Unformatted text preview: low to the three NMR MR Scanner shells.
The DMRP processing is a weighted average of NMR porosity and density porosity:
DMRP = W * POROdensity + (1-W) * PORONMR (1) W is a function of gas polarization and HI
The NMR measurement is a discrete measurement and in the case of MR Scanner, each measured volume represents distinct
non-overlapping shells with DOI’s of 1.5”, 2.7” and 4” (Fig. 4). From tool radial responses in Figure 3, one can estimate the
weight to apply to each shell to obtain a similar radial response to density.
Equation 1 is now rewritten as:
DMRP=W * POROdensity + (1-W)*(.4* POROsh-1 +.35* POROsh-4 +.25* POROsh-8)
• (2) W is a function of gas polarization and HI
sh-1, sh-4 and sh-8 are the shells measured at 1.5”, 2.7” and 4” into the formation. Fig. 5 displays the results of equation 2. We can observe two issues:
• In thick zones and a few feet away from the bed boundaries, the DMRP matches core porosity better than the
previously computed density-neutron porosity, but differences remain. A possible cause is the reduced accuracy of
the deeper shell measurements due to lower signal to noise ratio (SNR).
• In thin zones or close to bed boundaries, the computed porosity is clearly wrong. This is a consequence of the
NMR’s lower vertical resolution compared to density. In order to properly compensate for invasion, we need to
combine measurements that investigate the same volume of rock, both radially and vertically.
To address the first point, let’s look closely at the NMR measurement and how the MR Scanner achieves different depth
measurements. The permanent magnet in the tool creates a magnetic field which decreases laterally away from the tool.
Likewise, the Larmor frequency (spinning of the Hydrogen) also decreases laterally from the tool. By tuning the antenna to
lower frequency, we derive a deeper measurement. But there is a drawback, as the magnetic field decreases, less hydrogen is
polarized and as the measurement propagates deeper, the antenna signal travels further and becomes more attenuated. This
attenuation is compensated via a combination of more antenna power and signal amplification but at a cost of reduced SNR.
Fig. 6 displays SNR for different shells in the environmental conditions encountered in the reservoirs in the South China Sea.
One can observe that shell 1 is good even with low stacking but shell 8 has poor SNR. One can improve SNR via increased
stacking but this would result in a decreased vertical resolution, which we wish to avoid.
Via forward modeling, one can examine how reduced SNR affects the measurement.
Fig. 7 displays a simple fluid model representative of this reservoir. This is a gas reservoir at irreducible water saturation, so
the only fluids are bound water, oil-based-mud (OBM) filtrate and gas. The forward modeling computes the tool response for
a given fluid model and borehole environment, then NMR inversion is performed.
Fig. 8 and Fig. 9 display respectively the result...
View Full Document
- Fall '12
- South China Sea, SPE, mr scanner, apparent fluid density