Studies are structured multi-experiment campaigns managed by the AI. You describe what you want to investigate, and SimPilot designs the experiments, batches and runs them, evaluates results against your criteria, and decides what to run next -- all automatically.
Study types
SimPilot supports five categories of engineering studies:
How studies work
Describe the study
Tell the AI what you want to investigate. For example: "Run a parametric sweep of inlet velocity from 5 to 25 m/s in steps of 5" or "Optimize the fin height and spacing to maximize heat transfer while minimizing pressure drop."
Experiment design
The AI designs the experiment matrix based on your description -- selecting parameter ranges, sampling strategies, and success criteria.
Batch execution
Experiments are batched and run with up to 3 concurrent simulations. Each simulation follows the full engineering pipeline: case setup, mesh generation, solver execution, and error recovery.
Evaluation
As results come in, the AI evaluates them against your criteria. For optimization studies, it decides which experiments to run next based on what it has learned so far.
Dashboard generation
After completion, an interactive dashboard is automatically generated with chart types appropriate to the study: parallel coordinates for sweeps, Pareto fronts for optimization, tornado plots for sensitivity analysis.
Concurrency
Studies run up to 3 simulations concurrently. For studies with more experiments than the concurrency limit, jobs are queued and launched as earlier ones complete. You can monitor progress in real time from the chat interface.
After completion
When a study finishes, you get:
- Summary report with key findings and recommendations
- Interactive dashboard with appropriate visualizations (see Interactive Dashboards)
- Comparison tables ranking all experiments by your target metrics
- Raw data available for export and further analysis
Studies can be chained. Run a sensitivity analysis first to identify the most influential parameters, then follow up with a focused optimization study on just those parameters.