Pore scale analysis of NMR diffusion-relaxation responses of sandstones saturated by complex fluids

The interpretation of relaxation responses typically assumes fast diffusion, weak coupling, and homogeneous surface relaxivity for permeability estimation. Fluid typing, using D-T2 responses, relies on the distribution of the peaks in D-T2 maps with regard to the water line (constant diffusion coefficient) and an alkane line (relaxation-diffusion linear correlation). For crude oils, in particular with higher asphalthene content, the NMR response is considerably more complex. In this case, internal gradients, restricted diffusion, slow diffusion and non-zero surface relaxivity may strongly influence the results. In order to support the interpretation of NMR responses in such conditions, numerical simulations can be a convenient tool.

In recent years digital core technology has made significant advances. It is now possible to obtain registered sets of high-quality tomograms of rock samples from reservoirs at various resolutions, field of views and for different saturation conditions. In previous work we considered the analysis of NMR responses by numerical methods on the basis of topological partitions and introduced the simulation of more complex NMR responses. In this work we combine more complex simulation methods with topological analysis to interpret diffusion-relaxation responses for a set of rocks saturated by complex fluids.

We consider a set of samples covering different pore size ranges and thus diffusion lengths. Measurements of D-T2 for the complex fluids are carried out to determine the simulation parameters for the fluids themselves. We then compare measurements and simulation of the fluid saturated samples. Whilst modelling the crude as a set of independent fluid components, we analyse the spatial support of each different fluid component in the confined pore space and interpret the NMR measurements with regard to restricted diffusion, slow diffusion regime and potential permeability correlations.