CarbonWatch: Rock Physics-Based Machine Learning Solution for End-to-End Monitoring of CO2 Storage Site

C53-21d-271370CCS is a key technology for the mitigation of greenhouse gas emissions. Finding suitable geological storage sites, demonstrating effective CO2 containment, and establishing long-term monitoring of storage reservoirs are paramount to effective and safe storage of CO2. The risk of CO2 leakage can be reduced through real- time monitoring and early prediction of possible leaks. Early prediction of leaks involves identifying the onset of a leak and immediately visualizing the potential migration pathways. At present this goal is not achievable by available technologies like seismic inversion on time-lapse seismic data because the inversion process typically takes weeks to months, and the results are not reusable for newly acquired data. This limits its applicability to provide value for agile decision making that is critical for CO2 storage management. In addition, evaluating the CO2 concentration within the plume is desired to improve the storage efficiency and overall economy of carbon management operations. In this proposed research work, we plan to incorporate the large amounts of multiscale and multimodal data available in EDX for various formations to create machine learning-aided rock physics models. The aim is to use the data in the EDX platform to develop an end-to-end mapping between the rock and seismic properties. Based on the change in rock properties, the CO2 plume migration path can be readily visualized. Rock physics modeling (RPM) development is driven by data availability; therefore, this study is only possible with a wide variety of data available in EDX for various storage formations. One of the biggest challenges in the RPM is different uncertainties in the model. While rock physics modeling forms the skeleton of what connects rock and seismic properties, supervised machine learning models become an essential component to reduce the uncertainties between the properties’ mapping and improving the accuracy of the models. Thus, we advocate a physics-based machine learning approach to RPM. The proposed CarbonWatch technology has huge economic and technological benefits, while minimizing adverse environmental impacts. In upcoming Phase II of the project, we will first improve the CO2 saturation estimation from seismic data by adding dispersion and attenuation into the rock physics modeling to increase the model sensitivity to fluid saturation. We will apply the workflow to real seismic data and develop its capability to deal with noises. We also plan to expand the concept to other data types like Distributed Acoustic Sensors (DAS) to further raised its applicability for real-time monitoring. Finally, we will develop the software package and commercialize the technology as a field-ready product.