Oral Presentation Sub22 Conference

Carbon capture and storage projects in deep reservoirs have significant monitoring obligations associated with regulatory needs and social licence to operate. Downhole methods are becoming increasingly attractive because of their low surface footprint and ease of stakeholder engagement. When several wells are available, monitoring from VSP-based seismic methods is possible using downhole DAS fibre optics, and even lower footprint seismic methods are possible using surface orbital vibrators and DAS.  Using well completions in the reservoir interval, direct water injection testing methods are also possible with multi-well pressure gauges, a method known as pressure tomography (PT). PT can be employed in a time-lapse setting with surface requirements comprising only pumps, water facilities, and instrumentation. Inversion of either kind of data is a viable method of tracking gas saturation in the reservoir interval.  

A more ambitious goal is joint inversion of both data types in a consistent model.  This can be framed as a multivariate Bayesian inverse problem wherein gas saturation couples via acoustic impedance to seismic amplitudes, and via fluid compressibility to water-test pressure responses using Darcy's law.  We have developed a joint seismic-PT inversion framework using a single-phase effective-media flow model for the pressure, effective media and wavelet convolutional models for the seismic reflectivity, and use of adjoint methods for the derivatives.  Figure 1 shows applications on experimental data from the 2020-2021 Otway stage 3 project, using wells 2-7, where 15kT of CO2-rich gas was injected in well 3 in the Paaratte formation at ~1500m depth, with monitoring via PT and seismic methods (Fig.1(a)). Monitoring shows the plume goes updip to the east, and the the legacy gas plume around well 2 is has to be remobilised to match the pressure data in Fig.1(c). The spatial distribution of saturation explains all data types and known injection volumes.