Geophysics and Applied Mathematics for Exploration and Safe Production

The GAMES project focuses on innovative methods for exploration and safe production of hydrocarbon resources offshore Norway. We develop geophysical and statistical methodologies that are also applicable to petroleum resources elsewhere, as well as in other types of applications relying on spatio-temporal prediction and monitoring. When sedimentary rocks are uplifted or buried the seismic velocities will change. To know and quantify such changes are important to evaluate reservoir properties prior to drilling. In a recent study we used laboratory measurements of a synthetic sandstone formed under stress to study how velocities vary when the sample is loaded and unloaded. We find that the P-wave velocity is more sensitive to stress changes during uplift. Furthermore, the P-wave anisotropy transforms from negative to positive. In another study we performed a DAS (Distributed Acoustic Sensing) field test in the Trondheimsfjord, where we compared acoustic data recorded on a fibre optic cable crossing the fjord to conventional hydrophone data. The comparison shows that the shallow geology is imaged on both data sets, and that the quality of the DAS-data is slightly less than the corresponding hydrophone data. I We have studeid variaous methods for Bayesian inversion of seismic data and well information at the Alvheim field in the North Sea. Here, the goal is reliable predictions of oil and gas in the reservoir unit. By coupling geological prior information with a geophysical model for the data, we discover potential opportunities for increased oil and gas production in the vicinity of existing wells. At the same time, the clay content appears to be rather high in some of these zones (dirty sands), and this could be a challenge for production. Our predictions and associated uncertainty quantification of these reservoir variables support future decision making at this field. An ultrasonic test tank (2 by 4 m) has been developed as a part of an Eccsel-project (supported by EU). In this tank we can simulate a scaled (2000 times smaller than reality) version of CO2-injection into a subsurface storage volume. We use a grid of ultrasonic transducers for this purpose. A Games post doc is currently working on practical methods to move the ultrasonic devices in a continuous manner to achieve an acceptable imaging of the CO2-plume. For this purpose we use a 3-dimensional plastic model corresponding to the overburden at the Sleipner CO2 storage site. This work is relevant in order to obtain improved images of the CO2-injection process and also for early leakage detection from a CO2 storage volume.

The project is focused within exploration and production of hydrocarbons, and by combining geophysical and mathematical methods we will develop new methods and insight. We aim to get added value by i) using statistical machine learning techniques to improve the way geophysical data are integrated in subsurface models, ii) realistically constraining spatio-temporal statistical models by rock physics relations. This project proposal focuses on innovative methods for exploration and safe production of hydrocarbon resources offshore Norway, developing geophysical and statistical methodologies that are also applicable to petroleum resources elsewhere, as well as other applications relying on spatio-temporal prediction and monitoring. PhD candidates from the current project will use basic disciplines (mathematical sciences and geophysics) to develop insight that is directly relevant to petroleum-related industries. The generic form of knowledge is also applicable to other domains of earth sciences (mining, oceanography, meteorology, etc.), as well as other industries related to energy or new digital-type companies. Key areas that will be investigated are: - geophysical analysis of uplift - new geophysical monitoring methods - elastic seismic migration and inversion Six PhDs and one postdoc will be employed by the project. Seven companies have the intent to support the project both financially and by sharing data and hosting students. Four international highly reputed academic institutions have stated their willingness and interest to participate in the project.