Th C3 08 Capturing Structural Uncertainty in Fault Seal Analysis A Multi-throw Scenario Approach M. Giba* (DEA) Summary An intrinsic challenge of fault seal analyses is the large uncertainties that have to be considered. Parameters used in fault seal analyses - sedimentary architecture, Vshale distribution and fault geometry all have large uncertainties. Commonly, uncertainty in sedimentary architecture and Vshale distribution are addressed by building multiple 3D property models (Freeman et al., 2010). Uncertainty in fault geometry, however, is often not treated adequately since only one 3D structural framework model or one fault interpretation is mostly being utilised. In this study, a grid modification tool is used to capture structural uncertainty to assess likely throw scenarios and their impact on fault weak points. This multi-throw scenario analysis was performed on a fault bounded prospect in the Gulf of Suez, Egypt. Seismic imaging in this area is poor and the structural uncertainty of the seismic interpretation is therefore high. Juxtaposition analysis with the original seismic interpretation showed a good lateral seal, whereas a major fault weak point was detected at 70% and 60% throw scenarios. In consequence, the structural uncertainty was found to be too high and the prospect was discarded. This study therefore shows how valuable testing of structural uncertainty is to make better choices in prospect assessment.
Introduction An intrinsic challenge of fault seal analyses is the large uncertainties that have to be considered. Parameters used in fault seal analyses - sedimentary architecture, Vshale distribution and fault geometry all have large uncertainties. Commonly, uncertainty in sedimentary architecture and Vshale distribution are addressed by building multiple 3D property models (Freeman et al., 2010). Uncertainty in fault geometry, however, is often not treated adequately since only one 3D structural framework model or one fault interpretation is mostly being utilised. In this study, a grid modification tool is used to capture structural uncertainty to assess likely throw scenarios and their impact on fault weak points. Standard fault seal analysis approach The standard approach in 3D fault seal analyses is to use a single structurally QC ed fault network and to either build a 3D grid (e.g. in Petrel) or to directly project up-scaled parameters from well logs onto fault surfaces (e.g. in TrapTester). In both approaches, accounting for uncertainties is mainly related to petropysical properties of the grid or the well logs. In good practice studies, different Vshale scenarios are used and different facies model scenarios are tested (e.g. James et al. 2004). Accounting for uncertainty is therefore often restricted to petrophysical parameters and sand/shale distributions. Fault weak points are, however, fundamentally related to the geometry of the fault. Yet still, accounting for structural uncertainty is often neglected. Advances in 3D grid building have allowed fast and easy modifications of 3D grids. For instance, with the throw modifier (implemented in Petrel) it is possible to automatically change throws by bending the grid adjacent to faults without modifiying the entire grid. As a result, multiple fault throw scenarios can be tested with the same grid and the time-consuming approach to build several different grids to test structural uncertainty is not required. The throw modifier is mainly designed for high quality 3D seismic cases in which fault geometry and fault throws can be well determined - with the exception of small-scale fault zone complexities which are below seismic resolution (e.g. multiple slip surfaces, normal drag, etc). Multiple slip surfaces and normal drag close to faults may significantly reduce the effective throw and therefore generate altered juxtapositions of reservoir units which have an important impact on fluid flow behaviour (Fig. 1) (e.g. Wood et al. 2015). Incorporation of sub-seismic structural uncertainty into standard fault seal analysis is therefore highly important since the overall fault geometry has a first order impact on the location of fault weak points. To assess small-scale fault zone complexities a standard grid populated with petrophysical parameters has to be built around the analysed fault. Using the throw modifier, the grid is bent close to the fault (Fig. 2) to reduce throws by given values (e.g. 90%, 80%, 70%, etc.). In this way, several throw scenarios can be tested to assess structural uncertainty in high quality seismic datasets. Assessing structural uncertainty in poor quality seismic datasets With poor quality seismic data, structural uncertainty can be large since horizon picks may be less certain due to poor imaging (Fig. 1). This study uses the same approach to test structural uncertainty in an area of poor seismic data as described for high quality seismic (Fig. 3). In contrast to the above described approach, the geometry and throw of the evaluated fault is only poorly resolved. The key uncertainty is therefore structural and a standard fault seal analysis with multiple scenarios of petrophysical parameters and facies models does not address the main challenge. The fault seal analysis therefore has to focus on the structural uncertainty first.
