Fluid and rock characterization using new NMR diffusion-editing pulse sequences and two dimensional diffusivity-T2 maps
Hirasaki, George J.
Doctor of Philosophy
New down-hole nuclear magnetic resonance (NMR) measurement and interpretation techniques have substantially improved fluid and reservoir characterization. These techniques take advantage of the magnetic field gradient of the logging tools to make diffusion sensitive NMR measurements. In this work, new NMR pulse sequences called "diffusion-editing" (DE) are used to measure diffusivity and relaxation times for a variety of samples. We use a new inversion technique to obtain two-dimensional maps of diffusivity and relaxation times, and propose new interpretation approaches for these maps. Two DE NMR pulse sequences are of particular interest. First is the CPMG-DE Pulse sequence, based on the Carr-Purcell-Meiboom-Gill pulse sequence in a magnetic field gradient. Results presented here demonstrate that CPMG-DE measurements can be used to determine the saturation of partially saturated samples, detect wettability change, and observe the presence of internal field gradients or restricted diffusion. The second DE sequence of interest is PFG-SE DE, based on the pulse field gradient stimulated echo (PFG-SE) sequence. The PFG-SE DE sequence is particularly well suited for analyzing restricted diffusion, which occurs when the spins are prevented from diffusing freely by the presence of confinement such as pore walls. The interpretation of pore-size distributions from PFG-SE DE measurements can be simplified by approximating the pore network as a system of spheres of varying sizes. A procedure is developed for determining the optimal parameters for measuring spheres of a selected size. A technique for combining the sensitive portions of multiple measurements using masks is also developed. Experiments are performed on grain packs, vuggy carbonates, and emulsions. In most cases, the range of length scales the system is sensitive to is too narrow using the available range of gradients to provide interesting information about the systems, but in the case of emulsions the results are positive when evaluating droplet size distributions.