Reverse Engineering Marine Propellers

Marine propellers sit at the intersection of precision hydrodynamics and brutal real-world service conditions. When one comes out of the water worn, damaged, or simply undocumented, you need geometry you can trust - not a rough digitization that a foundry will reject. Our workflow covers the full range from 30 cm recreational two-bladers to 4-meter CPP hubs in dry-docks, and addresses every failure point that generic scan shops miss. Here’s exactly how we do it, what it costs, and what you need to send us.

---

Why Marine Propellers Are a Reverse-Engineering Challenge Unlike Any Other

Most machined parts have flat faces, symmetric features, and known tolerances. A marine propeller has none of those luxuries.

Every blade is a compound curve where rake, skew, pitch, chord, and camber all change at every radius station from root to tip. Cast geometry is never truly symmetric blade-to-blade - bronze casting tolerances mean you’re always working with the actual artifact, not a theoretical ideal. After years of service, cavitation erosion, barnacle pitting, and tip strikes layer additional asymmetry on top of what was already a complex surface.

Then there’s the material. Ni-Al-Bronze alloys are highly reflective. A standard structured-light scanner aimed at a polished bronze blade face will blow out highlights and return garbage data in those zones. The fix is Aesub Blue matte spray - it applies in seconds and the active coating sublimes over several hours with no residue left on the metal. Skip this step or use the wrong coating and you’ll be re-scanning.

The critical distinction our engineers make on every propeller job is design-intent reconstruction vs. copy-exact digitization. A worn propeller tip missing 8 mm of material from a strike is not the geometry the foundry should cast. Copying the damage exactly into a STEP file produces a defective replacement. We document what’s worn, reconstruct the blade section series to design intent for the damaged zones, and deliver a deviation map that shows both - the as-found artifact and the idealized model the foundry receives.

Size range matters for gear selection. A recreational three-blade at 30 cm requires bench-top structured-light scanning. A 4-meter fixed-pitch commercial prop requires a completely different kit and a field team in the dry-dock. And across that full range, there’s regulatory weight: SOLAS, ABS, and DNV classification rules require documented propeller geometry for replacement qualification. A CAD file with ISO 484-1 section plots satisfies that paper trail; a scan-shop STL mesh does not.

---

Scanning Equipment Selection for Marine Hardware

Gear selection is driven by prop diameter, surface condition, and whether we need hull context geometry in the same session.

The Creaform MetraSCAN accuracy figure reflects the current BLACK+ (0.035 mm) and BLACK+ Elite (0.025 mm) models per the Creaform MetraSCAN 3D Technical Specifications. The Leica RTC360 3D point accuracy is per the Leica RTC360 datasheet. The Trimble X7 3D point accuracy is per the Trimble X7 datasheet.

For a typical 4-blade commercial fixed-pitch prop in the 0.6-1.5 m range, we run our Creaform MetraSCAN for the blade surfaces and deploy our Trimble X7 to capture the surrounding hull - A-bracket position, strut geometry, shaft axis - in the same session. The TLS data feeds the clearance verification model; the handheld scanner data feeds the CAD reconstruction.

Reference targets are non-negotiable. We press magnetic nests onto the boss hub to hold a consistent frame of reference across all blade setups. Blade tip and trailing-edge targets prevent the “sail effect” - the stitching error that occurs when you register two scans taken at opposite ends of a long, thin surface using only ambient geometry. Without tip targets on a narrow-chord blade, registration errors of 0.5-1.0 mm are common and go undetected until you’re in CAD wondering why the leading-edge NURBS won’t blend cleanly.

Shop conditions are the other variable most people underestimate. Vibration from adjacent machinery and direct sunlight on reflective bronze are the two most reliable data quality killers. Our pre-scan site brief covers both: we time scans around adjacent equipment cycles and use matte spray plus careful scan angle management to kill specular reflection.

---

The 7-Step Scan-to-CAD Workflow for a Marine Propeller

This is our field-to-deliverable process, not a marketing diagram.

Step 1 - Condition survey and photography. Before any cleaning or prep, we photograph and document every tip erosion zone, blade crack, pitch marking, and hub stamping. These photographs become the damage baseline for the engineer and the client - critical if there’s ever a question later about what was worn vs. what was reconstructed.

Step 2 - Surface prep. Light degreasing to remove marine growth residue, followed by Aesub Blue on highly polished blade faces. We apply coded photogrammetric targets to the hub boss and one blade leading edge for global reference.

