NMR PUTS A DIFFERENT SPIN ON THE INTERPRETATION OFTIGHT GAS RESERVOIRS

摘要

Great uncertainties in reservoir properties and fluidvolumes can be expected with traditional petrophysicalanalysis of tight-gas sandstone / siltstone reservoirs.Sometimes it’s a struggle to differentiate betweenprospective and non-prospective zones.Nuclear Magnetic Resonance (NMR) technology can beof great help in this type of reservoirs. These toolsdetect directly the types and volumes of fluids withoutrelying on models that require precise knowledge ofe.g. formation water resistivity Rw.The acquisition of NMR data in tight gas reservoirs isnot without challenges. The small amounts of gas in thematrix in combination with the low hydrogen index forgas, reduce the signal to noise ratio of the NMRmeasurements. The hardware and softwareimprovements seen in many of the present day NMRlogging tools have convinced us that NMR can play acrucial role in the interpretation of tight-gas reservoirs.This is a territory that was previously considered offlimits when NMR logging was introduced in the lateeighties / early nineties.Baker’s MREXSM and Halliburton’s MRIL have metand exceeded our expectations in the Pinedale field,where both tools provided us with direct measurementsof gas in the reservoir. Total porosity derived fromNMR runs matched the actual porosity very closely(after the gas-filled pore volume was appropriatelycorrected for the reduced hydrogen index of gas), yetgas-filled porosity from NMR appeared to be too low.Moreover, we found that these different NMR toolsfrom different service companies acquired very similarraw data and differences in the final results could beattributed to different approaches in processing.A lower limit for reliable gas detection via NMR seemsto have been reached by one NMR acquired in one ofour tight Western Canada cretaceous reservoirs. In this,NMR produced again reliable porosity measurements,but was unable to quantify the gas when the totalporosity was lower than 3 pu. At higher porosities, thetool also hinted at the presence of gas, but again thedetermined gas volume was lower than expected,similar to our Pinedale experience.