Figure 1 Schematic illustration of structural uncertainty in high and poor quality seismic data cases. Structural uncertainty in high quality seismic data cases is mainly related to fault zone complexities which are below seismic resolution (e.g. multiple slip surfaces, normal drag, etc ). In poor quality seismic datasets structural uncertainty can be large and is mainly related to seismic interpretation uncertainties due to poor imaging. Figure 2 Description of multi-throw scenario approach. (A) The 3D grid is bent close to the fault to change throw values. (B) A schematic section across a fault is shown with original horizon interpretation in purple and modified horizon in gray. For high quality seismic cases, this methodology is used to investigate small-scale fault zone complexities. (C) In poor quality seismic cases, the same methodology can be used to test different throw interpretations and their impact on fault weak points. Case study a simple multi-scenario approach The method was applied to a prospect in the Gulf of Suez, Egypt. This fault-bounded prospect is located adjacent to an oil field which has been producing from fault-bounded tilted blocks for several decades. Fault seal issues have been known to represent a key risk in this area. The main challenge in this area is that target reservoirs (> 4000 m depth) are located beneath a thick and continuous salt layer. At target depth seismic imaging is therefore extremely poor (Fig. 3). The main prospect bounding normal fault has an interpreted maximum throw of 800 m and is oriented approximately N-S. Reservoir properties and distributions are well-understood with laterally continuous interbedded sands and shales.
Figure 3 Section from Gulf of Suez showing the quality of seismic imaging at target depth which illustrates the level of uncertainty associated with seismic interpretation. Vertical scale is in ms. Fault length vs displacement plot modified after Kim and Sanderson (2005) on the right. Blue star shows interpreted maximum fault throw (800 m) for interpreted fault length (9 km). Throw value range (80-500 m) inferred from Kim & Sanderson data for given fault length is shown in red. 150% throw scenario is used for a fault length of 30 km, i.e. fault in question is interpreted as a single segment of a large fault zone. Figure 4 Fault plane views of 100% throw scenario (top) and 60% throw scenario (bottom). Blue line indicates top reservoir horizon (bold in FW and dashed in HW). Dashed horizontal red line indicates location of fault bounded prospect. Red circle shows area of fault weak point.
A grid populated with lithological descriptors derived from two wells (1,5 and 2 km away from the fault) was built around the fault. For the multi-throw scenario analysis, the 3D grid was modified using the throw modifier in Petrel (150%, 100%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 10% throw scenarios) to assess likely weak points on the fault plane (Fig. 4). Results Juxtaposition analysis using the original seismic interpretation (i.e. 100% throw scenario) shows sealing behaviour on the main fault. Tight carbonate layers in the hangingwall are juxtaposed with the main sandstone target in the footwall (Fig. 4, top). A tilted fault block prospect with a valid lateral seal can be postulated from this throw scenario. Using different throws, however, yields dramatically altered results; at 70% and 60% throws, a large sand-sand juxtaposition area can be observed which represents a key fault weak point (Fig. 4, bottom). This juxtaposition of a sand layer in the hangingwall with the target reservoir in the footwall, was also seen in the 100% throw scenario but here, it does not represent a fault weak point since the trap at 100% throw is located structurally higher. Conclusions A multi-throw scenario analysis was performed on a fault bounded prospect in the Gulf of Suez, Egypt. Seismic imaging in this area is poor and the structural uncertainty of the seismic interpretation is therefore high. In order to better assess this uncertainty, several throw scenarios were run. Juxtaposition analysis with the original seismic interpretation showed a good lateral seal, whereas a major fault weak point was detected at 70% and 60% throw scenarios. Seismic data was not able to clearly resolve the best throw scenario. In consequence, the structural uncertainty was found to be too high and the prospect was discarded. This study therefore shows how valuable testing of structural uncertainty is to make informed decisions and better choices in prospect assessment. References Freeman, S.R., Harris, S.D., Knipe, R.J. [2010] Cross-fault sealing, baffling and fluid flow in 3D geological models: tools for analysis, visualization and interpretation. From: Jolley, S. J., Fisher, Q. J., Ainsworth, R. B., Vrolijk, P. J. & Delisle, S. J. (eds) Reservoir Compartmentalization. Geological Society, London, Special Publications, 347, 257-282. James, W. R., Fairchild, L. H., Nakayama, G. P., Hippler, S. J., Vrolijk, P. J. [2004] Fault-seal analysis using a stochastic multifault approach. AAPG Bulletin, V. 88, No. 7, PP. 885-904. Kim, Y-E., Sanderson, D.J. [2005] The relationship between displacement and length of faults: a review. Earth-Science Reviews, 68, 317-334. Wood, A.M., Paton, D. A., Collier, R. E. LL. [2015] The missing complexity in seismically imaged normal faults: what are the implications for geometry and production response? From: Richards, F.L., Richardson, N. J., Rippington, S. J., Wilson, R. W., Bond, C. E. (eds) 2015. Industrial Structural Geology: Principles, Techniques and Integration. Geological Society, London, Special Publications, 421, 213-230.