The Enhanced Diffusion Method8, 10 (EDM) types and quantifies oils with viscosities ranging between 1 and 50 cp. The EDM relies on diffusion contrasts to differentiate the fluids. Use of a properly selected long TE enhances the diffusion effect during echo-data acquisition and allows water and oil to be separated on a T2 distribution generated from the logged data. The EDM can use CPMG measurements acquired with the following activations:
•
standard-T2 logging with a long TE•
dual-TE logging with a single long TW•
dual-TW logging with a single long TEAn understanding of the EDM principle depends on an understanding of the factors affecting the relaxation rates of water and oil in rock pores. If two echo trains are acquired during dual-TE logging, the resulting TEL and TES T2 distributions will both include water and oil signals. TEL can be selected so that oil and water signals will separate on the TEL T2 distribu- tion, thereby providing a quick-look EDM result at the wellsite.
As discussed in Chapter 3, the relaxation rates of fluids in rock pores observed with CPMG measurements are related to bulk, surface, and diffusion relaxation mechanisms:
128 Answer Products Derived from MRIL Stand-Alone Analysis Chapter 6 Figure 6.15—This log
contains results from the application DIFAN to MRIL data from an Indonesian well. Track 1 includes conventional gamma ray, spontane- ous potential (SP), and caliper curves. Track 2 presents deep, medium, and shallow resistivity data and MRIL perme- ability. Track 3 contains the long-TE T2 distribu- tion. Track 4 contains the short-TE T2 distribu- tion. Track 5 displays answer products from the DIFAN calculations.
Answer Products Derived from MRIL Stand-Alone Analysis 129 Chapter 6 diffusion 2 surface 2 bulk 2 CPMG 2
1
1
1
1T
=
T
+
T
+
T
(6.5)The T2 measured by a CPMG sequence is smaller than the T2 calculated for any of the three relaxation mechanisms. Because T2bulk is always much larger than T2surface and T2diffusion, T2bulk can be ignored in practical applications. If T2surface is smaller than T2diffusion, then surface relaxation dominates the observed relaxation. Otherwise, diffusion relaxation dominates.
The diffusion effect can be manipulated to an extent by the choices of MRIL-tool operational parameters. In particular, the strength of the field gradient G is a function of the operating frequency and tool type, and the inter-echo spacing TE can be selected by the logging engineer at the wellsite. G and TE can be chosen so that the diffusion mechanism dominates water relaxation and thus so that the upper limit of T2 for water in rock pores is T2diffusion,w. This upper limit, designated T2DW, is given by
(
)
[
2]
2 12 CD G TE
TDW = w γ (6.6)
Therefore, T2DW constitutes the absolute upper bound for the measured T2 of water, and all T2 relaxation times associated with water will be less than or equal to T2DW.
The T2 of oil in water-wet rock pores is determined by both bulk and diffusion relaxation and
is given by oil diffusion, oil bulk,
T
T
T
2oil1
21
21
=
+
(6.7)The selection of TE and G can be often further refined so that
T2DW << minimum {T2oil values expected over the formation}
(6.8) In reality, because of noise effects, TE and G are usually selected so that
2T2DW << minimum {T2oil values expected over the formation} (6.9) Thus, the existence of a signal on the T2 distribution longer than T2DW unambiguously indi-
cates the presence of oil in the formation. Fig. 6.16 shows how this observation is used to recognize pay on EDM log displays.
The use of the EDM is quite straightforward. A T1 contrast is not necessary, and depending on oil NMR properties and the job objective, EDM data processing can be done in either T2 domain or time domain. If the EDM objective is to discriminate pay from non-pay, then a single CPMG measurement with long TW (for full polarization) and long TE (for diffusion enhancement) is sufficient. Thus, standard-T2 logging with a long TE can be used. If the EDM objective is to quantify fluids in a pay zone, then dual-TE logging will be required. The short-
TE measurements will provide correct MPHI and BVI. If the T2 contrast over the zone of interest is not expected to be large enough to separate the T2 values of water and oil, then dual-TW logging with a single, long TE may be required to obtain data for TDA processing. Thus, job planning is critical for EDM success.
Appendix: TDA Mathematical Model
11An understanding of how TDA determines the oil- and gas-filled porosity from the differen- tial spectrum can be based on discussions in Chapter 3. In particular, Eq. 3.17 indicates that
130 Answer Products Derived from MRIL Stand-Alone Analysis Chapter 6 Figure 6.16—In this log
display, the T2 distributions
from TE = 1.2, 3.6, and 4.8 ms are shown in Tracks 3, 4, and 5 respectively. On the EDM results in Track 4 and 5, significant energy to the right of the T2DW line
indicates obvious oil zones. Also note the increased separation in Track 5 because of the increased TE.
the amplitude of an echo train acquired with a CPMG sequence for a water-wet rock satu- rated with water, oil, and gas can be given by Eq. 6-A.1, when both oil and gas are assumed to have single T2 values.
