Lippmann–Schwinger equation from a Fredholm to a Volterra integral form. The Born series of a Volterra integral equation converges absolutely and uniformly in the entire complex plane. We present a further study of this new mathematical framework. A Volterra inverse scattering series (VISS) using both reflection and transmission data is derived and tested for several acoustic velocity models. For large velocity contrast, series summation techniques (e.g., Cesàro summation, Euler transform, etc) are employed to improve the rate of convergence of VISS. It is shown that the VISS method with summation techniques can provide a relatively good estimation of the velocity profile. The method is fully data-driven in the respect that no prior information of the model is required. Besides, no internal multiple removal is needed. This one dimensional VISS approach is useful for inverse scattering and serves as an important step for studying more complicated and realistic inversions.

Numerical simulation of bubble plumes in overlying water of gas hydrate in the cold seepage active region

To study the seismic responses produced by gas hydrate bubble plumes in the cold seepage active region, we constructed a plume water body model based on random medium theory and acoustic velocity model of bubble medium. The plume water body model was forward simulated by finite difference. Seismic records of single shot show the scattered waves produced by the plume. The scattered wave energy is strong where the plume exists. Where the scattered wave energy is stronger, the minimum of travel time is always above the plume, which has no relationship with the shot’s position. Seismic records of shot gathers were processed by prestack time migration. The migration section shows that the scattered waves produced by plumes can be imaged distinctly with higher accuracy. These researches laid a foundation for further study on the seismic responses produced by plumes and provided a new approach for the identification of gas hydrate.

T-wave source level and earthquake fault parameters: Empirically derived scaling relationships

Since 1991, NOAA/PMEL has routinely detected Tertiary waves from NE Pacific ocean earthquakes using the U.S. Navy’s SOSUS hydrophone arrays. Tertiary (T) waves are seismically generated acoustic waves (between 5–50 Hz) that propagate over great distances in the oceanic sound channel with little loss in signal strength. Low attenuation, combined with the availability of detailed ocean acoustic velocity models, have made T waves extremely useful for monitoring seafloor earthquakes;T waves allow for highly accurate epicentral locations, and a reduction of the detection threshold by almost two orders of magnitude. However, it is still not well understood how seismic energy generated at the earthquake source propagates through the seafloor–ocean interface and couples to the sound channel. Hence, the relationship between earthquake source parameters and the T-wave amplitude recorded at the hydrophone are not directly known. The goal of this paper is to empirically compare earthquake magnitude, seismic moment, and fault orientation (radiation pattern) to an event’s T-wave source level to see if systematic relationships exist. T-wave source level is estimated by deconvolving instrument gain and the transmission loss from the T-wave signal, thus resulting in the earthquake’s acoustic energy at the seafloor–ocean interface.

SEISMIC AND ELECTRICAL PROPERTIES OF UNCONSOLIDATED PERMAFROST1

ABSTRACTA model has been developed to relate the velocities of acoustic waves Vp and Vs in unconsolidated permafrost to the porosity and extent of freezing of the interstitial water. The permafrost is idealized as an assemblage of spherical quartz grains embedded in a matrix composed of spherical inclusions of water in ice. The wave‐scattering theory of Kuster and Toksoz is used to determine the effective elastic moduli, and hence the acoustic velocities. The model predicts Vp and Vs to be decreasing functions of both the porosity and the water‐to‐ice ratio. The theory has been applied to laboratory measurements of Vp and Vs in 31 permafrost samples from the North American Arctic. Although no direct measurements were made of the extent of freezing in these samples, the data are consistent with the predictions of the model. Electrical resistivity measurements on the permafrost samples have demonstrated their essentially resistive behaviour. The ratio of resistivity of permafrost in its frozen state to that in its unfrozen state has been related to the extent of freezing in the samples.Electromagnetic and seismic reflection surveys can be used together in areas of permafrost: firstly an EM survey to determine the extent of freezing and then the acoustic velocity model to predict the velocities in the permafrost. The necessary transit time corrections can thus be made on seismic reflection records to compensate for the presence of permafrost.