A dual-method nuclear magnetic resonance (NMR) interpretation analysis helped characterize Niger Delta reservoirs.

Even when several development wells have been drilled in certain Niger Delta reservoirs, key questions about the true irreducible water saturation and the point of the actual oil-water contact (OWC) remain. A lack of resistivity contrast may exist from high irreducible water saturation and from wireline sensor response masked by the invasion of drilling fluids such as oil-based mud filtrate (OBMF).

One of a dozen prior wells, Well-Aa was far along in the development of this Niger Delta field and its log response was typical of some Niger Delta reservoir packages. Although the resistivity profile clearly showed an upper reservoir oil column, the fluid type in the sand reservoirs below remained questionable. Conventional log analysis indicated the lower reservoirs to be water bearing, but a lengthy testing program eventually proved the lower reservoirs to contain oil. The resistivity values may result from decreasing formation water salinity with depth, which is not uncommon in Niger Delta reservoirs.

NMR logging of Well Bb

As the same reservoirs found in Well Aa were to be produced in Well Bb, the first objective was to measure irreducible water saturation. The second objective was to determine hydrocarbon saturations and the presence and location of an OWC in Well Bb reservoirs independent of conventional porosity and resistivity measurements. While an increasing formation water resistivity with depth would complicate conventional analysis, a high mud overbalance resulting in deep OBMF invasion would complicate NMR analysis. Pressure data from wireline formation testers in both wells revealed a ~2,500 psi average differential between wellbore hydrostatic pressure and formation pressures near total depth (TD). However, due to tool sticking the formation test program was shortened with only pressure taken at the very top of the main reservoir package. To help overcome complications from OBMF invasion, a dual wait time and dual echo spacing acquisition and analysis approach was used. This approach is possible in a single log pass by multi-frequency NMR tools.

NMR

The time domain analysis (TDA) method provides information relative to hydrocarbon versus water saturation; another method was necessary to differentiate the native oil signal from the OBMF signal in the NMR spectrums. Several longer echo spacing (Te) responses were modeled to predict if it was possible to separate native oil from OBMF T2 signals. From this, a long Te of 3.6 ms was selected for application of the enhanced diffusion method (EDM).

For a Te and a particular field strength, the longest T2 for free water may be calculated, termed T2DW. The 0.9 ms short Te and 3.6 ms long Te T2 spectrum of water, gas, free water, native oil and OBMF were predicted. Using the long 3.6 ms Te allows diffusion to be the dominant mechanism for the resulting T2 spectra. A T2DW of 41 ms was determined as the longest T2 for free water, while the T2 signals for native oil and OBMF were separated by a distinguishable margin. For the 3.6 ms Te, the native oil produced a T2 of 221 ms, while the more diffusive OBMF has a shorter T2 signal of ~118 ms. This method enabled prediction of a T2 separation of the native oil and invading OBMF T2 signals to estimate where native oil was present in the Well Bb reservoirs.

Lower log section

As the Well Bb traversed the same lower reservoirs as Well Aa, the second NMR logging objective was fluid typing via two NMR methods to confirm the presence and location of an OWC. In the lower reservoir, a very similar resistivity profile and nuclear porosity response is observed in Well Bb when compared to Well Aa (Figure 1).

From the conventional data, high resistivity over the upper reservoir interval indicates the presence of oil, but the lower reservoir sections display a profile similar to Well Aa, an inconclusive resistivity profile. The neutron and density data also have a similar profile as in Well Aa and may indicate the presence of hydrocarbons; again the data provides nearly as inconclusive evidence as in the prior well. Calculated formation water salinities also indicated a similar water freshening trend with depth.

Tracks

The EDM is another robust "NMR only" method of fluid identification performed independently of resistivity measurements. A seven-track display (Figure 2) shows all T2 spectra acquired from a simultaneous TDA-EDM log pass along with differential spectrums and also included raw NMR porosities compared to density and neutron data in track 7.