Step 3 - Data capture. Minimum 8 scanner setups per blade: suction face, pressure face, leading edge, trailing edge, root fillet, tip, plus two intermediate angles to capture the chord camber transition zones. The hub is scanned separately with reference overlap to the blade dataset. For a 4-blade prop this means 32+ individual scan positions before we pack up.

Step 4 - Point cloud registration and QC. We align all scans in Creaform VXelements and run a full deviation map of blade-to-blade registration. Any blade showing a stitch error above 0.3 mm gets flagged and re-scanned before we leave site. We do not attempt to fix registration problems in post.

Step 5 - Mesh processing in Geomagic Wrap. Pitting voids from cavitation erosion are filled conservatively - we preserve the erosion record in the mesh rather than over-smoothing it into oblivion. The goal is a watertight mesh that accurately represents the artifact, not a cosmetically cleaned surface that obscures damage extent.

Step 6 - Parametric CAD reconstruction. In Geomagic DesignX (or Siemens NX reverse-engineering module for clients with existing NX workflows), we extract blade cross-sections at 10-15 radius stations, typically r/R = 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 0.95, and 1.0. Each section is fit as a NURBS curve. We then loft between sections to produce Class A NURBS blade surfaces and merge them with the hub solid. For damaged tip zones (typically r/R > 0.85 on worn props), the section geometry is extrapolated using the blade section series - Wageningen B-series coefficients where the original OEM spec isn’t available.

Step 7 - Deliverables package. Every propeller project ships with: STEP (AP214) + IGES neutral files, native Parasolid .x_t or CATIA V5 if specified, a 2D inspection drawing with ABS/DNV blade section pitch plots, a scan deviation color-map PDF showing as-found vs. design-intent, and the raw .e57 point cloud archive. The client owns all of it.

---

Accuracy Benchmarks and ISO/Class Society Tolerances

ISO 484-1 defines manufacturing tolerance classes for ship propellers. Here’s where scan-derived CAD sits against those benchmarks:

Note on ISO 484-1 Class S and Class II tolerance figures: the specific numerical values for pitch tolerance and Blade Area Ratio tolerance for Class S and Class II could not be independently verified from a primary source with a publicly accessible URL. Consult the ISO 484-1:2015 standard directly for the full tolerance table. Class I pitch tolerance of ±1.0% per station is confirmed by multiple industry references.

Our metrology-grade handheld scanner achieves ±0.025-0.035 mm on blade faces - well inside Class S pitch station tolerances. The critical measurement for a replacement propeller is helical pitch extracted at r/R stations 0.7 and 0.9, which are the primary hydrodynamic load stations. We compare suction-face pitch against pressure-face pitch at both stations. If the asymmetry exceeds 0.3%, we flag the blade for design-intent reconstruction rather than copy-exact digitization.

Blade Area Ratio (BAR) and Expanded Area Ratio (EAR) are computed directly from the finished NURBS model and cross-checked against hub stampings or the original specification sheet when available. This cross-check has caught discrepancies on separate jobs where previous owners had already swapped a non-original prop - the stamped hub data didn’t match the blade geometry we measured.

For hull context, our Trimble X7 TLS captures shaft axis, A-bracket, and strut positions to within its specified range accuracy of 2 mm. That is more than sufficient to verify replacement propeller tip clearance under full rudder deflection - a calculation the naval architect needs before the casting order is placed.

---

Scan-to-Mold: Producing a Replacement Casting Pattern from Scan Data

The CAD model leaving our system is not ready for the foundry yet. Bronze propellers experience shrinkage during casting, and the exact shrink allowance depends on the foundry’s specific process and alloy heat. Our CAD work builds the foundry-specified shrink factor directly into the STEP file before release - a step that a raw STL mesh cannot accommodate without re-meshing, which is why we tell every client that an STL is not a foundry deliverable.

Two pattern routes are available depending on timeline and budget:

Route A - CNC-machined foam or epoxy tooling pattern from the STEP file. Fastest option, ±0.5 mm pattern accuracy, 2-4 days from STEP approval to pattern in the foundry’s hands. Suitable for standard commercial props where casting variability is acceptable within ISO 484-1 Class I.

Route B - Direct 5-axis CNC of Ni-Al-Bronze blank from the STEP file. Eliminates casting variability entirely. Preferred for high-value props, safety-critical CPP blades, or applications where Class S tolerances must be met on the final machined component.

