Skip to content

AURA Validation Chain Registry

The open evidence-gated validation registry for aerostatic gas-bearing rotor-dynamic engineering software.

Maintained by Breshev Engineering. Schema and documentation under CC-BY-4.0. Validation infrastructure code under MIT license. Repository: github.com/lfnavigator/aura-registry.


What this is

A schema-enforced public record of validation evidence for rotor-dynamic engineering software, with a focus on aerostatic gas-bearing systems (precision spindles, microturbomachinery, cryogenic engineering, and high-reliability rotating machinery).

The registry implements the Breshev Validation Chain (BVC) methodology:

  • A two-layer registry separating solver-sanity evidence (analytical anchors at machine precision) from engineering-evidence chains (multi-step evidence terminating in measured or experimentally grounded physical data).
  • A four-level acceptance policy (analytical sanity, public external benchmark, external diagnostic, forensic/blocked) that prevents evidence from being promoted beyond what its provenance supports.
  • A two-gate model combining structural completeness (schema gate) and content completeness (runtime quality gate).
  • Deterministic SHA-256 hashes over canonical serialization for promoted registry entries, enabling third-party re-validation without access to private systems.

The methodology paper is in preparation for external review and submission. The registry is active and already contains its first public scoped Breshev gas-bearing method-chain anchor.


Current state — May 2026

Registry status snapshot

The registry now contains a first public scoped engineering-evidence anchor for the Breshev conical aerostatic bearing method chain: VCR_BRESHEV_CONICAL_P05_METHOD_CHAIN_001.

Layer Current state Public status
Solver-sanity 10 analytical rotor-dynamic anchors in the active regression track Level A foundation
Engineering-evidence 1 public scoped Breshev gas-bearing method-chain anchor Public scoped anchor
Diagnostic (Level C) OS-V:1010 external diagnostic comparison Diagnostic only, not public benchmark
Forensic (Level D) Nelson-McVaugh reconstruction record Forensic/internal, not active evidence

What this means in practice:

  • The solver-sanity layer exercises specific solver capabilities against closed-form mathematical references. These anchors form the mathematical verification foundation under AURA's dynamic workflow.
  • The engineering-evidence layer now includes a first public scoped gas-bearing anchor: a Breshev conical aerostatic bearing method-chain case linking industrial spindle evidence, experimentally validated FEM results, and the Breshev perturbation method used in AURA.
  • Diagnostic and forensic records are deliberately preserved when they are useful for methodology development, but they are not promoted into engineering-grade evidence unless their provenance supports that promotion.

For the full registry status, see Registry overview.


Getting started

If you are... Start here
New to BVC and want the conceptual frame Methodology overview
An engineer evaluating AURA for pilot use Engineering-evidence layer and Lifecycle states
A contributor wanting to add an anchor Anchor submission guide
Implementing the schema in your own tooling Schema v1.1 reference
Citing AURA in academic work Citation guide

Key design principles

Evidence over assertion. Every claim in the registry must be backed by a declared source, a comparison target, residuals, and an explicit lifecycle state.

No silent promotion. Diagnostic comparisons and forensic records are valuable, but they do not become public benchmark anchors unless the acceptance policy requirements are met.

Useful limitations. A limitation is not a defect if it is explicit. The registry records what each anchor proves, what it does not prove, and where future work is required.

Engineering decisions, not just solver outputs. The long-term purpose of the registry is to support trust verdicts that are appropriate to the consequences of engineering decisions.