Use this URL to cite or link to this record in EThOS:
Title: Upscaling for thermal recovery
Author: Abubakar, Shamsudeen
ISNI:       0000 0004 6347 1176
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Thermal recovery is one of the most widely applied EOR methods worldwide. It is used mainly to improve the recovery of more viscous oils. Heat is used primarily to reduce the viscosity of oil but may also improve recovery by other mechanisms such as oil expansion, steam distillation and reduction in capillary forces. As a result, the prediction of thermal recovery performance requires more complex and computationally demanding numerical simulations than is the case when evaluating a waterflood. The simulator has to evaluate the impact of both heat transfer and compositional behaviour on the fluid flows. In most cases the engineer will have to use coarser grids in the simulation models than would be needed for a waterflood in order to keep simulation times within acceptable limits. However, it has been shown in the modelling of other EOR processes that in fact much finer grids may be needed to capture the frontal dynamics of the displacement than would be the case in waterflooding. As a result, the simulation of steamflooding on the field scale is unlikely to capture the impact of small scale heterogeneity on flow, as well as being severely affected by numerical dispersion. The ideal solution is to use homogenization to capture the effects of sub-grid-block heterogeneity combined with advanced numerical techniques such as dynamic/adaptive gridding or higher order schemes to reduce the impact of numerical diffusion and dispersion. However, such methods are not currently available in the commercial simulators available to most engineers. The alternative is to use upscaling, but as yet there are no upscaling methodologies suitable for thermal oil recovery methods. Upscaling is a way of altering the inputs to numerical flow models so that simulations run on a coarse grid. These will a) predict the impacts on flow of sub-grid heterogeneity (homogenization) and b) will be less affected by numerical dispersion. For waterflooding this is typically achieved by calculating the effective absolute permeability for each coarse grid cell. In many cases, it is also necessary to upscale the well index and the transmissibilities of the well blocks. Upscaling the absolute permeability, the well index and the well block transmissibilities, ensures that the overall pressure drop for single phase flow between wells is correctly predicted in the coarse grid, whilst upscaling the relative permeabilities ensures that the two-phase flow effects (pressure drop, breakthrough time and water cut development) are correctly modelled. Although there is a significant literature on single-phase upscaling and the development of pseudo relative permeabilities for waterflooding applications, these methods do not capture the heat transport or the impact of temperature on the oil mobility. In this research, a combination of pseudoisation procedures (by making reasonable assumptions) with the Buckley-Leverett solution approximated for thermal EOR processes is used to derive analytical pseudo-relative permeabilities for both hot water and steam flooding. The methodology involves upscaling both the oil viscosity dependence on temperature and the relative permeabilities to compensate for the increased numerical dispersion that occurs in coarse grid simulations. The methodology is demonstrated by comparing the fine and coarse grid simulations. The approach provides significantly improved predictions compared with performing coarse grid simulations without upscaling. For the 3D cases, we found out that the near well upscaling was very important in order to capture pressure and minimise grid orientation effects. For upscaling the well index we used the Near-Well Arithmetic Averaging (NWAA) and the results showed that upscaling the well index improved the ability of the coarse grid simulation when compared without the upscaling procedure. The investigation also suggests that steamflooding in a combination of gravity and viscous forces with little capillary effects, and that gravity forces dominate in the steam zone while viscous forces are dominant in the hot water zone. The study showed numerical smearing of the temperature and flood fronts occur in the coarser models, leading to misrepresentation of the phase distributions within the model. These misrepresentations have an effect on the oil recovery and breakthrough times. This study also showed that grid orientation error adversely affects the temperature and flood fronts in steamflooding, using a non-uniform grid for steamfloods alters the grid orientation error. For the phase behaviour of steam, it is shown that it is more important for upscaling to capture the fine grid pressure changes in steamflooding than in water waterflooding, because of the influence that pressure has on phase behaviour.
Supervisor: Muggeridge, Ann ; King, Peter Sponsor: Ministry of Petroleum Resources ; Nigeria
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral