This question aims to build a decision framework that identifies relevant gaps in petrophysics and evaluates—using technical criteria and clear metrics—when and why AI is appropriate to address them, including considerations of data, physics, uncertainty, interpretability, risks, and expected outcomes.
Some open problems in petrophysics where AI could help include: Better reservoir property prediction – using limited well logs or seismic data to more accurately estimate porosity, permeability, and fluid saturation. Automated core and thin-section image analysis – speeding up and improving grain size, texture, and mineral identification. Data integration – combining well logs, seismic, production, and core data for more consistent interpretations. Uncertainty quantification – providing probabilistic forecasts instead of single-value estimates. Real-time monitoring – detecting changes in reservoir performance from streaming sensor data.