The anelastic effects of the earth can cause frequency dependent energy attenuation and phase distortion, especially when gas clouds are present. To correct these unwanted effects for proper imaging, both the velocity and quality factor (Q) models need to be accurately estimated. With FWI offering the capability to obtain higher-resolution models than tomography, visco-acoustic FWI (Q-FWI) is highly desirable for inverting both Q and velocity models together.
The visco-acoustic wave propagation in an anisotropic medium and the gradient computation for model parameters can be implemented in the framework of FWI. However, the similar radiation patterns between velocity and Q make the joint inversion non-trivial (
To compensate for the phase distortion and amplitude loss due to gas absorption, both the velocity and the quality factor (Q) models need to be accurately estimated for proper imaging (Xie et al., 2009). Besides imaging, the Q model also helps to provide useful information about rock characteristics, such as saturation. Full-waveform inversion (FWI) has established the capability for high-resolution and high-fidelity velocity model building. Q estimation in the framework of FWI is highly desired, and the greatest challenge for Q and velocity inversion in FWI is the crosstalk between the parameters since their radiation patterns are similar. Here we describe a visco-acoustic full-waveform inversion (Q-FWI) approach to jointly estimate velocity and Q models. We demonstrate with both synthetic data and a field data example from the Norwegian North Sea that Q-FWI can invert for high-resolution velocity and attenuation models, providing superior imaging using an attenuation compensating pre-stack depth migration (QPSDM).
Wang, Min (Texas A&M University) | Wei, Chenji (Research Institute of Petroleum Exploration & Development, PetroChina) | Song, Hongqing (University of Science and Technology Beijing) | Efendiev, Yalchin (Texas A&M University) | Wang, Yuhe (Texas A&M University)
In this paper, we couple Discrete Fracture Network (DFM) and multi-continuum model with Generalized Multiscale Finite Element Method (GMsFEM) for simulating flow in fractured and vuggy reservoir. Various scales of fractures are treated hierarchically. Fractures that have global effect are modeled by continua while the local ones are embedded as discrete fracture network based on the geologic observation. For independent vugs, a continuum is used to represent their effects with specific configuration that there's no intra-flow of this continua. GMsFEM enables us to systematically develop an approximation space that contains prominent sub-grid scale heterogeneous background information based on the multi-continuum and DFM model. Conforming unstructured mesh is used to surrender the application of random discrete fracture networks. This paper targets on the improvement of the flow simulation performance in complex high-contrast domain by extending the ability of multiscale method to modeling arbitrary discrete fracture network. This advancement by GMsFEM is motivated by the limited capability of Multiscale Finite Element Method (MsFEM) on modeling discrete fractures when multiple fracture networks present in same coarse block. Multiple numerical results are shown to validate the efficiency of our coupled method.
Yang, Fan (Wuhan University of Technology) | Wang, Lizheng (Wuhan University of Technology) | Wang, Jiamei (Wuhan University of Technology) | Chen, Shunhuai (Wuhan University of Technology) | Luo, Liang (Wuhan University of Technology) | Wang, Min (Wuhan University of Technology)
Aiming at the drag reduction of a river-sea bulk cargo by gas film, the original ship has been scaled 31.98 times smaller to be the calculation model in order to compare the simulation results with the experimental ones, which turns to solve the incompressible unsteady turbulent flow and the gas-liquid two phase flow. On the basis of numerical filtering of the jetting direction and position, the simulation shows that with three slots jetting simultaneously, the total drag coefficient is reduced by 19% and frictional drag coefficient decreased by 22%, which corresponds to the results of completed towing experiment.
Lots of investigations about numerical simulation of drag reduction by gas film have been made. Liming Lin (2002) of Wuhan University of Technology calculated the uniform bubble flow at the velocity of 4m/s, the initial bubble concentration of 30% and the bubble diameter of 1000μm for the drag reduction of a plate. The results show that the drag reduction rate is between 50% and 60%. Chengsheng Wu and Shulong He (2005) established a simplified physical model of two-phase flow around a high-speed bubble boat. Numerical calculations were carried out using the VOF method and the finite volume method. The results show that the bubble length increases with the increase of the flow velocity which is beneficial to the formation and stabilization of the bubble. Weitao Zheng, Ziqing Ma, Keqiang Chen and Peng Yang (2009) used the mixture model and the finite volume method to simulate the relative motion between bubble and water on the basis of fluent software. The effect of microbubbles on the drag of the ship under different floating states was studied. The results show that the drag reduction effect is the best, and the drag reduction rate reaches 35.2% under the condition of air bubbles. Yong Li (2011) investigated numerical simulation of micro-bubble drag reduction for a plate with the Mixture model while initial bubble concentration, velocity at the entrance and the size of bubbles had been studied. It shows that the drag reduction rate of local friction resistance can reach 86.55%, and the total drag reduction rate can reach 26% when applied on a ship model. Bingliang Wang (2012) applied the Euler multi-phase flow model and considered the influence of the interphase force, and selected the general drag law model to calculate the microbubble drag reduction for the plate and bulk ship model. The calculation results show that the number of meshes needed in the mixture model calculation is small and the convergence is good, but the error is large. Euler multiphase flow model with high accuracy has large computational complexity and unstable value. K Mohanarangam and SCP Cheung (2009) used the population balance model to study the effect of bubble breaking and polymerization on the bubble drag reduction. The numerical simulation results are in good agreement with the experimental results. HJ Park and Y Tasaka (2015) demonstrated that the repetitive jetting of air bubbles can achieve a considerable drag reduction compared to the conventional continuous jet method by using a bubble repetitive jet method to generate a stable bubble group to ensure a high bubble concentration. EL Amromin (2016) proposed a new model for exploring the interaction of the bottom cavitation with the boundary layer, which includes the incompressible air flow at the cavitation, the compressible air-liquid mixing flow at the boundary layer, and the incompressible exterior water flow without rotation. The numerical calculations based on this model took the observed effects at the early stage into account and the prediction of the required air flow is consistent with the experiment.
