Tu B3 15 Multi-physics Characterisation of Reservoir Prospects in the Hoop Area of the Barents Sea

Transcription

1 Tu B3 15 Multi-physics Characterisation of Reservoir Prospects in the Hoop Area of the Barents Sea P. Alvarez (RSI), F. Marcy (ENGIE E&P), M. Vrijlandt (ENGIE E&P), K. Nichols (RSI), F. Bolivar (RSI), R. Keirstead (RSI), M. Smith (RSI), H.-W. Tseng (RSI), S. Bouchrara (RSI), L.M. MacGregor* (Rock Solid Images), J. Rappke (ENGIE E&P) Summary We present a case study from the Hoop area of the Barents Sea, in which seismic, well log and controlled source electromagnetic (CSEM) data were integrated within a rock physics framework, to provide a robust assessment of the prospectivity of the area. Combining seismic and CSEM results can resolve the ambiguities that are present when only a single data type is considered. In this example, although seismic data identified potentially hydrocarbon bearing sands, the saturation was uncertain. In this area and at shallow depth, the main focus is on (very) high oil saturations. Adding the CSEM data in this setting allows us to distinguish between high saturations (> 70%), and low and medium saturations (< 50%): it is clear that saturations similar to those observed at the nearby Wisting well (>90%) are not present in this area. However, because of limitations on the sensitivity/recoverability of the CSEM data in this high resistivity environment, it is not possible to distinguish between low and medium saturations. This remains an uncertainty in the analysis.

2 Introduction Accurate pre-drill characterisation of reservoirs is essential, especially in more remote areas where drilling costs are high. In cases where only a single geophysical measurement is considered, uncertainty in this characterisation, particularly in the determination of hydrocarbon saturation can be high (e.g. MacGregor, 2012). There are many documented cases of sub-commercial wells drilled on the basis of seismic alone. Incorporating resistivity, derived from controlled source electromagnetic (CSEM) data, into quantitative interpretation workflows can help resolve ambiguities that remain when only seismic data are considered. Geology of the survey area In this paper we consider the quantitative integrated interpretation of seismic and CSEM data from PL723 in the Barents Sea (figure 1). Since the discovery of the nearby Wisting field in 2013, and the subsequent Hanssen discovery in 2014, there has been increasing interest in the region. The Wisting discovery penetrated nearly 60m of oil charged sand in the Jurassic Sto and Tubaen formations, which form part of the Realgrunnen subgroup. The goal of the study presented here was to characterise the Realgrunnen subgroup in PL723, and assess its prospectivity. Figure 1: Left: PL723 is located in the Northern Barents Sea. Right: Top Realgrunnen structural map showing the area in which pre-stack seismic data were available. Black squares denote CSEM receivers across which the CSEM transmitter was towed. One well, 7324/2-1 (Apollo) was available at the edge of the survey area, to provide calibration. Quantitative interpretation of seismic data The pre-stack seismic were first conditioned (Singleton, 2009) to ensure they were suitable for input to the seismic inversion. A deterministic simultaneous impedance inversion was then used to invert angle stacks for volumes of P- and S-impedance. These formed the basis of the seismic interpretation. The multi-attribute rotation scheme (MARS) is a methodology that uses a numerical solution to estimate a transform to predict petrophysical properties from elastic attributes (Alvarez et al., 2015). The transform is computed from well-log-derived elastic attributes and petrophysical properties, and posteriorly applied to seismically-derived elastic attributes. In this case porosity within the Sto interval was estimated from input volumes of lr and acoustic impedance, using the Apollo well for calibration (figure 2a). A statistical rock physics workflow was then applied to classify the seismic data and determine the most likely seismic facies (figure 2b). The results agree well with the Apollo well, which encountered low saturation gas. An area with a high probability of hydrocarbon was also identified in the Northern part of the area, consistent with a structural closure. However, on the basis of the seismic data alone the saturation within this body, and hence whether it is commercial or not, cannot be determined.

3 Figure 2: Left: Average porosity in the top Realgrunnen to top Fruholmen interval, derived from seismic inversion attributes using the MARS approach of Alvarez et al Note the decreasing porosity towards the Northeast, which correlates with an increase in resistivity in this interval. Right: The most likely litho-fluid facies in this interval. Note that whereas potentially hydrocarbon bearing bodies can be identified, the saturation within these cannot be determined using seismic alone. CSEM sensitivity and recoverability considerations Saturations in the Sto were extremely high at the Wisting well (7324/8-1): greater than 90%, leading to extremely high measured resistivity. This coupled with the shallow depth of the Sto in this area (240m below mudline at Wisting) makes this interval, in principle, an ideal target for CSEM methods. However, when assessing the ability of CSEM data to resolve sub-surface resistivity structure, both the resistivity of the target, and importantly the resistivity of the background structure must be considered. Background resistivity in the Northern Barents is notoriously high even in the shallow sedimentary structure (often in excess of 20 Wm), with a high degree of electrical anisotropy. In this environment it is important to understand not only whether CSEM data are sensitive to a given structure (i.e. does the structure produce a measureable effect on the survey data?), but also whether a CSEM inversion process is likely to be able to recover a given resistivity structure from survey data - the so-called recoverability problem (MacGregor & Tomlinson, 2014). This recoverability question is illustrated in the left panels of figure 3. The upper panels show the vertical resistivity of an anisotropic model constructed to represent the background observed in the survey area. A high resistivity zone is embedded just beneath the top Realgrunnen to represent a hydrocarbon charged target. The bottom panels show the vertical resistivity derived from 2D anisotropic inversion of noise contaminated synthetic data derived from these models. Whereas 6000 Wm 2 can be easily recovered by the inversion, a 3000 Wm 2 target is considerably more challenging and considered to be at the limit of what is likely to be recoverable. The meaning of this recoverability limit in terms of hydrocarbon saturation is shown on the right of figure 3, which illustrates the variation of transverse resistance with hydrocarbon saturation, calculated from Archie s law calibrated at the nearby Wisting well, and assuming a 20m thick hydrocarbon charged sand. Three different porosities are considered covering the range likely to be encountered in the area. The results suggest that for a hydrocarbon accumulation to be clearly imaged by the CSEM data, the saturation must be 60% or greater. Integrated analysis of seismic and CSEM data Before seismic and CSEM data can be integrated, they must be transformed to a consistent scale, and converted into a common domain. Here we convert the seismically derived results to the electrical domain by applying Archie s law to the seismically derived porosity shown in figure 2 for the case of a water wet sand. The resistivity at other saturations can be calculated by fluid substituting in areas

