Factors affecting injection well performance and fracture growth in waterflooded reservoirs
MetadataShow full item record
Waterflooding involves the injection of water to displace oil from oil and gas reservoirs. Well over 80% of oil reservoirs will undergo waterflooding at some point in their life. It is, therefore, important to understand some key aspects of this process that have hitherto not been well studied. This dissertation investigates the following aspects of waterflooding: (i) the filtration of solids and oil-in-water emulsions in fractured and unfractured injection wells, (ii) the generation and filtration of oil-in-water (O/W) emulsion droplets in the near-well region or in the fracture, (iii) the height-growth and containment of injection-induced fractures, and (iv) the stress reorientation induced by water injection when waterflooding reservoirs. These aspects are investigated as separate physical phenomena, but their impacts are integrated using the platform of a comprehensive waterflooding injection well model. The first phenomenon investigated is filtration in frac-packed injectors. During long-term water injection, solid particles in the injection water may deposit in the proppant pack of frac-packed injectors. Researchers have not fully understood whether particles will travel without plugging the frac-packs or deposit in the near-well area under the high-velocity flow conditions in the proppants. Filtration behavior under frac-pack flow conditions is the most important factor that determines overall injector performance. In this dissertation the filtration of injected solids under these conditions was experimentally studied, and the effect of frac-pack filtration on the injector performance was predicted. The flow of dilute oil droplets in a porous medium under near-well conditions was experimentally investigated. When the porous medium has a residual oil saturation, oil droplets can be generated by viscous forces overcoming entrapping capillary forces. The generated oil droplets will subsequently participate in filtration processes along with injected oil droplets. If this occurs in the near-injector area, the injectivity can severely decline and this may require expensive remediation processes. In this study, prediction of O/W emulsion flow was improved by experimental observations of the rates of generation and filtration of oil droplets. In a larger scale problem, a 3-dimensional model of water-injection-induced fracture was developed to predict the fracture height growth. If a fracture breaches the bounding layers, the sweep efficiency can be significantly impaired and it could have severe environmental consequences (such as contamination of shallower aquifers or the seabed). During long-term water injection, fracture growth can only be simulated properly when the filtration near fractures, thermo-elastic stress changes and reservoir fluid flow behavior are all concurrently calculated. Based on this new model, the impact of reservoir stress conditions, mechanical properties, and injection-water quality on fracture growth was studied. On a reservoir-scale, the stress reorientation caused by injection-production activities during waterflooding was investigated. A new finite-volume multi-phase reservoir simulation with poro- and thermo-elasticity was developed. This model was applied to various waterflooding well patterns, such as five-, nine-spot, line-drive and horizontal well pairs, and the critical geomechanical responses by injection-production activities during waterflooding operations were analyzed. The model can be used to predict the direction of induced fractures, design infill well locations and configurations and optimize the reservoir sweep. Through the use of both experimental observations and numerical models this work has elucidated various physical phenomena affecting fracture growth and injection-well performance. The findings in this dissertation provide critical data and models that help us to more confidently specify injection water quality, the design of pumping and water treatment facilities, and the optimization of well planning. The models developed in this work can be used to substantially improve the predictions of injection well performance and improve reservoir oil recovery by waterflooding.