Integrated Work-flow For Reservoir Characterisation Nguyen Nam - Halliburton Landmark Software & Services, 12 Dec 2010

This article details a reservoir characterisation study in which an integrated work-flow is applied. The objective is to evaluate reservoir quality from a new discovery well and calibrate to seismic to predict hydrocarbon presence away from well. Figure 1 illustrates the original discovery well that was drilled based on a poststack-gas 'bright-spot' conformable with a contour of 2,200 ms at the crest of the structure. In addition to the moderate amplitude extent that is conformable at 2,300 ms contour on the structure high, there is a bright-amplitude anomaly in the flank side requiring further detailed analysis.

A log composition from the discovery well is presented in Figure 2. The changes in the caliper log flags a 'wash out' zone around 2,200 m TVDSS, just above the reservoir zone. The high values of the resistivity log suggest the reservoir is saturated with hydrocarbons. The crossover between NPHI-RHOB confirms that it is a gas-bearing zone. The mud-filtrate invasion in the reservoir zone is detected by the difference between resistivity curves: the deep resistivity at true reservoir condition is higher than the shallow resistivity in the flushed zone. In this study, reservoir properties at true reservoir condition will be estimated. The results are then calibrated with seismic data to understand the reservoir's extents.

Method

Figure 3 illustrates the integrated work-flow diagram, starting with petrophysical analyses to estimate reservoir properties. Next, AVO modelling is applied to calculate gas/oil/brine offset-synthetics cases to perform seismic-towell calibration which yields a good match between modeled synthetics and the real seismic data. Near-mid-far seismic data are then used to generate AVO envelopes and fluid-factor attributes highlighting hydrocarbon presence. Finally, a volume interpretation technique is performed on these AVO attributes to capture 3D geo-bodies of hydrocarbon distribution.

1 - Petrophysical study

The petrophysical analysis estimates the reservoir properties in the flushed zone and at true reservoir condition beyond flushed zone. The typical work-flow is to select a petrophysical model (shaley-sand model) and then decide on petrophysical parameters for this model (shale density, shale-hydrogen index, sand density, etc.). The next step is to evaluate the reservoir intervals to calculate porosity, shale volume, flushed zone saturation, and undisturbed zone saturation. It is assumed that shale volume and porosity are the same in the flushed and the undisturbed zones. Integrating shallow resistivity in an invaded zone and deep resistivity at true reservoir condition, reservoir properties are accordingly calculated (see Figure 4).

2 - AVO modelling

AVO modelling is a combination of fluid substitution and seismic forward modelling. The purpose of this process is to experiment how the 'elastic logs' are, and hence, the seismic are affected by different replacement fluids (brine, oil, and gas). Fluid substitution uses the petrophysical results of reservoir properties to calculate the P-wave, S-wave, and density logs response for brine, oil, and gas scenarios at true reservoir condition. This process uses the standard Gasman Equation to correct the effects of mud invasion on velocity and density logs. The output set of fluid substituted elastic log curves are used as input for synthetic seismic forward modelling to create offset synthetics for comparison with the real seismic data.

Figure 5 shows results of three cases of fluidsubstituted density logs and forward modelling synthetics. The brine-case synthetic is low amplitude, and the oil case is an AVO class-3 with moderate amplitude. The gas-case response is an AVO class-3 with high amplitudes, which ties very well with the real CDP gathers at the well location.

In Figure 6, CDP gathers at the three selected locations are compared with offset synthetics to validate the assumption of two possible fluid contacts at 2,200 ms and 2,300 ms on structure high. Below 2,300 ms contour, the first location gather shows amplitude decreases along the offset, which is similar to the brine-case synthetic. The second location, which is in between two contours, matches with the oil case of moderate AVO class-3, and the third location above contour of 2,200 ms is similar to the gas case of AVO class-3 high amplitude. The results suggest the assumption of fluid contacts makes sense. The next step is applying AVO attributes to map out the similar oil/gas AVO class-3 response in the reservoir throughout the whole 3D seismic survey.

3 - AVO attributes

AVO attributes, including volumes intercept, gradient, AVO envelope, and fluid factor, are generated using near-mid-far partial offset stacks. This technique has proven to be robust, even in noisy data areas. Near-mid-far angle/ offset partial stacks are very common dataset available for every interpreter, which are key inputs for AVO analysis when pre-stack gathers aren't available or in the noisy data areas where partial stack traces commonly have better signals to noise ratio than the raw gather.

As shown in Figure 7, reservoir horizon is required to be interpreted on each volume individually to capture precise amplitude changes at the reservoir interval. A common technique is to pick a seed horizon on full stack and then let the software performing auto-pick onto the three partial stacks. These three horizons are used to extract the amplitude from the partial stacks for AVO characterisation, as shown in Figure 8. The area of the amplitude increases versus offset (AVO class-3 - oil and gas cases) limits on the structure high above the contour of 2,300 ms, while the bright amplitude anomaly on the flank responds to decreasing amplitude versus offset, which is similar to the brine case.

In Figure 9, the horizon-based attributes are also used to calculate on-the-fly AVO envelope attributes to highlight the AVO class-3 distribution. In addition, the fluid factor attribute is calculated on-the-fly on the selected sections where high value of amplitude typically indicates hydrocarbon presence. The colour code used for the fluid factor are red for gas (high amplitude), green for oil (moderate to high amplitude), and blue represents brine (low-amplitude). It is observed that high-amplitude cut-off for gas and oil cases are consistent with the possible fluid contacts at the 2,200 ms and 2,300 ms contours. The final volumes of AVO envelope and fluid factor attributes are generated to map out hydrocarbon presence in the reservoir throughout the entire 3D seismic survey.

4 - Volume interpretation

Volume interpretation is a key component of this work-flow. In the early stage, it was applied to map out the seed horizon of top reservoir on a full-stack dataset. A volume interpretation technique is performed to capture the 3D geobodies of hydrocarbon distribution from AVO attribute volumes. The area and volume of the pay reservoir can be estimated directly from the interpreted 3D geo-bodies. As shown in Figure 10, a similar colour code is applied: the red coloured geo-body represents the gas extent, and the green coloured geo-body illustrates the oil extent in the reservoir. The geo-bodies of oil and gas extents are conformable with the potential fluid contacts in the reservoir.

Conclusion

The reservoir characterisation study applying integrated work-flow is effective for evaluating reservoir properties, improving seismic welltie error owing to mud-filtrate invasion, and understanding the hydrocarbon presence away from the well. Results from the study show a good match between modeled synthetics and the real seismic data. Good porosity (PHI ~25%) and high-saturation (Sw ~10%) hydrocarbonbearing sands respond similar to that of AVO class-3, but brine sands exhibit a dimamplitude response with offset. Hydrocarbon AVO anomalies are conformable with structure which successfully verifies the potential fluid contacts in the reservoir.

Acknowledgement

I wish to thank Halliburton for their sponsorship and permission to publish this article. The integrated work-flow in this article uses Landmark R5000 Technologies, including OpenWorks®, PetroWorks®, SpecDecomp®, GeoProbe®, Well Seismic FusionTM, ProMAX®, PostStackTM family, and DecisonSpace® Desktop applications 

comments powered by Disqus