4 which are identified as potentially hydrocarbon bearing in the seismic facies classification. The results are shown in figure 4. Figure 3: Left plots: Synthetic CSEM recoverability analysis. The upper plots show the vertical resistivity of an anisotropic resistivity structure representing the survey area. The lower plots show the results of 2.5D anisotropic inversion of synthetic data. Right plot: Transverse resistance vs hydrocarbon saturation for a reservoir typical of the area. Saturations greater than 60% are required for the reservoir effect to be clearly recovered from CSEM data. Note that there is a scaling factor of approximately ten between the seismic and CSEM results. This scaling has two main causes: firstly the different resolution of the seismic and CSEM techniques. Whereas the seismic data can resolve the reservoir sand, which when wet has a relatively low resistivity, the CSEM data resolve only the bulk resistivity across the reservoir zone including the background resistivity which is considerably higher. Secondly, applying Archie s law, calibrated to the horizontal resistivity measured in a well, to the seismic data gives an estimate of the horizontal resistivity, whereas the CSEM data are sensitive to the vertical resistivity. Taking these effects together provides the necessary scaling factor, which is calibrated at the Apollo well. Comparing the CSEM derived transverse resistance in figure 4 (left) with the water wet seismically derived case, it can be observed that many of the resistivity variations can be explained by variations in porosity (see for example the higher resistivity to the north, and around the centre of area which correlate well between the CSEM and water wet seismically derived cases). There is little evidence for higher resistivity around the area where seismic data predict hydrocarbon charged sands. Indeed when the transverse resistance is calculated from the seismic data assuming 60% hydrocarbon saturation (right panel in figure 4, the limit of recoverability) it is clear that this does not agree with the CSEM data. Figure 5 shows the results of a litho-fluid facies classification incorporating both seismic and CSEM results. Cross-plotting probability of hydrocarbon charged sand derived from seismic data against CSEM derived transverse resistance gives a domain in which commercial hydrocarbon sands are clearly separated from residual hydrocarbons and other confounding resistors, such as carbonates or low porosity sands. Taking cut-offs based on well log calibration and sensitivity analysis and projecting these back into the data, gives the map shown in the right of figure 5. The areas of low porosity reservoir are clearly mapped out, as is the residual gas encountered at the Apollo well. Further areas of residual gas are observed within the structural closure identified in the seismic as being potentially hydrocarbon bearing. Conclusions Combining seismic and CSEM results can resolve the ambiguities that are present when only a single data type is considered. In this example, although seismic data identified potentially hydrocarbon bearing sands, the saturation could not be determined. Adding the CSEM data resolved this ambiguity to an extent: it is clear that saturations similar to those observed at Wisting (>90%) are not present in this area. However, because of limitations on the sensitivity/recoverability of the CSEM data in this

5 high resistivity environment, it is less clear whether the saturation is very low (<20%) or higher (potentially up to 40-50%). This remains an uncertainty in the analysis. Figure 4: Left: Transverse resistance within the top RealgrunnenFruholmen interval, derived from 3D anisotropic inversion of the CSEM data. Middle: Seismically derived transverse resistance at wet conditions. Right: Seismically derived transverse resistance at 60% hydrocarbon saturation. Figure 5: Left: Cross plotting probability of hydrocarbon charge from seismic against transverse resistance from CSEM gives a domain in which commercial and noncommercial saturations can be distinguished. Right: Litho-fluid classification from seismic and CSEM data, suggesting only residual hydrocarbon saturation in the area. Acknowledgements The authors would like to thank Engie and their partners in PL723 for permission to publish this abstract. We thank EMGS for acquiring the CSEM data used in this study. References Alvarez, P., Bolivar, F., Di Luca, M. & Salinas, T., 2015, Multi-attribute rotation scheme: A tool for reservoir property prediction from seismic inversion attributes, Interpretation, 3, SAE9-SAE18 MacGregor, L.M., 2012, Integrating seismic, CSEM and well log data for reservoir characterisation, The Leading Edge, March 2012, MacGregor, L. & Tomlinson, J., Marine controlled source electromagnetic methods in the hydrocarbon industry: A tutorial on method and practice. Interpretation, 2, SH13-SH32 Singleton, S, 2009, The effects of seismic data conditioning on pre-stack simultaneous impedance inversion, The Leading Edge, July 2009, th EAGE Conference & Exhibition 2017