The assessment of time lapse marine controlled-source electromagnetics (CSEM) for dynamic reservoir characterisation
MetadataShow full item record
Marine controlled-source electromagnetics (CSEM) techniques can be used to detect subsurface resistivity anomalies to discriminate hydrocarbon filled reservoir from the water saturated sediments in pre-drill appraisal of seismic anomalies in hydrocarbon exploration. The governing physics of marine CSEM is electromagnetic induction/diffusion therefore it has poor structural resolution. Current time – lapse CSEM feasibility studies for reservoir monitoring assume that the intrinsic limitation of CSEM has little impact on the dynamic fluid discrimination, as more structural constraining information are available at a producing oilfield. However, basic resistivity model is used without rigorous rock physics model, and is thus lacking in dynamic reservoir characterisation. Recent efforts at utilising simulation models combined with rock physics for realistic water-flooding front did not include reservoir management issues. In this thesis, CSEM is presented from the perspective of a reservoir manager, the end – user of this technology. A review of various hydrocarbon production mechanisms and scenarios showed that water – related mechanisms are ideally suited for time lapse CSEM applications as a complimentary tool to seismic in reservoir monitoring because of the resistivity anomaly generated as water replaces hydrocarbon. Channelized turbidite system for the North Sea oilfield model is used, such that the laminar lithological arrangement of sand and shale indicates that a linear arithmetic summation of resistivities of shale and sand will be a good representative of electrical rock physics model. Using this electrical rock physics model, three hydrocarbon provinces are assessed for the technical risk of time lapse CSEM project, in similar manner as done in 4D seismic projects. The North Sea province has highest technical risk, followed by the Gulf of Mexico, while the West Africa province has the least technical risk. A simulation to electromagnetic (sim2EM) workflow is then incorporated into the simulation to seismic (sim2seis) workflow. The sim2EM workflow is used to first examine the impacts of overburden complexity and sea water resistivity stratification on CSEM data. It is observed that the structural impacts are more pronounced on the static CSEM images than on its dynamic images. Then, coupled forward modelling of inline CSEM data and seismic amplitude data from a 3D fluid flow reservoir simulator is performed. The simulator serves the dual purpose of common oilfield in which production is aided by water injection, and of an interpretational constraint involving correlation of CSEM and seismic anomalies with injection and production activities at well locations (here called dynamic well tie). The time-lapse in-line CSEM amplitude change, modelled using dipole 1D, shows linear correlations of 64 to 68% with the change in water saturation. It is more responsive and consistently more linearly related to the change in water saturation than the seismic, despite the possible detrimental effects of reservoir heterogeneity. This is not surprising as seismic is responsive to a combination of changes in saturation and pressure. Coupled interpretation of seismic and CSEM modelled data show that time – lapse CSEM is a definite indicator of water saturation changes. For instance, when seismic softening due to rise in pressure masks increase in water saturation, or when seismic hardening due to pressure drop gives false increase in water saturation. The importance of brine mixing on the acoustic and electrical properties, during secondary and tertiary oil recovery, is examined. The seismic and EM rock physics are adjusted to cater for effective mixed brine resistivity, bulk modulus and bulk density, as functions of temperature and salinity for the injected and formation brines. Modelling of three scenarios of different combinations of injected and formation brines around the world, calibrated with a reference model in which brine properties were kept constant, indicate that EM is more responsive than the seismic, to the brine chemistry. Fluid flow modelling of sea water injection in the North Sea field shows that temperature effect is restricted to the vicinity of injector; while salinity effect travels farther from the injector along the water flooding front. The time-lapse EM could theoretically distinguish extreme brines. For instance, low salinity water injected into oil-wet reservoir with saline formation water; or moderately saline subsurface aquifer water injected into very saline formations of the Middle Eastern carbonates produced between -15 and 7% change in inline CSEM amplitude. In this thesis, 1D dipole forward modelling has generally highlighted values of EM in reservoir monitoring and management. Finally, repeat 3D EM data modelling produced time-lapse amplitude change of 0.3%, which is too small to be detected by the current CSEM acquisition. Thus, high precision EM field sensor will be required for practical application of 4D CSEM to reservoir monitoring. Only about 46% of this small 4D signature is interpretable for the change in transverse resistance of between -800Ωm2 and -1050Ωm2 (equivalent to resistivity reduction of between 13Ωm to 18Ωm). Broad qualitative information about the water flooded areas is provided, but fine detailed information about bypassed oil and early warning of water breakthrough could not be properly imaged.