Quick comparison
| Method | Equipment buy-in (USD) | Training to competence | Finds | Role in your program |
|---|---|---|---|---|
| Visual (VT) | $200–1,000 basic; $2k–10k with pro borescope | 2–4 days; $500–2,500 | Surface defects only | Backbone — 100% of welds, two sets of eyes |
| Leak testing (LT) | $200–2,000 | <1 day; ~$0–500 | Through-wall leak paths | Every tank & watertight compartment |
| Ultrasonic (UT) | $20k–30k new; $6k–12k used | ~10 classroom days + 3–12 months mentored practice; $3k–8k in courses | Internal planar & volumetric defects | Spot checks on critical butt welds |
| Phased Array (PAUT) | $40k–80k | UT Level II first, + 5–8 days PAUT; $2.5k–5k | Same as UT, with imaging & records | The modern replacement for X-ray |
| Dye penetrant (PT) | $100–2,500 | 4–6 days total; $1k–2.5k | Surface-breaking cracks & porosity | Aluminum's primary surface method |
| Eddy current (ET) | $8k–20k | 5–8 days; $2k–5k | Near-surface cracks, through paint | Best for in-service (painted) survey |
| Radiography (RT) | $50k–150k+ plus licensing | 15+ classroom days + radiation safety + licensing | Volumetric (2D projection) | Skippable for aluminum boats — see below |
1. Visual Testing (VT)
The cheapest, fastest, and most productive method. On a well-run aluminum boat build, the majority of all weld problems are found by eyes and gauges.
| Equipment cost | Weld gauge set (Gal Gage etc.) $50–150 · good inspection light $50–200 · mirrors and magnifier $50 · USB borescope $150–800 · professional video borescope (Evident/Olympus, Waygate, Wohler, MEDIT) $2,000–10,000. Aim for ≥500 lux of light at the weld surface. |
|---|---|
| Training | 2–4 days: TWI's CSWIP visual weld inspector (~3 days), ASNT VT Level I/II (2–3 days), or an AWS Visual Inspection workshop. Course + exam typically $500–2,500. The full AWS CWI (Certified Welding Inspector) is far deeper — a week-long seminar plus years of experience eligibility — and is worth it for whoever leads your QC team. Also give every welder basic VT training; they self-inspect each pass. |
| Defects detected | Surface only: undercut, cracks, arc strikes, crater defects, surface porosity, spatter, underfill, poor profile/weld size (check with fillet gauges), misalignment (hi-lo), burn-through, distortion. Finds nothing internal. |
| Top sources | Gauges: Gal Gage Company · Borescopes: Evident (Olympus IPLEX), Waygate Technologies, Wohler, MEDIT · Training: TWI (CSWIP), ASNT, AWS · Standards: ISO 17637, AWS D1.2 (Structural Welding Code — Aluminum), AWS D3.7 (Guide for Aluminum Hull Welding), AWS B1.10. |
Fit: mandatory 100% coverage by the yard's QC and your inspector. The only prerequisites are good access, good light, and written acceptance criteria — define weld quality levels (e.g., ISO 10042) in your build spec before production starts.
2. Leak Testing (LT) — air pressure and water fill
The definitive functional test of tanks and watertight boundaries, and a classification-society requirement when built to class.
| Equipment cost | Digital manometer or gauge (Dwyer, Testo, Ashcroft) $100–600 · soap solution $20 · plugs, fittings, hoses $100–500 · access to compressed air · vacuum box (useful on lap/fillet seams) $500–3,000 · water-fill testing: hoses plus water. |
|---|---|
| Training | Half a day, in-house. Classification rules and your written procedure specify the pressures and hold times. Cost ≈ $0–500. |
| Defects detected | Only through-wall defects: interconnected porosity chains, through cracks, missed root penetration or lack-of-fusion that creates a leak path. It won't find buried flaws that don't yet leak — but it will absolutely find anything that would let water in. |
| Top sources | Procedures: classification society rules (ABS, DNV, Lloyd's Register) and the IMO High-Speed Craft Code · Instrumentation: general suppliers (Dwyer, Testo, Grainger). |
Fit: test every fuel, water, and grey tank, every watertight bulkhead, and every closed compartment — before fairing compound or insulation goes on. Typical air test per class rules is low pressure (~0.2–0.35 bar / 3–5 psi) with a soap solution; never exceed rule pressure, as thin aluminum panels and baffles can buckle. Hydrostatic fill to the specified head is the alternative.
