Together with Füsun Servin Tut Haklıdır, I have worked on data-driven methods for geothermal energy systems: fuzzy control of calcium carbonate and silica scaling (Geothermics, 2017), machine-learning prediction of reservoir temperatures from hydrogeochemical data (Natural Resources Research, 2020), deep-learning prediction of geothermal boron contamination (Environmental Earth Sciences, 2020), synthetic hydrogeochemical data generation with generative adversarial networks (Stanford Geothermal Workshop, 2025) and rank–Gaussian copulas (Geothermics, 2026), prediction of dissolved lithium in geothermal fluids with calibrated uncertainty quantification (Earth Science Informatics, 2026), agent-based reinforcement learning for geochemical control (2026) and agentic AI for dynamic reservoir management (World Geothermal Congress 2026).
Earlier papers explored hydrogen production from geothermal sources and from H₂S in the Black Sea.