Michael R. Gross

TL;DR: Researchers develop a machine learning method to detect and characterize hydrogen leakage from underground storage using sparse seismic data, achieving high accuracy and potentially serving as a cost-effective geophysical tool.

TL;DR: This laboratory study investigates cyclic underground hydrogen storage in porous media, revealing high withdrawal efficiency (up to 95%) and evaporation-induced salt precipitation, with potential for monitoring H2 plumes using 4D seismic surveys.

TL;DR: Monitoring subsurface fracture flow using unsupervised deep learning of borehole microseismic waveform data classifies different types of microseismic events and finds that low-frequency LD events occur only during the proppant injection period.

TL;DR: The preliminary results show that the ML-models developed based on PIML workflow have good quantitative predictions and are significantly faster yet provide accurate predictions for real-time history matching and forecasting at shale-gas sites.