Next generation CSEM inversion and modelling

Water conducts electricity and oil acts as an electrical resistor. This simple truth has made remote electrical resistivity mapping of reservoir rocks from the seabed using marine 3D controlled source electromagnetic (CSEM) an important tool in the explorer's toolbox for the Norwegian Continental Shelf. Although this geophysical method has shown remarkable prediction strength in light of well results, oil companies find it difficult to embrace new technology that requires them to rethink their exploration and decision making workflows. This calls for making new technology "easy" and reducing the uncertainty associated with it. In the case of 3D CSEM, this translates to generating a clearer resistivity image of the earth that can be interpreted by explorers with high confidence. In this project, we have developed 3D modelling and inversion solutions for 3D CSEM data that define a step change in imaging quality and reliability, and demonstrated the improvement these solutions offer over industry standard techniques by successful application to datasets from the Barents Sea and the US Gulf of Mexico. Accurate and fast modelling, i.e. simulation, is key to imaging of 3D CSEM data by inversion. Finite-difference techniques are popular for modelling physical systems due to their robustness and versatility, two properties that are essential to solving geophysical inverse problems. However, when performing 3D CSEM modelling, implementing finite difference techniques can be challenging because the earth model resistivities may vary over several orders of magnitude, e.g. from less than 1 Ohm.m to more than 1 000 Ohm.m. As a result, standard finite-difference formulations may not give satisfactory results in practice. We have finalized research and implementation of an iterative finite-difference frequency domain (FDFD) solver based on a multigrid approach. The approach addresses the numerical problems associated with high earth model resistivities and has been demonstrated to simulate shallow and deep water 3D CSEM surveys effectively. The new iterative FDFD solver is ideal for generating modelling results for earth models containing highly resistive geologies such as salt structures that are often associated with petroleum systems. On the imaging side, we have implemented a next generation 3D inversion software using a powerful mathematical formulation previously thought unfeasible to apply in practice on modern 3D CSEM datasets due to exorbitant computational cost and complexity even on modern high performance computing systems. Important scientific achievements have been made developing strategies for reducing the number of earth model parameters that the inversion needs to solve for, numerical approximations that are motivated by the physics of CSEM, data compression technologies and parallel numerical linear algebra methods for solving large scale optimization problems. The software has been tested successfully on 3D CSEM datasets over several Barents Sea discoveries as calibration targets. The resulting resistivity images show stunning improvements in image accuracy compared to standard commercial CSEM inversion solutions, clearly pinpointing and outlining the discoveries. The developed modelling and inversion innovations form the basis for new geophysical products and solutions that we will be offering to oil companies in Norway and internationally in 2016. We consider this an important milestone in establishing 3D CSEM as a standard tool in offshore exploration and field appraisal programs, helping to reduce finding and field development cost, as well as reducing environmental impact by achieving more with less wells being drilled.

Subsurface resistivity mapping based on Controlled Source Electromagnetic (CSEM) measurements is an attractive technology for exploration as it offers the possibility to distinguish between hydrocarbon and brine bearing prospects where conventional seismi c methods tend to prove inconclusive. This enhances the explorationist´s understanding of the target geology and assessment of pre-drill risk. Using the same principles as electromagnetic well logging, marine CSEM uses the responses of electromagnetic fie lds to probe buried structures from the seafloor. CSEM employs a strong electric field source towed near the seabed to sample electric and magnetic field responses over a large area. From this data CSEM yields information about subsurface resistivity at l arge scales over geometries typical of oil and gas fields. The goal of this R&D project is to increase the value proposition of marine CSEM as an efficient and trustworthy exploration technology. The planned innovations will significantly improve the ima ges of the subsurface that are available prior to drilling, and as a result enable increased drilling success rates for the energy companies, throughout the E&P workflow. The impact of the planned innovations is most prominent in more accurate and effici ent 3D CSEM inversion methods. This will result in less time spent on building initial models, increased confidence in inversion results and optimal value of full-azimuth 3D data. Easier interpretation due to superior inversion results and improved integr ation with seismic data will make EMGS a technology leader in 3D CSEM inversion. The main innovations in this project are the expected research results with respect to numerical algorithms for inversion of extremely large non-linear systems. Although the solution algorithms will be studied in the context of CSEM, the underlying numerical methods will fundamentally impact other communities researching inversion of large non-linear systems.