Sam Zandong Sun

The theory and application of DEM-Gassmann rock physics model for complex carbonate reservoirs

The storage spaces of carbonate reservoirs in the Tarim Basin are dominated by secondary pore porosity such as dissolution caves, holes, and cracks (Song et al., 2001). Research has shown that pore shapes can significantly influence the velocities of elastic waves (Eshelby, 1957; Kuster and Toksoz, 1974; Xu and White, 1995; Yan et al., 2002; Sun et al., 2004). Empirical VP-VS relationships such as Castagnas (1985) mudrock line, Hans relations (1986), and Greenberg-Castagnas (1992) relationship that ignore the effect of pore geometry are not applicable to carbonates (Xu and White, 1995; Wang et al., 2009). Cheng and Toksoz (1979) imaged various pore space structures using SEM, and proposed a pore aspect ratio spectrum that can be used to explain velocity prediction. Han (2004) measured velocities of 52 different carbonate rocks, indicating that caves influence seismic velocities due to high rigidity while micropores or cracks significantly decrease velocities of rocks due to the low rigidity. Regarding ...

Origin and architecture of fractured-cavernous carbonate reservoirs and their influences on seismic amplitudes

Ordovician carbonate is important for petroleum exploration and production in the Tarim Basin. The matrix porosity of the carbonate reservoirs is usually less than 2% because of their age-old sedimentary history and depth (deeper than 5500 m). The storage spaces are secondary dissolution pores and fractures which are dominated by visible dissolution caves, holes, and fractures (Zhou et al., 2006). The reservoirs are therefore called fractured-cavernous reservoirs. Good quality reservoir usually appears as a strong reflection (called “bead-like reflection” or “BR”) on poststack seismic sections, and the corresponding reservoir is called a fracture-cave body (FCB) which is the main drilling target of current E&P in Tarim Basin. The FCBs are unevenly distributed in carbonate strata, and the storage spaces inside an FCB are also unevenly distributed. Therefore, the reservoirs have both macroscopic and microscopic heterogeneities, which bring great challenges to petroleum exploration.

Studying heterogeneity and anisotropy via numerical and physical modeling

Palaeozoic carbonate rocks are widely distributed in China. The storage spaces of these carbonate reservoirs are mainly secondary and include dissolution caves, holes, and fractures. The original porosity is less than 1% in most formations. Hydrocarbon distribution is controlled by the systems of fractures and caves, which have a quasi-layered shape. The reservoir media are heterogeneous (caves and holes) and have varying degrees of anisotropy degree (fractures). So it is important to evaluate the heterogeneity and to check the accuracy of anisotropy inversion.

Two promising approaches for amplitude-preserved resolution enhancement

In western China, carbonate reservoirs are deeply buried and the surface is very loose. Due to Earth filtering, the high-frequency component of seismic waves is seriously absorbed. This causes a decrease in the dominant frequency and a narrowing of the frequency bandwidth of the seismic signals from a deep target. In the Tarim Basin, the most prospective reservoirs are below 6000 m. The seismic data recorded in these areas always show weak energy, low frequency, and phase distortion. Additionally, due to the desert surface, the energy of the seismic waves suffers serious absorption.

A Full-frequency Band Kuster-Toksöz Model And Its Application In Velocity Dispersion Analysis

Summary A modified microstructural dispersion model is introduced into the original high-frequency Kuster-Toksoz model, and a new full-frequency band Kuster-Toksoz model is established. Then the new model is employed to model the frequency-dependent P- and S-wave velocities (velocity dispersion) for a Fox Hills sandstone sample. The highfrequency velocities predicted and the low-frequency velocities corrected from the model are consistent well with the laboratory data measured at high- and low-frequency respectively, proving the practicality of the model. Additionally, the full-frequency band Kuster-Toksoz model is employed to study effects of fluids, porosity, viscosity, permeability, and crack parameters on velocity dispersion and frequency-dependent attenuation.

Q estimation using modified S transform based on pre-stack gathers and its applications on carbonate reservoir

Pre-stack seismic data is acknowledged to be more favorable in estimating Q values since it carries much more valuable information in traveltime and amplitude than post-stack data. However, the spectrum of reflectors can be strongly altered by nearby reflector or side lobes of the wavelet, which thereby degrades the accuracy of Q estimation based on the pre-stack spectral ratio method. To solve this problem, we propose a method based on the modified S-transform (MST) for estimating Q values from pre-stack gathers, in which Q values can be obtained with regression analysis based on the relationship between spectral ratio slope and the square of offset. Through tests on a numerical model, we first prove advantages of this pre-stack spectral ratio method compared to the traditional post-stack method. Besides, it is also shown that application of MST would lead to a much more focused intercept, which is the kernel for the pre-stack method. Therefore, the accuracy of Q estimation using MST is further improved when compared with that of conventional S-transform (ST). Based on this Q estimation method, we apply relevant processing methods (e.g. inverse Q filtering and dynamic Q migration) in practice, in order to improve imaging resolution and gathering quality with better amplitude and phase relationships. Applications on a carbonate reservoir witness remarkable enhancements of the imaging result, in which features of faults and deep strata are more clearly revealed. Moreover, pre-stack common-reflection-point (CRP) gathers obtained by dynamic Q migration well compensate the amplitude loss and correct the phase. Its ultimate pre-stack elastic inversion result better characterizes the geologic rules of complex carbonate reservoir predominated by secondary-storage-space.

