First Break - Volume 42, Issue 2, 2024

Protecting natural and near-natural ecosystems is becoming increasingly important in an ever more densely populated and intensively used earth surface. Climate change-induced extreme weather events have accelerated the environmental degradation that has been taking place for centuries. Comprehensive and precise environmental monitoring is therefore essential, especially around mining, post-mining and industrial sites. Traditional in-situ measurements are inadequate for wide areas, necessitating the integration of satellite and unmanned aerial vehicle (UAV) remote sensing data for a more comprehensive monitoring.

This multi-level approach utilises satellites for large-scale and high-temporal remote sensing and UAV data for medium-area and high-precision monitoring, while in-situ measurements serve as validation for both data sources. Different case studies at the Research Center of Post-Mining demonstrate the approach’s effectiveness in geomonitoring post-mining processes and risk management in the oil and gas industry.

Integrating diverse data sources enables comprehensive monitoring and analysis, enabling the creation of user-friendly web applications to facilitate efficient risk management decisions. This multi-level monitoring concept offers an efficient approach to understanding and addressing environmental changes and risks in various industries and conservation projects.

Ensemble history matching adjusts multiple geomodels used for reservoir simulation, conditioning them to historical data. It reduces and quantifies the uncertainty in the unknown model parameters to increase the models’ reliability for decision support. In this study, we adapt the latest generation of generative artificial intelligence algorithms, SPADE-GANs. In geosciences, Generative Adversarial Networks (GANs) learn to simulate complex geological patterns. SPADE (SPatially Adaptive DEnor-malisation) layers in the GAN generator learn conditioning to the coarse geological structure provided as coarse-scale maps, enabling explainable output, stable training, and higher variability of resulting outputs. Our statistical method, an iterative ensemble smoother, assimilates data into an ensemble of these maps, interpreted as the channel proportions. This Bayesian data assimilation conditions the ensemble of GAN-geomodels to a combination of well data and flow data, thus extending the usability of pretrained SPADE-GANs in subsurface applications. Our numerical experiments convincingly demonstrate the method’s capacity to replicate previously unseen geological configurations beyond GAN’s training data. This proficiency is particularly valuable in data-scarce scenarios typical for renewable geo-energy, where the GAN captures realistic geology, but its output geomodels must be adjusted to match observed data. Furthermore, our fully open-source developments lay the foundation for future multi-scale enhancements of history matching workflows.

The extended abstract for this article is published in the proceedings of the Fifth EAGE Conference on Petroleum Geostatistics (November 27–30, 2023; Porto, Portugal).

This study illustrates a field application of a robust and reliable approach for assessing geomechanical parameters during drilling operation or as a post-mortem analysis. The methodology leverages surface logging drilling data (such as Rate of Penetration, Rotation Per Minute, Weight on Bit, Torque, Standpipe Pressure, and Flow rates) along with well log data (including Sonic log, Bulk Density log, and Gamma Ray log) as input of the methodology. This model is grounded on the integration of various data processing techniques and machine learning algorithms (encompassing Multiple Linear Regression, Support Vector Regression, Random Forest, Artificial Neural Network, and XGBoost), ensuring a comprehensive and accurate evaluation of geomechanical parameters in the area under evaluation.

The methodology is applied to a dataset of six wells drilled in the same geological units of an area located towards the eastern limit of the Neuquén Basin, north of the ‘Dorsal de Huincul’ (Gonzalez et al. 2016). The results associated with Young’s Modulus, Density, and UCS, here presented, provide evidence of its successful use in a challenging geological context such as the unconventional plays in Argentina.

We have developed a data structure called GEOH5 with the objective of integration and storage of geological models, data, and metadata where dissemination, ease of access, and persistence are required without commercial encumbrance. Our emphasis is on the needs of the mineral industry which, unlike the upstream oil and gas industry, otherwise lacks common data exchange formats with a scope encompassing most exploration and production data types. Although only a few years old, the data structure is already in use by many thousands of users with increasing acceptance across mineral geoscience and engineering. This includes industry, academia, and geological surveys that use GEOH5 as a documented, public, easy-to-use, vendor-neutral, and permanently accessible means of storage and communication. GEOH5 is open source and free to use. It is based on open-source HDF5 technology because of its many advantages: wide acceptance across numerous data-intensive industries, self-describing behaviour through integration of data and metadata, fast I/O, excellent compression, file merging, cross-platform capability, unlimited data size, and access to libraries in a variety of programming languages. It provides professionals, researchers, and the public at large with a robust means of managing, exchanging, and visualising large quantities of diverse mineral geoscience and engineering data.