Nghiem, Long X. (Computer Modelling Group Ltd.) | Mirzabozorg, Arash (University of Calgary) | Chen, Zhangxin John (University of Calgary) | Hajizadeh, Yasin (Computer Modelling Group Ltd.) | Yang, Chaodong (Computer Modelling Group Inc.)
History matching of reservoir flow models based only on production data may not reveal deficiencies that affect future predictions. Incorporating saturation and temperature profile data that come from 4D seismic surveys in the history matching process can reduce the uncertainty of reservoir models for the prediction stage. We constructed a field reservoir model from which production history, saturation and temperature profile history were obtained. We started the history matching process with a base reservoir model, the petro-physical properties of which were substantially different than those of the field reservoir model. We propose a new methodology for matching the fluid and temperature profiles by adjusting reservoir petro-physical properties. In this methodology, some grid blocks in a reservoir model were selected judiciously to capture the overall saturation and temperature distribution profiles. In addition to well production data, we included the saturation and temperature profiles at these grid blocks as extra objective functions during the history matching process. The DECE optimization is used to reduce the objective function. We applied this method in a Steam Assisted Gravity Drainage (SAGD) process and matched the saturation and temperature profiles with an average error of less than 2%.
Among of the new inventions on thermal recovery, Fast-SAGD was introduced as the next generation of SAGD with greater amounts of bitumen and lower injected steam. However, there are still many suspicions about the successful of this technology such as the incremental bitumen recovery of Fast-SAGD is from the SAGD production well or combined with the offset well? It is very difficult to conclude that Fast-SAGD is better than conventional SAGD when numerical simulation of two processes was conducted in different well pattern as well the amount of operated well.
This paper presented a comparative evaluation between conventional SAGD and Fast-SAGD in three typical formations (McMurray, Clearwater, and Bluesky) of Alberta's Oil Sand. Three reservoir models with over one hundred numerical simulations under various operation conditions were developed to achieve the most unprejudiced comparison between two recovery processes. The simulation results proved that significantly recoverable bitumen was originally produced from offset well in Fast-SAGD system and leads to higher recovery factor. But, there is only slight increase in cumulative oil recovery when two processes were performed in same pattern with similar number of production wells. The result also indicated that the difference of 10kPa between steam injection pressure and reservoir pressure in literature is not enough for both SAGD and Fast-SAGD operations. And then, this study presented a numerical investigation for evaluating the potential applicability of Fast-SAGD recovery process under complex reservoir conditions such as shale barriers, thief zones with bottom and/or top water layers, overlying gas cap and fracture systems in Clearwater formation.
Reservoir simulation for a full field heterogeneous model with millions of grid blocks demands significant computational time so improving the computational efficiency becomes crucial in designing a reservoir simulator. Graphics Processing Unit (GPU), a new high-profile parallel processor with hundreds of microprocessors, stands out in parallel simulation because of its efficient power utilization and high computational efficiency. Also, its cost is relatively low, making large-scale parallel reservoir simulation possible for most of desktop users.
In this paper several GPU-based parallel preconditoners, in conjunction with a new GPU-based GMRES algorithm, are proposed and coupled with an in-house black-oil simulator to speedup reservoir simulation. In particular, massively parallel ILU preconditioners (ILU(0), ILUT, block ILU(0), block ILUT), which are usually regarded as data dependence and highly sequential preconditioners, are developed on GPUs.
In the numerical experiments performed, the SPE 10 problem, a 3D heterogeneous benchmark model with over one million grid blocks, is selected to test the speedup of our GPU solver and preconditioners. On the state-of-the-art CPU and GPU platform, the new GPU implementation can achieve a speedup of over eight times in solving linear systems arising from this SPE 10 problem compared with the CPU based serial solver. Moreover, our GPU solver is successfully coupled with the in-house black-oil simulator to test the performance of the whole parallel simulation process, with a speedup of about six times. The excellent speedup and accurate results demonstrate that the GPU-based parallel linear solver and preconditioners have the great potential in parallel reservoir simulation.
In this paper, we present our work on developing parallel preconditioners for iterative linear solvers on NVIDIA Tesla GPU. A unified triangular solver for ILU-like preconditioners is developed. A new matrix format and the corresponding parallel algorithm are designed. Based on this triangular solver, block ILU(0), block ILUT and domain decomposition preconditioners are implemented. Numerical experiments show that these GPU based parallel preconditioners are around 8 times faster than our CPU based preconditioners.
The injection of different solvents, such as propane and CO2, into bitumen, has proven to be an effective method in the production of these kinds of reservoirs. However, in some cases, the prediction of large solvent requirements can make it uneconomical. The formation of a second liquid phase has been observed when the solvent is propane or CO2, with the second liquid phase mainly composed of the solvent itself.
The objective of this research is to understand the importance of this second liquid phase and its effect on production. Also, a simulator that can allocate an individual phase to this liquid phase would allow for prediction of the amount of solvent that can be produced and recycled. This makes the cost evaluation of solvent injection processes to be more realistic.
Depending on the reservoir fluid distribution, a three- or four-phase flow can occur in the absence or presence of water. A compositional simulator based on an equation of state is designed to simulate these multiphase situations. This simulator has a four-phase flash and stability subroutine, which make it more realistc compared to other compositional simulators. In fact, it can handle a maximum of three hydrocarbon phases and one aqueous phase. Relative permeability plays an important role in multiphase flow; numerical results indicate that, by increasing the number of phases, there is an increase in project life. It is valuable to mention that the results of this research can be also used in CO2 sequestration.
Heavy Oil and Bitumen
Heavy oil and bitumen are viscous mixtures of hydrocarbons and other organic compounds. In general, heavy oil and bitumen are classified according to API and viscosity. Crude oils with API less than 20º and viscosity less than 10,000 cp are known as heavy oils while those with viscosities greater than 10,000 cp are termed as extra heavy oil or bitumen (Das, 1995; Miller 1994).
Canada has huge heavy oil and bitumen resources. Original Oil In Place, OOIP is estimated to be more than 400 billion m3 (approximately 2.5 trillion bbl) which is approximately twice that of the total conventional oil reserves in the Middle East (Dusseault, 2001; Farouq Ali, 2003).
These reserves exist in unconsolidated sand and carbonate sedimentary formations of Athabasca, Cold Lake, Peace River and Wabasca regions in Alberta, Saskachewan and British Columbia. Heavy oil and bitumen are becoming more and more important considering the depletion of conventional oil reserves in the world. However, production of heavy oil and bitumen is more challenging and more expensive than that of conventional oils because of immobility of them in reservoir conditions due to their high viscosity.
There are two methods for recovery of heavy oil and bitumen: open pit mining by using large trucks and shovels and insitu recovery through wells. The first method is very effective (more than 90% recovery). However, this is only suitable for shallow reservoirs with overburden formation depth of less than 75m. In Canada, most of the reservoirs are deep enough not to be exploited by open pit mining.
At present time cold heavy oil production (CHOP), cyclic steam stimulation (CSS) and steam-assisted gravity drainage (SAGD) are the major in-situ recovery methods used in Canada.