Draft angle analysis is a prerequisite for either route. We analyze the blade geometry in Geomagic DesignX to identify the natural parting planes and flag any negative-draft zones - typically root fillets and the hub face - before the pattern is cut. A foundry that discovers a negative-draft zone after the pattern is machined loses 2-4 days and several thousand dollars. This analysis takes us roughly two hours in DesignX and costs the client nothing on a standard job. For more on recovering and validating draft geometry from cast parts, see our related guide on recovering draft angles from cast parts.

The QC loop closes with a post-machining scan of the finished casting, deviation-mapped against the design-intent model. Acceptance criteria: ±1.5 mm on blade faces, ±3.0 mm on the boss. Props that fall outside get flagged for re-machining or re-casting before delivery.

Timeline summary: Scan + CAD reconstruction: 3-5 business days. Foundry pattern CNC: 2-4 days. Casting lead time: 3-6 weeks for Ni-Al-Bronze, foundry-dependent. Total to a qualified replacement prop in hand: typically 5-8 weeks from scan day.

---

Other Ship Components That Follow the Same Workflow

The scan-to-CAD discipline we developed for propellers applies directly across the hardware list that marine yards deal with regularly:

Shaft flanges and couplings. Bolt-circle and face-flat tolerances are critical for alignment. We capture the installed shaft axis with our Trimble X7 and deliver an alignment report alongside the CAD model.

Rudder stocks and pintles. Corrosion-wasted bearing surfaces are a common find. The scan captures residual geometry so the machinist knows the exact material removal required before hard-facing and re-grinding - no guesswork, no undershooting the bore.

CPP hub mechanisms. Internal oil-distribution passages and servo geometry can’t be captured optically from the outside. For these, we cross-reference with CT scanning to capture internal features. See our companion article on CT scanning for internal features for how that workflow integrates with the optical scan data.

Anchor windlass drums and wildcats. Worn gypsy teeth on wildcat drums are a judgment call - recutting vs. full replacement. A deviation scan gives the surveyor real numbers rather than a visual estimate.

Historic vessel components. Museum ships, tall ships, and working schooners carry ironwork that was never documented to modern CAD standards. The same shrink-allowance and draft-angle discipline applies to original historic hardware where surface fidelity matters more than pitch tolerances.

Impeller and pump casings from auxiliary seawater systems. Same Ni-Al-Bronze alloys, same reflectivity challenge, same shrink-allowance requirement. If it’s a marine bronze casting and the original drawing is gone, the workflow is the same.

---

Cost Ranges and What Drives the Quote

Line-item add-ons that move the number:

• Foundry shrink-allowance build-in: +$400-$800
• 2D inspection drawing to ISO 484-1 format: +$600-$1,200
• On-site mobilization to dry-dock or shipyard: +travel and per diem (quoted by location - US East Coast, Gulf, and West Coast are all serviceable)
• Urgent 24-48 hr delivery: +30-40% on base price

The urgent premium is common. Dry-dock time is expensive - daily costs for a commercial vessel vary substantially by size and yard. Paying a scan premium to avoid extending a dry-dock window is straightforward math for any owner or fleet manager.

For a detailed breakdown of exactly what information and files to assemble before requesting a proposal, see our reverse engineering quote guide.

---

How to Prepare Your Component and Engage Capture

What to send for a remote quote. Four photographs minimum: suction face full spread, pressure face full spread, leading edge full span, hub stamp with any pitch or serial markings visible. Add the diameter, blade count, material if known, and the end-use (re-casting, direct CNC, drawing-only, or classification documentation). That’s enough for us to return a fixed-price proposal with a turnaround guarantee within 24 hours.

On-site vs. ship-to-lab. Props under roughly 80 kg can be pulled and shipped to our lab. This eliminates dry-dock scheduling complexity and is the fastest route for smaller commercial and recreational props. Larger props - anything above 1.5 m - require a field team. We mobilize to US East Coast, Gulf, and West Coast yards and work directly with the yard scheduler to book a scan window that minimizes the vessel’s out-of-service time.

Dry-dock coordination. Our on-site scan session for a 4-blade commercial prop runs 4-8 hours. The vessel does not need to be fully dry - waterline access is sufficient if the hub and blade roots are exposed. We bring our own reference target hardware, spray consumables, and power. The yard needs to provide only lighting and safe working access to the prop.

