Among the various risks associated with CO2 storage in deep geologic formations, wells are important potential pathways for fluid leaks and groundwater contamination. Injection of CO2 will perturb the storage reservoir and any wells that penetrate the CO2 or pressure footprints are potential pathways for leakage of CO2 and/or reservoir brine. Well leakage is of particular concern for regions with a long history of oil and gas exploration because they are top candidates for geologic CO2 storage sites. This review explores in detail the ability of wells to retain their integrity against leakage with careful examination of the coupled physical and chemical processes involved. Understanding time-dependent leakage is complicated by the changes in fluid flow, solute transport, chemical reactions, and mechanical stresses over decade or longer time frames for site operations and monitoring.Almost all studies of the potential for well leakage have been laboratory based, as there are limited data on field-scale leakage. Laboratory experiments show that CO2 and CO2-saturated brine still react with cement and casing when leakage occurs by diffusion only. The rate of degradation, however, is transport-limited and alteration of cement and casing properties is low. When a leakage path is already present due to cement shrinkage or fracturing, gaps along interfaces (e.g. casing/cement or cement/rock), or casing failures, chemical and mechanical alteration have the potential to decrease or increase leakage risks. Laboratory experiments and numerical simulations have shown that mineral precipitation or closure of strain-induced fractures can seal a leak pathway over time or conversely open pathways depending on flow-rate, chemistry, and the stress state. Experiments with steel/cement and cement/rock interfaces have indicated that protective mechanisms such as metal passivation, chemical alteration, mechanical deformation, and pore clogging can also help mitigate leakage. The specific rate and nature of alteration depend on the cement, brine, and injected fluid compositions. For example, the presence of co-injected gases (e.g. O2, H2S, and SO2) and pozzolan amendments (fly ash) to cement influences the rate and the nature of cement reactions. A more complete understanding of the coupled physical–chemical mechanisms involved with sealing and opening of leakage pathways is needed.An important challenge is to take empirically based chemical, mechanical, and transport models reviewed here and assess leakage risk for carbon storage at the field scale. Field observations that accompany laboratory and modeling studies are critical to validating understanding of leakage risk. Long-term risk at the field scale is an area of active research made difficult by the large variability of material types (cement, geologic material, casing), field conditions (pressure, temperature, gradient in potential, residence time), and leaking fluid composition (CO2, co-injected gases, brine). Of particular interest are the circumstances when sealing and other protective mechanisms are likely to be effective, when they are likely to fail, and the zone of uncertainty between these two extremes

Summary: This paper describes a fully coupled finite element/finite volume approach for simulating field‐scale hydraulically driven fractures in three dimensions, using massively parallel computing platforms. The proposed method is capable of capturing realistic representations of local heterogeneities, layering and natural fracture networks in a reservoir. A detailed description of the numerical implementation is provided, along with numerical studies comparing the model with both analytical solutions and experimental results. The results demonstrate the effectiveness of the proposed method for modeling large‐scale problems involving hydraulically driven fractures in three dimensions. © 2016 The Authors. International Journal for Numerical and Analytical Methods in Geomechanics published by John Wiley & Sons Ltd.

Lattice Boltzmann methods (LBM) are used for the computational simulation of Newtonian fluid dynamics. LBM-based simulations are readily parallelizable; they have been implemented on general-purpose processors, field-programmable gate arrays (FPGAs), and graphics processing units (GPUs). Of the three methods, the GPU implementations achieved the highest simulation performance per chip. With memory bandwidth of up to 141 GB/s and a theoretical maximum floating point performance of over 600 GFLOPS, CUDA-ready GPUs from NVIDIA provide an attractive platform for a wide range of scientific simulations, including LBM. This paper improves upon prior single-precision GPU LBM results for the D3Q19 model by increasing GPU multiprocessor occupancy, resulting in an increase in maximum performance by 20%, and by introducing a space-efficient storage method which reduces GPU RAM requirements by 50% at a slight detriment to performance. Both GPU implementations are over 28 times faster than a single-precision quad-core CPU version utilizing OpenMP.