Dubai-based JOME Engineering reproduced an existing marine propeller from a digital scan using Caracol's Vipra AM robotic wire-arc additive platform. The case matters because the output was not a display geometry but a replacement intended to return to service: a 685 mm, 26 kg SS316L component carried from scanning through final balancing.

The physical part was the only drawing

No CAD model or tooling accompanied the request. JOME scanned the original propeller and created a digital master. This is not a file conversion exercise: blade profile, pitch, hub transitions and wear-related deviations all require engineering decisions. The model can support a future rebuild, but only if critical dimensions and intended function have been separately verified.

Twenty hours of deposition, one hundred of machining

The hub was deposited first, precision machined and subjected to dye-penetrant inspection. Three blades were then added. Deposition on the Vipra AM XP took 20 hours, while machining required roughly 100. JOME re-scanned the part at each stage to track the master and preserve machining allowance, then completed hand grinding, pitch verification and balancing.

That ratio is the practical lesson. Printing was the smaller share of roughly 120 direct production hours. Fixturing, locating and machining freeform blades remained the bottleneck. A WAAM cell alone does not create a marine-spares capability; scanning, CAM, machining, non-destructive inspection and balancing remain part of the system.

What the supplier case study claims

Caracol and JOME report about 60% shorter lead time and 40% lower cost versus a one-off casting route, attributing the gain to avoiding new tooling and a foundry queue. These are partner case-study figures without a public cost breakdown, competing quotation or independent service data. The description of the part as the UAE's first is also a company claim.

Practical conclusion

The route fits one-off or low-volume large metal spares where tooling is expensive and downtime matters. Procurement should price the complete chain rather than the arc-on time: data capture, model approval, deposition, heat treatment, machining, NDT, dimensional inspection, balancing and release to service. This project is a working reference, not a substitute for the vessel operator's qualification plan.

Detailed workflow update, 21 October 2026

Later trade reporting clarified the sequence. JOME scanned the existing propeller because no drawing, CAD file or tooling was available. The digital master had to capture not only the 685 mm envelope but blade twist, pitch and blade-to-hub transitions. These freeform surfaces control hydrodynamic behaviour; visual similarity is insufficient.

The final SS316L part weighs 26 kg. Production began with the central boss, which was machined and then checked by dye-penetrant inspection for surface-breaking cracks and porosity. Only after that gate were the three blades deposited using Caracol's Vipra AM XP robotic WAAM system.

Deposition was one sixth of total time

Metal deposition took 20 hours; machining and finishing took about 100. Additive build time was therefore only one sixth of the roughly 120-hour operation. This corrects any business case that compares printing time alone with casting.

The propeller was rescanned after each main stage, compared with the digital master and checked for sufficient machining stock. Repeated measurement helped control twisted freeform blades with few usable references. Hand grinding, pitch verification and balancing were process steps rather than cosmetic additions.

What the 60% and 40% claims establish

Caracol reports about 60% shorter lead time and 40% lower cost than a one-off casting route. Avoided tooling and foundry queues provide a credible mechanism, but the percentages remain partner case-study claims. The baseline quotation, energy, labour, inspection and downtime tables are not public.

The retained digital file may avoid rescanning a future replacement only if the master geometry, edits, material, WAAM path, heat input, machining programme and inspection results remain versioned. A lone STL or point cloud is not a manufacturing master record.

Updated editorial conclusion

The case is valuable not because WAAM deposited metal in twenty hours, but because scanning, intermediate verification, nondestructive testing, one hundred hours of machining, pitch checking and balancing formed one spare-part workflow. The next evidence level is independent documentation of properties, tolerances and service loading against a classification requirement.

Sources and transparency

Dimensions, weight, processing times and savings come from the joint Caracol–JOME case study. VoxelMatters and 3Druck independently reported the event and workflow. No independent durability or in-service marine performance dataset was published.