M(t) = Σ[M0iexp(-t/T2i)] + Moilexp(-t/T2oil) + Mgasexp(-t/T2gas) (6-A.1) When polarization effects are considered, M0i, Moil, and Mgas can be expressed as
M0i =M0i(0)[1- exp(-TW/T1i)]
Moil =Moil(0)[1- exp(-TW/T1oil)]
Mgas =Mgas(0)[1- exp(-TW/T1gas)] (6-A.2) The amplitudes of echo trains for TWL and TWS are then given as
MTW (t)=Σ{M0i(0)[1- exp(-TWL/T1i)]exp(-t/T2i)} + Moil(0)[1- exp(-TWL/T1oil)]exp(-t/T2oil)
+ Mgas(0)[1- exp(-TWL/T1gas)]exp(-t/T2gas) (6-A.3)
om001616
Answer Products Derived from MRIL Stand-Alone Analysis 131 Chapter 6
and
MTW (t)=Σ{M0i(0)[1- exp(-TWS/T1i)]exp(-t/T2i)} + Moil(0)[1- exp(-TWS/T1oil)]exp(-t/T2oil)
+ Mgas(0)[1- exp(-TWS/T1gas)]exp(-t/T2gas) (6-A.4) ∆M(t)=MTWL(t) - MTWS(t)
= Σ{M0i(0)exp(-t/T2i) [exp(-TWS/T1i)- exp(-TWL/T1i)] } + Moil(0)exp(-t/T2oil) [exp(-TWS/T1oil)- exp(-TWL/T1oil)]+
Mgas(0)exp(-t/T2gas)[exp(-TWS/T1gas)- exp(-TWL/T1gas)] (6-A.5) Polarization functions are then defined for water (∆αwi), oil (∆αo), and gas (∆αg) as follows:
∆αwi =[exp(-TWS/T1i)- exp(-TWL/T1i)] (6-A.6)
∆αo =[exp(-TWS/T1oil)- exp(-TWL/T1oil)] (6-A.7)
∆αg =[exp(-TWS/T1gas)- exp(-TWL/T1gas)] (6-A.8)
Eq. 6-A.5 subsequently becomes
∆M(t) =Σ[M0i(0)exp(-t/T2i) ∆αwi ]+ Moil(0)exp(-t/T2oil) ∆αo + Mgas(0)exp(-t/T2gas) ∆αg
(6-A.9) If TWS was selected to fully polarize the protons of water in rock pores, then ∆αwi ≅ 0. This
condition allows the difference of the two echo trains to be written as
∆M(t) = Moil(0)exp(-t/T2oil) ∆αo + Mgas(0)exp(-t/T2gas) ∆αg (6-A.10) A differential porosity function is then defined by
∆φ(t) = φ*oilexp(-t/T2oil) + φ*gasexp(-t/T2gas) + noise (6-A.11) where
noise = noise during CPMG measurement of the two echo trains
∆φ = difference in hydrocarbon-filled porosity obtained from the echo trains φ∗oil = apparent oil-filled porosity obtained from the difference of the two
echo trains
φ∗gas = apparent gas-filled porosity obtained from the difference of the two echo trains
132 Answer Products Derived from MRIL Stand-Alone Analysis Chapter 6 Finally, the apparent porosities are related to the true porosities (φoil and φgas) by
φ*oil = [Moil(0)/M100%(0)] ∆αo= φoilHIoil∆αo (6-A.12) φ*gas= [Mgas(0)/M100%(0)] ∆αg = φgasHIgas∆αg (6-A.13) where
M100%(0) = amplitude of CPMG echo train at time zero obtained from an MRIL water- tank (i.e., 100% porosity) calibration
HIoil = hydrogen index of oil
HIgas = hydrogen index of gas Therefore,
•
If the values of T2oil and T2gas are known at reservoir conditions, then Eq. 6-A.11 can be used to calculate the apparent porosities, φ*oil and φ*gas.•
If the values of T1oil, T1gas, HIoil, and HIgas are also known, then Eqs. 6-A.12 and 6-A.13 can be used to calculate the true porosities, φoil and φgas.The actual TDA procedure includes the following steps: 1. Acquire two echo trains with dual-TW activation.
2. Estimate the bulk properties (T1,T2,and HI) of oil and gas at reservoir conditions (e.g., temperature, pressure, and oil viscosity).
3. Subtract the echo trains from one another.
4. Search T2 for gas and oil, and search T1 for oil at reservoir conditions. 5. Calculate the apparent porosities (φ*oil and φ*gas)using Eq. 6-A.11.
6. Calculate the true porosities (φoil and φgas) using Eqs. 6-A.12 and 6-A.13, the bulk properties estimated in Step 2, and the apparent porosities found in Step 5. (Note that the T1 values estimated in Step 2 or measured by a triple-TW12 activation are used to calculate the oil and gas polarization functions.)
