±0.05 mm at 1,100°C: In-Line 3D Scanning Reaches the Hot Forging Line

A 3D scanner now measures hot die forgings at temperatures up to 1,100°C with ±0.05 mm accuracy, a 600 mm × 450 mm single-shot field, and a 0.40 s capture time — on the line, while the part is still hot. That specification, delivered by Wuhan VISION3D Technology Co., Ltd. (VISION3D), turns what was a post-process sampling chore into an in-process measurement, and it changes what forging engineers can legitimately claim about their process data.

Why cooled-part data doesn't answer the real question

Forgings change dimensions with temperature, so a part measured cold tells you about the part after cooling — not about the conditions the die actually produced. Only measurements under real hot-forging conditions feed die design and process optimization with representative numbers. The old method — forge with generous machining allowances, machine later — disrupts metal flow lines (the grain structure that gives forgings their load-bearing performance in aerospace and automotive parts) and wastes material and energy.

Hot measurement has four standing obstacles: production-line vibration corrupts data acquisition, complex geometries defeat naive path planning, no markers can be attached to a glowing surface, and the line's equipment resists coordination. All four, it turns out, yield.

What did the four-fringe algorithm change?

The physics rests on fringe projection profilometry (structured light, where projected fringe patterns and their phase shifts encode surface height). A 2022 study in the journal Forging & Stamping Technology proposed a fast phase retrieval algorithm needing only four fringe projections at different spatial frequencies, with no phase shifting at all. The projected image count dropped from the traditional 12 to 4. That cut calculation time and, importantly, made the measurement more robust to on-site vibration — less exposure time, less vibration smearing. The method was validated on actual high-temperature forgings.

From there, the vendor built a complete solution: 3D measurement head, automated path-planning software, automated data-processing software, and in-line equipment. Hence the numbers in the opening paragraph. Vendor figures, mind you; but they are anchored to a peer-reviewed algorithm and deployment at named component categories.

What did the deployment data show?

The shift is from sampling inspection to 100% dimensional inspection of hot forgings, feeding a closed loop of 'measurement–analysis–feedback–optimization.' Application data from a Chinese company reports:

• First- and last-part hot-state inspection time: from 1–2 hours down to 2 minutes, removing batch-scrap risk.
• Product pass rate: up 10%–15%, thanks to timely detection of line abnormalities.
• Measurement path planning efficiency: up 5×; measurement viewpoints down 30%.
• Stitching accuracy of the automated data-processing software: 0.05 mm + 0.02 mm/m.

Note what these are: application results, not datasheet bullets. The stitching spec is the one metrological hard number, and it's stated with conditions. Good.

How does this fit legacy lines?

Most forging lines worldwide still run the allowance-and-machine approach. The upgrade path here doesn't replace the line; it adds sensors and digital systems to existing equipment. Deployment requires coordinated planning with the line's machinery, dynamic compensation of robot end-effector poses in hot environments, and feedback control of process parameters from dimensional statistics — the interface work that reconciles measurement with takt time.

The equipment is already running at global manufacturers of steering knuckles, front axles, and crankshafts, plus makers of aero-engine blades, heat-conducting inner rings, and gears, supporting lightweight vehicles and next-generation aircraft programs.

The credential trail is unusual for a vendor piece. Professor M. Saka, President of The Japanese Society for Non-Destructive Inspection, commented on the fringe image research: it "established a more accurate gamma distortion model for fringe images and proposed a more accurate gamma correction method." An expert acceptance review by the Ministry of Education Innovation Team 'Advanced Forming Technology and Equipment' concluded the results "reach the international advanced level as a whole, with some reaching the international leading level." Gamma distortion correction, for the record, addresses the nonlinear projector-camera intensity response that otherwise warps the fringe phase — a classic structured-light error source.

What's the verdict?

There's a green ledger too. Out-of-tolerance parts get segregated before downstream processing burns energy on them, and optimized processes shrink machining allowances, improving raw material utilization and cutting emissions. Not trivia — forging is energy-intensive, and allowances are paid for twice, once in material and once in machining heat.

The deeper point survives the vendor framing. A forging's dimensions vary with temperature, so the value of the dataset depends on how faithfully it captures hot-state behavior. In-line measurement at temperature gives engineers that representative basis for the first time at production scale. The question buyers should now put to their metrology budgets: is a sampling plan on cold parts still defensible when a 0.40 s full-field scan at 1,100°C exists?