Repeated seismic surveying and data analysis (time-lapse seismic) is a key technique in monitoring and optimizing the depletion of hydrocarbon reservoirs and it will play a similar key role in monitoring future CO2 and energy storage projects. Traditional time-lapse seismic acquisition for reservoir monitoring typically requires close repetition of the acquisition geometry, i.e., source and receiver positions. We propose a paradigm change in time-lapse seismic acquisition and processing that will benefit both traditional hydrocarbon monitoring but also the rapidly growing fields of CO2 and energy storage. We aim at eliminating the need for strong repeatability. Thanks to advances in data processing capability, we aim to relax geometric repeatability, thereby significantly reducing the costs and environmental impact of acquisition while improving data quality for time-lapse monitoring. This will contribute to optimum hydrocarbon reservoir production and can be regarded as an enabler for large-scale future CO2 and energy storage projects. Since widespread field data is unavailable, and very costly to acquire, the project will primarily focus on extensive modelling and processing synthetic data to evaluate how well different combinations of acquisition and processing techniques can resolve the known time-lapse signal while accounting for uncertainties in reservoir behavior. Strategies for optimum acquisition design will be implemented and tested. We will adapt and use value-of-information concepts to identify the best trade-off between added value and cost, including the contribution to reducing CO2 emissions related to time-lapse seismic acquisition. Although synthetic experiments will be the primary focus of the R&D, available field data from the Sleipner CO2 monitoring project will be used to validate some of the implementations and findings from the synthetic exercise toward the end of the project.
The CLEAN project seeks to develop innovative approaches to address the challenge of enabling time-lapse surveys using any acquisition technique that provides sufficient illumination to image the reservoir (or a specific target within the reservoir) to the required resolution, without the need to closely reproduce the acquisition geometry of previous surveys as per current standard. A realistic 3-D model of a hydrocarbon reservoir and its overburden will be built. Using visco-acoustic finite-difference modelling in this model, seismic data will be synthesized for different candidate acquisition geometries. Reservoir properties will then be modified to simulate the effects of production of oil or gas from a hydrocarbon reservoir, and of injection of CO2 into the reservoir and subsequent plume evolution. Further synthetic seismic datasets, representing monitor surveys, will be modelled with different acquisition geometries which have improved survey efficiency and/or reduced impact. The differences between acquisition geometries lead to substantial deviations from current state-of-the-art survey repeatability. Survey combinations will be processed using advanced imaging methodologies to establish how well such deviations can be mitigated in processing, allowing to relax constraints on repeatability in acquisition, and how well different combinations of acquisition and processing techniques can resolve the known time-lapse signal while accounting for uncertainties in reservoir behaviour. Based on these results, strategies for optimum acquisition design will be implemented and tested. We will adapt and use Value Of Information (VOI) concepts to identify the best trade-off between added value and cost, including the contribution to reducing CO2 emissions related to time-lapse seismic acquisition. The implementation and findings will be validated using synthetic and real data.