7. Calculate water porosity and effective porosity.
The following summarizes several TDA assumptions that were discussed earlier in this appendix:
•
In Eq. 6-A.1, each of the oil and gas signals exhibits a single-exponential decay. This single-exponential decay is a reasonable approximation for the decay of gas and many low-viscosity oils.•
In Eq. 6-A.10, TWS should be selected so that the water protons are fully polarized. Otherwise, a water-polarization correction is needed, and the analysis process will be more complicated.Answer Products Derived from MRIL Stand-Alone Analysis 133 Chapter 6
•
In Eq. 6-A.11, ∆φ, the difference between the porosities derived from the two individual echo trains is dependent on both the true porosity of the rock and the T1 contrast between water and light hydrocarbons. If ∆φ is not sufficiently large, say ∆φ < 1.5 p.u., then fitting the difference signal with a single- or two-exponential function may be difficult because of the noise level under which the MRIL data are acquired.•
A significant T1 contrast must exist between water and light hydrocarbons.•
Gas and oil have significantly different T2 values to allow separate signals to be recognized.These assumptions are generally valid for high-porosity, water-wet reservoirs containing light hydrocarbons (gas and light oil). In such reservoirs, TDA should be possible, provided that TWL and TWS are carefully selected to magnify the T1 contrast between water and light hydrocarbons. Thus, job planning is critical to the success of TDA. TDA requires only MRIL data to provide porosity, permeability, and hydrocarbon typing: no other conventional log data are needed.
References
1. Prammer, M.G., et al., 1995, Lithology-independent gas detection by gradient-NMR logging, SPE 30562, 1995 SPE Annual Technical Conference and Exhibition
proceedings, v. Ω (Formation evaluation and reservoir geology), p. 325–336.
2. Akkurt, R., Prammer, M., and Moore, A., 1996, Selection of optimal acquisition parameters for MRIL logs, paper TT, 37th Annual SPWLA Logging Symposium
Transactions, 13 p. Later published in 1996 in The Log Analyst, v. 37, no. 6,
p. 43–52.
3. Akkurt, R., Moore, A., and Freeman, J., 1997, Impact of NMR in the development of a deepwater turbidite field, paper SS, in 38th Annual SPWLA Logging Symposium
Transactions, 14 p.
4. Akkurt, R., et al., 1995, NMR logging of natural gas reservoirs, paper N, 36th
Annual SPWLA Logging Symposium Transactions, 12 p. Later published in 1996, as
Akkurt, R., et al. in The Log Analyst, v. 37, no. 5, p. 33–42.
5. Moore, M.A., and Akkurt, R., 1996, Nuclear magnetic resonance applied to gas detection in a highly laminated Gulf of Mexico turbidite invaded with synthetic oil filtrate, SPE 36521, 1996 SPE Annual Technical Conference and Exhibition
proceedings, v. Ω (Formation evaluation and reservoir geology), p. 305–310.
6. Mardon, D., et al., 1996, Characterization of light hydrocarbon-bearing reservoirs by gradient NMR well logging—a Gulf of Mexico case study, SPE 36520, 1996 SPE
Annual Technical Conference and Exhibition proceedings, v. Ω (Formation evalua-
tion and reservoir geology), p. 295–304. Also published in 1996 in condensed form in Journal of Petroleum Technology, v. 48, no. 11, p. 1035–1036.
7. Mardon, D., et al., 1996, Experimental study of diffusion and relaxation of oil-water mixtures in model porous media, paper K, 37th Annual SPWLA Logging Symposium
134 Answer Products Derived from MRIL Stand-Alone Analysis Chapter 6 8. Akkurt, R., et al., 1998, Enhanced diffusion: expanding the range of NMR direct
hydrocarbon-typing applications, paper GG, 39th Annual SPWLA Logging Symposium Transactions.
9. Mardon, D., Prammer, M.G., and Coates, G.R., 1996, Characterization of light hydrocarbon reservoirs by gradient-NMR well logging, Magnetic Resonance Imaging, v. 14,
nos. 7 and 8, p. 769–777.
10. Akkurt, R., et al., 1998, Determination of Residual Oil Saturation Using Enhanced Diffusion, SPE 49014.
11. Xiao, L.Z., 1998, NMR imaging logging principles and applications (in Chinese), Science Press, Beijing. 328 p.
12. Hou, L., et al., 1999, Enhanced NMR logging methods for accurately measuring volumes of gas and light oil in hydrocarbon reservoirs, SPE 56769, prepublication print, 14 p.
Answer Products Derived from MRIL Combinations with Other Logs 135 Chapter 7
Answer Products
Derived from
MRIL
Combinations
with Other Logs
Chapter 7
As stated in Chapter 6, MRIL stand-alone analysis, such as Time Domain Analysis and Diffusion Analysis, provides interpretation of the invaded zone because of the shallow depth of investigation of MRIL measurements. If MRIL data are combined with other logs, analysis can furnish even more information about the reservoir. For example, a combination of MRIL and deep-resistivity data provides a complete analysis of the fluids in the virgin zone. MRIAN is one of the interpre- tation models that uses this data combination.1 This chapter contains a complete discussion of MRIAN and its applications.
If MRIL data are combined with quad-combo data (neutron, density, sonic, and resistivity), then the combination can provide critical information for well completions. For example, when used in a model called StiMRILTM, such data can generate information about rock properties, formation lithology, and formation permeability. The StiMRIL model and its application in stimulation optimization are discussed in this chapter.