AI Reservoir Simulation
What AI reservoir simulation should actually do.
Learn how AI reservoir simulation combines natural-language engineering intent, simulation deck generation, deterministic validation, OPM Flow execution and reservoir-engineering judgment.
AI reservoir simulation applies language models and agentic workflows to the work around numerical reservoir simulators: interpreting an engineering request, constructing or editing model inputs, checking the model, running the simulator and analyzing outputs.
Where AI helps
- Translate engineering intent into structured simulator inputs.
- Repair syntax and task-level omissions.
- Automate repetitive model edits and sensitivity cases.
- Run simulators and inspect diagnostics.
- Return results with assumptions and provenance visible.
Why the simulator stays in the loop
A language model can produce a deck that looks correct while still being wrong. Simulator parsing, initialization behavior, run completion and physical-consistency checks provide stronger evidence than fluency alone.
CLARISSA's approach
CLARISSA combines AI model authoring with deterministic gates, OPM Flow execution, ECLIPSE-compatible deck workflows and the RIGOR benchmark.