Case versioning lets you snapshot, compare, and track the evolution of your simulation setups. Combined with golden baselines and trust scoring, it provides a complete audit trail of how a simulation changed and whether those changes are acceptable.
Case versioning
Every simulation case can be versioned. A revision captures the complete state of the case at a point in time: mesh settings, boundary conditions, solver parameters, numerical schemes, and convergence criteria.
Create a revision
Snapshot the current case state. The revision includes full reproducibility metadata -- everything needed to recreate the exact same simulation.
Compare revisions
Diff any two revisions to see exactly what changed. The comparison highlights differences in mesh configuration, boundary conditions, numerical settings, and solver parameters.
Branch for experiments
Create a branch from any revision to try experimental modifications without affecting the mainline case. Merge back if the experiment succeeds.
Golden baselines
A golden baseline is a revision that has been validated and promoted as the reference standard for a simulation case. It represents a known-good configuration that future changes are measured against.
To promote a revision to a golden baseline:
- Run the simulation and verify results meet acceptance criteria
- Complete any required validation checks (mesh convergence, experimental comparison)
- Promote the revision -- it becomes the reference point for all future comparisons
Once a revision is promoted to a golden baseline, it cannot be modified. This ensures the reference standard remains stable and auditable.
Trust scores
When you compare a new revision against a golden baseline, SimPilot computes a trust score that quantifies how closely the new revision matches the validated reference.
| Score range | Rating | Meaning |
|---|
| 0.80 -- 1.00 | Green | Changes are minor and within validated tolerances |
| 0.50 -- 0.79 | Yellow | Significant changes detected -- review recommended |
| Below 0.50 | Red | Major deviations from the baseline -- validation may be invalidated |
The trust score accounts for the type and magnitude of each change. A small mesh refinement has less impact than changing the turbulence model or boundary condition type.
Baseline compliance
Check whether a new revision complies with a golden baseline:
- Trust score: Overall numerical score quantifying deviation
- Deviation list: Itemized list of every difference with its impact assessment
- Compliance verdict: Pass, conditional pass (with noted deviations), or fail
This is particularly useful for regulated industries where changes to validated simulations require documented justification.
Branching
Create branches from any revision to explore modifications without risk:
- Experimental branches: Try a different turbulence model or mesh strategy
- What-if analysis: Test the effect of changing boundary conditions
- Parallel development: Multiple engineers can work on different branches of the same case
Each branch maintains its own revision history and can be compared back to the parent revision or the golden baseline at any time.