In this paper we develop dynamic-warping full-waveform inversion (D-FWI) to address the well-known cycle skipping problem in conventional full-waveform inversion (FWI). The dynamic warping technique is used to detect the traveltime difference between the predicted and the observed data. We make use of the timeshift to partially warp the observed data and thus generate a series of datasets that connect the predicted and the observed data. We then use the modified observed data to solve a sequence of conventional FWIs that avoid the cycle skipping issue. Synthetic and real data examples show that D-FWI can converge successfully by overcoming the cycle skipping problem, while conventional FWI results in an erroneous model. With this new approach, we can invert velocity models starting from a higher frequency and/or a poor starting model. This technology has the potential to save time in the processing sequence since it allows the velocity model building to start with minimum preprocessing on the seismic data and to be done in parallel with other pre-processing steps.
Presentation Date: Wednesday, October 19, 2016
Start Time: 8:50:00 AM
Presentation Type: ORAL
Zhang, Kai (University of Calgary) | Liu, Qingquan (University of Calgary) | Wang, Min (University of Calgary) | Kong, Bing (University of Calgary) | Lv, Jiateng (University of Calgary) | Wu, Keliu (University of Calgary) | Chen, Shengnan (University of Calgary) | Chen, Zhangxin (University of Calgary)
Shale gas production gets a tremendous breakthrough with the advent of horizontal well and massive hydraulic fracturing. There is still extensive gas stored in a reservoir after primary production so gas injection has a possibility to improve gas recovery for a shale play. In shale gas reservoirs, the gas can be in a free or absorbed state. In addition, there is a difference in the adsorption capacity between hydrocarbon and non-hydrocarbon components, therefore, it may cause gas recovery variation by hydrocarbon and non-hydrocarbon gas injection. In this paper, a Montney shale gas reservoir is modeled by Petrel based on data from Accumap.
Asphaltenes are of particular interest to the petroleum industry because of their depositional effect which creates problems for production, storage, transportation and refinery processes. A class of non-ionic polymeric surfactants has been developed to prevent the aggregation of asphaltene colloids in crude oils. A surfactant dosage rate as low as 25 ppm can be used to keep the asphaltenes dispersed at nearly 100 %. These polymeric surfactants are made from sustainable and biodegradable raw materials and free of BTEX, other aromatic solvents and phenol formaldehyde resin.
The polymeric surfactants were synthesized with a range of monomers at various ratios and under different conditions. The products were then tested in three crude oils from the USA and Canada (API: 45-11 °) to evaluate their performance in a range of systems. The inhibition effect was analyzed with an optical scanning device according to ASTM D7061-06. It was found that the chemical bonding and physical absorption between an asphaltene molecule and the polymeric surfactant played an important role in stabilizing the asphaltene colloids in crude oil. The hydrophobic chain of the polymeric surfactant provided steric hindrance between the asphaltene colloids while the polar groups gave multiple interaction points for bonding and absorption to the asphaltene. Achieving a balance between these aspects of the molecular design has created a new class of polymeric surfactants based on sustainable and biodegradable raw materials which efficiently inhibit the precipitation of asphaltenes from a range of crude oils at low dose rates.
Summary Extending our previous work on shallow water demultiple (SWD), we incorporate the multichannel prediction operator that is derived from SWD in to a modelling process for attenuating internal multiples generated by shallow reflectors. The resulting process is a fully data driven method that overcomes the commonly observed issue of missing data caused by the near offset gap in shallow water acquisition. In addition, no subsurface information is required in the multiple prediction process as the multichannel prediction operator is first derived from the water layer related surface multiples and then either the inverse scattering series or layer-based convolution method can be used for generating the internal multiple model. We demonstrate through field data examples that the internal multiples predicted by our approach exhibit better event continuity and higher fidelity with respect to frequency bandwidth. Consequently, it imposes less burden on the subsequent subtraction step which in turn leads to enhanced performance of internal multiple attenuation.
Internal multiple attenuation (IMA) has long been regarded as a challenging problem in seismic data processing. The major difficulty stems from the need of identifying the multiple-generating interfaces for IMA.We present and compare two advanced IMA methods that a-priori knowledge about the multiple-generating interfaces is not required in the modeling process. The two approaches are inverse scattering series (ISS) and our methodology based on the convolution-correlation process. We describe our implementation procedure that involves segmenting the input data into different time windows and using a sliding-window approach to ensure that all the top multiple-generating interfaces are included in the modeling process so that internal multiples of all orders can be predicted without any subsurface information. We also discuss and compare the different assumptions made in ISS and our approach for IMA. We show the application of the two methods for handling internal multiples on both the synthetic and field data from the Tupi oilfield in the Santos Basin and on the field data acquired offshore Australia.