GPU acceleration of amplitude-preserved Q compensation prestack time migration

Amplitude-preserved Q compensation prestack time migration (Q migration) is a new method that evolved from prestack Kirchhoff time migration (PKTM). Five algorithms are developed for Q migration on graphics processing units (GPUs). First, the principle of Q migration is briefly introduced. Second, one parallel strategy, namely, imaging domain parallel strategy, is proposed to accelerate Q migration on a single GPU by developing GPU algorithm. Results show that the imaging domain parallel strategy with the corresponding algorithm is superior to the CPU algorithm in several aspects, i.e., faster computing speed, shorter computing time, and higher computational efficiency. Third, based on the imaging domain parallel strategy, two division methods, namely, seismic data division method and velocity data division method, are presented to optimize the performance of Q migration on multi-GPUs and four algorithms are implemented by using Message Passing Interface (MPI)+Compute Unified Device Architecture (CUDA) and multi-thread+CUDA. An optimal algorithm is determined by comparing the performance of four algorithms. Results demonstrate that the optimal algorithm has the shortest computing time, which is 3.85 times shorter than that of a single GPU when four GPUs are all involved in computation and 300 times shorter than that of a 4-core central processing unit (CPU). Finally, a parallel computing framework on GPU cluster is established, which consists of imaging domain parallel strategy, seismic data division method and MPI+CUDA. This framework is suitable for all prestack time migration (PSTM) methods and has a short computing time and high speedup ratio on GPU cluster. Display Omitted We develop five algorithms on GPU for amplitude-preserved Q compensation prestack time migration.One parallel strategy is proposed to accelerate Q migration on a single GPU.We present two types of data division methods to further develop four algorithms on multi-GPUs.A parallel computing framework is established on GPU cluster from the optimal algorithm with the best performance, 300 speedup ratio.

Automatic fault extraction using a modified ant-colony algorithm

The basis of automatic fault extraction is seismic attributes, such as the coherence cube which is always used to identify a fault by the minimum value. The biggest challenge in automatic fault extraction is noise, including that of seismic data. However, a fault has a better spatial continuity in certain direction, which makes it quite different from noise. Considering this characteristic, a modified ant-colony algorithm is introduced into automatic fault identification and tracking, where the gradient direction and direction consistency are used as constraints. Numerical model test results show that this method is feasible and effective in automatic fault extraction and noise suppression. The application of field data further illustrates its validity and superiority.

3C-3D VSP vector wavefield separation with constrained inversion

Summary Vector wavefield separation of multicomponent offset VSP data based on an optimized localized parametric inversion was introduced in this paper. We developed and employed a constrained optimization search algorithm (COSA) in this inversion to determine the optimum incident angles and velocities. We firstly approximately evaluate the range of both the velocities and the incident angles of P and SV wave, which can respectively refer to the acoustic longing data and the geometry. Then the results were applied to the COSA. Synthetic and field data examples show that application of the COSA makes the inversion method more stable and fast to decompose multicomponent offset VSP wavefield. Introduction 1Vertical Seismic Profiles (VSP) has been applied to many fields of oil industry, such as imaging the vicinity of the well, analyzing anisotropic velocity, forecasting formation pressure and so on (Chen, 2007; Cao, 2008). 3C VSP wavefield separation is one of the most fundamental and important steps in the processing of 3D VSP. The reason is that wavefield separation prior to migration (reflection imaging) has historically been performed and the separated wavefields can be used in deterministic deconvolution (down-going P wavefield), true amplitude AVO analysis, event identification and time picking for tomography (Leaney, 2002). The method described in this paper is related to the method of Esmersoy (1998, 1990) and Leaney (1989, 1990). The distinction is that we develop a constrained optimization search algorithm to determine the range of optimum values for velocities and incident angles. The speed of convergence is obviously increased, and the wavefield decomposed is crisper. Modeling the VSP data as the superposition of a small number of plane waves (up- and down-going P, up- and down-going SV) well approximates to field seismic data. Additionally, this method can estimate the arrival angle and medium velocity at receiver locations more reliably (Esmersoy, 1990). The problem of inversion is usually ambiguity. Thus if we can acquire some prior information (such as the range of velocities; angles; the ratio of P and SV velocity; stratigraphic information; even Snell’s Law for isotropic formation), and add them in the inversion algorithm, the multiplicity of solution will be greatly diminished.

3D Seismic attribute optimization technology and application for dissolution caved carbonate reservoir prediction

Summary The fractured and caved Ordovician carbonate reservoirs in Tarim Basin are characterized by significant seismic anisotropy and heterogeneity. It is very difficult to predict this type of reservoir using common seismic attribute analysis technology because of the uncertainty of analysis results. On the basis of investigating the application condition and effectiveness of various seismic attributes, focusing on the heterogeneity of carbonate reservoir and taking logging data as the known training samples, this paper realizes the 3D optimization processing for multi seismic attribute data volume by combing of multi statistic methods such as multivariate discriminant analysis, kernel principal component analysis etc. The established reservoir discriminant model in this paper has the total coincidence rate of 90.13% by retrospective test, proving the validity of the model. The optimization results of multi seismic attribute data can finely sculpture the favorable distribution range and geometric shape of dissolution caved carbonate reservoir in 3D space, thus obtaining good geologic results. It is expected that this paper can provide a scientific and practical method for seismic multi-attribute optimization and dissolution caved carbonate reservoir prediction.