INDUSTRY FIELD GUIDE
Compressor manufacturing data for AI performance-point review
CAGI publishes performance-verification information and compressor data resources.
NAICS 333912 · Air and Gas Compressor Manufacturing
CAGI publishes performance-verification information and compressor data resources. A manufacturer can propose a task that keeps a reported observation attached to its exact model, condition and source scope. CAGI performance-verification context
Write fictional compressor cards, a test-point table and a customer request. Give two observations different explicitly stated conditions and power scopes. Ask the assistant which point answers the request. Include a nominal motor value beside a recorded package value and a report from another model with a similar family name. The key should identify the scope and model distinctions.
Annotate model, test point, property, unit, condition and measurement boundary. Hold out a new table arrangement for evaluation. Score wrong-condition comparisons, nominal component values used as package observations and wrong-model joins independently. Supply any intended calculation formula so the review can reproduce the result from cited inputs.
Use self-authored documents and invented values. Actual performance histories, licensed test methods, certification material and customer requirements need permission and qualified review. A licensing proposal should identify the point-selection task and checked examples. Confirm buyer scope before preparing commercial reports or commissioning further capture.
Discuss this original manufacturing sample with Fiund
Original Fiund exercise
Two reports do not bracket the same test configuration
Fictional compressor cards report an arbitrary index. Test-point identifiers are names only.
| Record | Source | What it records |
|---|---|---|
| 1 | Request | Find value at point P2 under setup C for unit U5. |
| 2 | Table A | U5 P2, setup D, index 21. |
| 3 | Table B | U5 P1, setup C, index 18. |
Question for the model
Can a reviewer infer a P2/C value from the tables?
Show the reference answer and checks
No. A in record 2 has the correct point but a different setup, while B in record 3 has the correct setup but a different point. Record 1 requires both together. There is no compatible pair or documented interpolation method for estimating the missing value.
- Interpolating across setup changes
- Choosing the nearest point without preserving condition
Fields to preserve
- test_point_id
- Identifies the requested observation
- setup_key
- Defines comparability
- unit_id
- Ensures evidence concerns the same specimen
Test another case
Provide three points under C with P2 explicitly recorded; select it instead of estimating from adjacent points.
No actual compressor settings, performance curves, pressure limits or operating advice.
Discuss a contribution
Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.