The future of the upstream industry lies in the adoption of innovative geoscience digitalization solutions to optimize operations. Machine learning applications for geoscience data have been in use for more than 25 years but have recently become critical due to massive growth in the amount of petrotechnical data being acquired. As machine learning evolves, it will play an increasingly visible role in analyzing surface and subsurface data.

The upstream oil and gas exploration and production industry is looking for ways to reduce costs while minimizing emissions, water use and other environmental impacts. Near-field exploration and development enables producers to leverage already depreciated costs in operating infrastructure and extend the life of a declining field by accessing new or previously bypassed reservoirs.

Global energy companies trust the AspenTech® Subsurface Science & Engineering portfolio to solve their most complex exploration and production challenges while reducing geological risks and minimizing impact on the environment. Get a quick look at our products, with QR codes linking to more details. Download now.

Leverage machine learning to perform quick and accurate amplitude inversions and rock property estimations when short project timelines exist. Aspen SeisEarth’s Neural Network Inversion feature provides a step-by-step workflow available for interpreters and non-specialists.

Full azimuth angle domain imaging provides an alternate way to map events in structurally complex areas. Information about continuous surfaces can be derived from specular gathers, while diffraction gathers are used to derive information about discontinuity i.e., faults and small-scale fractures.

This paper presents a method for maximizing fault information from depth migrated narrow-azimuth as well as full-azimuth seismic data. The study demonstrates that depth domain diffraction imaging can be used to generate higher resolution fault definition than conventional reflectivity volumes, or their derivative post-stack attributes.

Upstream companies today must achieve operational excellence by reducing emissions and utility demands, improving production at existing assets and replacing and expanding reserves while exhibiting capital discipline.

Discover how seismic volume-based unsupervised facies classification associated with advanced visualization and detection helps delineate the prospect’s potential, increase drilling success and reduce cost and risk.

This article shows how a machine-learning workflow based on a Self-Growing Neural Network (SGNN) was used by Aspen SeisEarth™ as an efficient and unbiased scanning tool for carbon capture and storage (CCS) monitoring, enabling faster identification of the confinement system.

Seismic data can provide useful information for prospect identification and reservoir characterization. Combining seismic attributes helps identify different patterns, thus improving geological characterization. Machine learning applied to seismic interpretation is very useful in assisting with data classification limitations.