Search

Accurate estimation of mineral composition is essential for refined reservoir characterization, thermal conductivity and mechanical determinations of sedimentary rocks, but is extremely challenging in shale units due to the mineralogical complexity, low porosity and ultra‐low permeability. Direct mineral measurements derived from laboratory X‐ray diffraction (XRD) analysis on core samples and borehole geochemical logging tool (GLT), and conventional geophysical logs from vertical wells penetrating sediments are widely available in some basins, which enables detailed mineralogical characterization of a well. A hybrid machine learning (ML) architecture that improves model training and validation by combining convolutional neural network (CNN) with XGBoost allows accurate description of the mineralogical compositions across a basin. We applied this ML approach to predict the mineral compositions using conventional well logs from the Horn River Basin, northeast British Columbia, Canada, where extensive drilling for shale‐gas and conventional hydrocarbon resources, complemented by high temperature geothermal energy potential is ideal for case testing. The predicted mineral compositions from the ML approach are consistent with the mineralogical readings from the GLT and are confirmed by the XRD mineral measurements. This allows basin‐wide mineral compositions mapping that reveals spatial trends of major mineral compositions and their relationship with the previously recognized geomechanical and geological features, which have important implications for thermal conductivity modeling, reservoir evaluation and extensive geological studies. Key Points: A hybrid machine learning approach is applied to develop log‐mineral models for mineralogy characterization using geophysical well logsThe model‐predicted major mineral compositions are consistent with the mineral data from geochemical logging tool and XRD analysisThe Horn River Group shale is dominated by quartz, clay, and carbonate and shows great variability among the three formations in the basin [ABSTRACT FROM AUTHOR]

Inversion is a fundamental step in magnetotelluric (MT) data routine analysis to retrieve a subsurface geoelectrical model that can be used to inform geological interpretations. To reduce the effect of non-uniqueness and local minimum trapping problems and improve calculation speeds, a data-driven mathematical method with a deep neural network was developed to estimate the subsurface resistivity. In this study, a deep learning (DL) inversion technique using a revised multi-head convolutional neural network (CNN) architecture was investigated for MT data analysis. We created synthetic datasets consisting of 100,000 random samples of resistivity layers to train the network's parameters. The trained model was validated with independent noised datasets, and the predicted results displayed reasonable accuracy and reliability, which demonstrates the potential application of DL inversion for real-world MT data. The trained model was used to analyze MT data collected in the southwestern Athabasca Basin, Canada. The calculated results from the DL method displayed a detailed subsurface resistivity distribution compared to traditional iterative inversion. Since this approach can predict a resistivity model without multiple forward modeling operations after the CNN model is created, this framework is suitable to speed up the computation of multidimensional MT inversion for subsurface resistivity. [ABSTRACT FROM AUTHOR]

Activation energy distribution from kerogen kinetics study contains some integral characteristics of the host source rock thermal history over geological time scales. Extracting relevant information from pyrogram is critical for reconstruction of source rock maturation and hydrocarbon generation histories. In this study, in order to improve understanding of the Bakken shale source rock, we present a new quantitative kinetic model by inversion of Rock-Eval pyrograms of the Bakken Formation in Williston Basin. A group of core samples from both Upper and Lower Members of the Bakken Formation were collected for estimating transformation ratio (TR) and pyrolysis analysis, and an inversion method with different kinetic statistic parameters was designed and conducted using the source rock T max -Hydrogen Index covariation combined with different burial histories. The analysis of Rock-Eval results on Williston Basin samples from both Canada and US sides demonstrate basin-wide organic matter characteristics of Upper and Lower Bakken and identify variation of hydrocarbon generation histories. The Upper and Lower Bakken shales have different kerogen type characteristics based on activation energy distributions derived from T max -HI pair relationship, which suggest variance of the preserved algal materials for Upper and Lower Bakken. The modeling results by the thermal history analysis with the Rock-Eval dataset indicate a higher remanent generable kerogen in Lower Bakken shale, and the thermal maturity results will be valuable for integrated assessment of the hydrocarbon systems of Williston Basin accurately. • Kinetics parameters are calculated by pyrogram inversion for Upper and Lower Bakken Members • Improved pyrolysis reaction model is created by combining Kissinger equation and T max -HI modeling • The characteristics of Upper and Lower Bakken suggest different algal input and OM preservation history [ABSTRACT FROM AUTHOR]

• Ground surface temperature (GST) monitoring network tested as a geothermal resource exploration tool. • Daily GST variation in the snow free season shows clear environmental and topographic footprints, and a snow curtain period of near zero GST in winter with thick snow. • Two GST anomalies associated likely with subsurface heat flow identified during snow curtain period when solar irradiation and surface weather effects were greatly suppressed. • Non-supervised machine learning partitions observation stations into mutually exclusive clusters statistically, useful in data analysis and interpretation. The southern Mount Meager area has been studied for its geothermal resource potential for a half century. Although attempts to commercialize the resource were unsuccessful in the past, a large volume of surface and subsurface geoscience data have been collected and made publicly available, making the Mount Meager area a unique natural laboratory for testing new ideas or tools for geothermal exploration. This study investigates the potential of using a ground surface temperature (GST) monitoring network as a tool in geothermal resource exploration. Twenty-two temperature data loggers were deployed in, and surrounding, the core area of the geothermal prospect for nearly one year. The collected temperature time series were analyzed statistically, and machine learning methods such as hierarchical clustering and k-means classification, were applied to data interpretation. The temporal variation of measured temperature can be divided into four segments: 1) snow free, 2) snow affected, 3) snow curtain, and 4) snow melt seasons. The intensity of daily ground temperature variation in the snow free season shows clear environmental and topographic footprints and is complicated by soil properties. The sites with a short period of near zero GST during the snow curtain period is spatially coincident with high heat flow areas. The measured mean GST shows a close affinity with subsurface heat flow and shallow ground water circulation, suggesting that GST time series contain information indicative of subsurface geothermal anomalies and that a GST monitoring network could be an effective tool for geothermal resource exploration in high latitude and/or elevation areas with seasonal snow cover. [ABSTRACT FROM AUTHOR]