NMR has been used extensively for oilfield exploration, in both laboratory settings and in the borehole. Relaxation time measurements (T1 and T2) have proved particularly valuable for measurements of rock permeability and fluid typing while signal intensity has been correlated to saturation. Recent advances allow for downhole measurement of diffusion coefficients, and it is known that diffusion coefficients can be a robust indicator of the hydrocarbon molecular size in a mixture of alkanes (Freed 2005). This work applies the previous work to actual downhole measurements and produces a running log of average molecular size versus depth in oil wells.
Multi-dimensional diffusion-relaxation experiments conducted at every depth in a well produce maps which are then analyzed by a computer algorithm. Bound fluid shows in the map at fast relaxation times; since this fluid will not flow, signal from this region is ignored. Moveable oil and gas often exhibit longer relaxation times and oil signals have slower diffusion constants than gas. These signals are identified by the algorithm to obtain the distribution of diffusion coefficients from which the mean carbon number (MCN) for the different phases can be calculated. Because data is taken at multiple frequencies and thus multiple depths of investigation (DOI), the chain length at different DOIs can be obtained. This DOI dependent MCN allows for the identification of invading fluids in shallower depths and the true formation fluids in the deeper depths. Furthermore, this method also shows some regions where significant oil and gas saturations are coexistent and independently finds an average hydrocarbon size for each component. We validate this method by comparing calculated chain length data with similar data collected from laboratory gas chromatography data and downhole spectroscopy measurements.
