Characterizing the
pore and fluid distribution is critical for
evaluating the reservoir potential of new areas. Nuclear magnetic
resonance (NMR) is considered as an experimental method capable of
full-scale characterization of pore characteristics. However, the T
2 spectrum of a saturated sample is affected
by a combination of sample and experimental parameters, and it is
important to confirm whether the T
2 spectrum
fully reflects the sample pore information. Eight tight sandstone
samples from the Julu area were selected for thin section identification,
mercury intrusion porosimetry (MIP), NMR, NMR cryoporometry (NMRC),
and centrifugation experiments to critically analyze the applicability
of the NMR results. Two methods, the similarity method and Kozeny’s
equation method, were used to calculate surface relaxivity, a critical
parameter for converting NMR T
2 signals
into pore information. The discussion focuses on the applicability
of the calculated surface relaxivity and the phenomenon of T
2 signal changes in a short relaxation range
after centrifugation. The main results are as follows: The surface
relaxivity values calculated using the different methods differed
significantly. The surface relaxivity calculated using the same method
reflected the relative magnitude of the true surface relaxivity of
the samples. For the samples with large surface relaxivity, there
may be partial misses of the short relaxation signal, the NMR porosity
was smaller than the gas-measured porosity, there was a variation
in the T
2 spectrum in the short relaxation
range after centrifugation, and the calculated surface relaxivity
was small. The surface relaxivity calculated using Kozeny’s
equation was nearly accurate, but perhaps smaller than the true value.
The T
2 spectra mainly reflected macropore
information. This study suggests that PSDs converted from T
2 spectra of saturated samples should be judged
with relative caution rather than solely based on the peak or range
correspondence between the two curves, and the minimum centrifugal
radius can be used as a constraint.