Data security. Raw scan files and all CAD deliverables are transferred via encrypted link. We do not retain or share any propeller geometry without a signed NDA. For vessels with sensitive hull geometry or proprietary propeller design, we are accustomed to operating under classification-society-level confidentiality requirements.

For the broader context of how scan data becomes an engineering model at every stage from capture through CAD, see our scan-to-CAD workflow explainer and our full overview of reverse engineering services. If you want to go deeper on the point cloud processing side, our point cloud to CAD services page covers the post-processing methodology in detail.

---

FAQ

Can you reverse engineer a propeller blade that has cavitation damage or missing tip material?

Yes - but the process is design-intent reconstruction, not copy-exact digitization, and that distinction matters enormously for the foundry. We fit the NURBS model to the undamaged inboard blade sections (r/R 0.2-0.8, where cavitation rarely reaches), and extrapolate tip geometry using the blade section series - Wageningen B-series coefficients, or the original OEM specification if the client can supply it. The deviation map we deliver documents exactly which zones were scanned from actual metal and which were reconstructed. The foundry receives an idealized model, not a digitized scar. Copying wear damage into a casting pattern produces a defective prop on the first pour.

What file format does the foundry or machine shop actually need?

Most foundries and CNC shops accept STEP (AP214 or AP242) as the universal neutral format. Some machining centers prefer Parasolid .x_t or IGES for legacy CAM systems. We deliver all three plus the QC inspection drawing as a PDF with ABS/ISO 484-1 blade section plots. A raw STL mesh is not sufficient for pattern-making - it lacks the parametric surfaces needed for foundry shrink scaling and parting-line analysis. Any shop asking for only an STL is planning to skip the shrink-allowance step, which will produce an undersized casting.

How long does the full scan-to-CAD process take for a commercial propeller?

Standard turnaround is 3-5 business days from scan completion to approved STEP delivery. Urgent 24-48 hr service is available at a 30-40% premium - the option most commonly used by vessels on tight dry-dock windows where the cost of an extra day on blocks dwarfs the scan premium. The scan session itself - on-site or in-lab - runs 4-8 hours for a 4-blade commercial prop, including setup, surface prep, all scan positions, and on-site registration QC.

Do you capture the surrounding hull geometry at the same time?

Yes, when we mobilize to a dry-dock. During the same session, we scan the A-bracket, strut, shaft axis, and rudder post with our Trimble X7. This context geometry feeds a clearance model that the naval architect uses to verify the replacement propeller clears the keel and rudder post under full deflection before the casting order is placed. Catching a clearance problem at the CAD stage costs nothing. Catching it after the prop is cast and delivered costs the price of a second casting plus delay.

Is the scan accurate enough to satisfy ABS or DNV classification requirements?

Our metrology-grade Creaform MetraSCAN handheld scanner achieves ±0.025-0.035 mm on blade faces. That comfortably exceeds ISO 484-1 Class I tolerances and approaches Class S. The deliverable package includes a deviation color-map PDF and blade-section pitch plots in the format classification surveyors recognize. The scan data provides the accurate, timestamped existing-conditions geometry that classification surveyors and the client’s own licensed professionals use to verify compliance with their documentation requirements. For new-build classification, the shipyard’s class society should be consulted on specific documentation requirements.

Can the same workflow produce a mold for a decorative bronze casting - a ship’s bell, capstan head, or architectural ornament?

Exactly the same discipline applies: surface prep, NURBS reconstruction, draft-angle analysis, shrink-allowance build-in, and a QC scan of the finished casting. The only meaningful difference is that ornamental pieces weight visual surface fidelity over ISO pitch tolerances. The same shrink-allowance and draft-angle rigor applies equally to historic ship hardware for museum restoration projects - the discipline is identical to propeller work.

---

Get a Fixed-Price Proposal in 24 Hours

Your vessel is on a schedule. Send us four photos of the propeller - suction face, pressure face, leading edge full span, and hub stamp - along with the diameter, blade count, material if known, and your dry-dock date. We will return a fixed-price proposal with a written turnaround guarantee within 24 hours.

For components that need a full reverse engineering services scope, or if you want to review what information to pull together before reaching out, see what to send for a reverse engineering quote. You can also read how point cloud data becomes a clean engineering model if you want to understand the full process before committing.

Contact us at weare-capture.com/services/scan-to-cad-reverse-engineering/ or email directly at [email protected]. We work with fleet operators, independent yards, naval architects, and foundries - and we’re used to the urgency that dry-dock schedules demand.