Tracks 1 and 2 contain the T2 spectra from the long 12.99-sec Tw, with short 0.9 ms Te and long 3.6 ms Te, respectively, highlighting any changes in the diffusivity of the fluids. Gamma ray curves are included in the depth track. Tracks 3 and 4 contain the T2 spectra from the short 1-sec Tw with short 0.9 ms Te and long 3.6 ms Te, respectively. Tracks 5 and 6 contain the differential spectra from the short Te and long Te, highlighting any changes in the diffusivity of just the long T1 fluids. The spectra in tracks 1-4 are displayed as variable density log waveforms because the peak locations and shifts in T2 time are the most important features. In contrast, the differential spectra in tracks 5 and 6 are displayed as normal waveforms, similar to the TDA display, as the amplitude variations are important features.

As the dominant mechanism acting on fluids from the 3.6 ms Te portion of the activation set, diffusion causes the water signal to "shift" to a much shorter T2 time than either the native oil or OBMF. The T2DW line present in the long Te spectra tracks 2, 4 and 6 represents the longest T2 time that water can occupy at this particular Te of 3.6 ms; only the signals of oil should be present to the right of this line.
The sum of all the porosity with a T2 longer than the T2DW line in the 3.6 ms Te long Tw track (Track 2) is the oil volume in the EDM saturation analysis, while the total porosity is derived from the NMR-density material balance.

Distinguishing native oil from OBMF is done by comparing the diffusivities of the two hydrocarbons at the long Te echo spacing. During job planning, the OBMF was predicted to be more diffusive than the native oil, and its T2 signal from the long Te measurement would "shift" to a shorter T2 than the native oil. Both OBMF and native oil have a T2 around 1,100 ms in the 0.9 ms long Te spectra, with the OBMF shifting to ~120 ms in the 3.6 ms long Te, and the native oil shifting to ~220 ms in the 3.6 ms long Te spectra. The difference between the two T2 values is still not much, only 100 ms, but due to the logarithmic nature of the spectral display, the T2 shift and amplitude in track 2 around x775 meters where the OWC exists is visually apparent.

Lower log section analysis

Due to the shallow depth of investigation of the NMR sensitive volume, both TDA and EDM may be flushed zone, invaded zone or even true formation interpretations depending on the depth of invasion. A
third form of analysis was also employed: Magnetic Resonance Imaging Analysis (MRIAN). MRIAN uses NMR total porosity from TDA along with true resistivity (Rt) to provide a true formation analysis (Figure 3) in track 5.

The EDM volumetric analysis in track 3 uses the NMR-density method to derive total porosity and the low hydrogen index (HI) fluid volume, and uses EDM to derive the oil volume. With a 12.99-second wait time for the NMR acquisition activation, there should be no under-call of NMR porosity except where reservoir fluid HI is low. A small section of reservoir with green shading represents either a small amount of gas on top of the oil reservoir or high gas-to-oil ratio (GOR) liquid hydrocarbon, different from the reservoir fluids below. This possible gas or high GOR fluid was not "seen" by the TDA method as the differential porosity was too low to allow the gas computation to be reliably applied in this short interval.

The TDA volumetric analysis in track 4 was performed using one set of oil properties to represent the oil, a T1 of 1,500 ms, T2 of 900 ms and HI of 0.95. The OBMF and native oil properties are too similar to allow TDA to make a robust computation differentiating them and is why subjective techniques were used to make this differentiation.

Ideally, the volume of oil computed from TDA would match that computed from EDM, as both computed total oil volume. In this case, the volume computed from EDM is slightly greater than that from TDA. Here, EDM indicates more oil than TDA as some OBMF signal is spread across T2DW and occurs longer than T2DW, where EDM adds it to the calculated oil saturation.

Conclusions

Well design and environmental parameters offered difficult conditions to achieve both sets of logging objectives in Well Bb. Careful pre-job planning, proper data acquisition and dual method data acquisition/analysis allowed NMR logging objectives to be met.

The methods outlined here confirmed the OWC in reservoirs highly invaded by OBMF and having an inconclusive resistivity contrast.

This is a condensed version of SPE paper 85654, presented at the 27th Annual SPE International Technical Conference and Exhibition in Abuja, Nigeria. The original paper is available from ron.balliet@halliburton.com.

Comments