3. Ultrasonic Testing (UT)
The workhorse for internal defects — but the steepest learning curve of any method here.
| Equipment cost | New flaw detector (Evident/Olympus EPOCH 650, Waygate USM Go, Sonatest Sitescan) $15k–25k · probes $100–400 each · couplant, aluminum calibration blocks, angle-beam kit — realistic all-in kit $20k–30k. Certified refurbished units $6k–12k. |
|---|---|
| Training | Level I: ~5 classroom days; Level II: ~5 more days plus exam — then the real cost: ASNT's recommended practice calls for hundreds of hours of documented supervised experience before a Level II is dependable. Courses $1,500–3,500 each. Realistically 3–12 months and $10k–20k per person before you can trust them on aluminum welds unsupervised. |
| Defects detected | Internal: lack of fusion (the #1 killer defect in aluminum GMAW), lack of penetration, cracks, oxide-film inclusions, clustered porosity. Also thickness gauging for corrosion-loss surveys in service. |
| Top sources | Instruments: Evident (Olympus), Waygate Technologies (Baker Hughes), Sonatest, Karl Deutsch · Training: Hellier (US), Lavender International (UK), TWI, manufacturer academies · Standards: ISO 17640, AWS B1.10 · Certification schemes: ASNT SNT-TC-1A / ACCP or ISO 9712. |
Fit: aluminum is harder to UT than steel — different sound velocity, weld-metal grain scattering, and lots of single-sided fillet geometry. Straightforward on plate butts; genuinely difficult on fillets. If your people will only do a handful of scans per boat, hiring a contractor is cheaper than building this skill in-house.
4. Phased Array UT (PAUT)
UT with an array of elements and electronic beam steering — S-scan images instead of one blip. This is the technology that makes skipping radiography realistic.
| Equipment cost | Instrument (Evident/Olympus OmniScan X3, Eddyfi Gekko or Mantis, Waygate Mentor PA, Sonatest Veo) $30k–55k · probe and wedge sets, scanners, software — realistic all-in $40k–80k. Rental (~few $k/month) and per-build contractor services are viable alternatives to owning. |
|---|---|
| Training | Requires conventional UT Level II first (see above), then a 5–8 day PAUT Level II course ($2,500–4,000) plus months of mentored use. Scan techniques for each joint family should be developed or reviewed by a contracted Level III consultant. |
| Defects detected | Everything UT finds, with far better coverage per pass, defect sizing (length and height), and — critically — permanent electronic records you can compare against at every future survey. Excellent detection of lack of fusion in butt welds. |
| Top sources | Evident (Olympus), Eddyfi Technologies, Waygate, Sonatest · Training: same NDT schools plus manufacturer courses · Acceptance: agree PAUT scope with your class surveyor in writing — ABS, DNV, and Lloyd's all accept PAUT for hull welds. |
Fit: the strongest modern substitute for X-ray on shell-plating butt welds. For a production line, consider a contractor campaign per boat or renting one instrument shared across builds.
5. Dye Penetrant Testing (PT)
Cheap, fast, license-free, and the go-to surface-crack method for aluminum (since MT can't work on it).
| Equipment cost | Solvent-removable three-can kit (cleaner/penetrant/developer, e.g., Magnaflux Spotcheck) $60–200 · fluorescent system (Magnaflux Zyglo) plus UV-A LED lamp (Magnaflux, Spectronics, UVATA) $400–2,500 · ongoing consumables are trivially cheap. |
|---|---|
| Training | Level I ~2 days, Level II ~2–3 days; $500–1,500 per level. The easiest formal certification on this page — train two or three staff members. |
| Defects detected | Surface-breaking only: fatigue cracks, crater cracks, HAZ cracks, porosity open to the surface, lack of fusion exposed at the weld toe. Finds nothing beneath the surface, and won't work through paint or moisture — the surface must be bare metal. |
| Top sources | Consumables: Magnaflux, Sherwin International, Met-L-Chek, MR Chemie · UV lamps: Magnaflux, Spectronics, UVATA · Standards: ISO 3452 series · Acceptance: ISO 10042 / class criteria. |
Fit: spot checks at stress concentrations — cross-deck/hull connections, engine beds, rudder and shaft areas, keel shoes, tank baffle corners, weld terminations — plus 100% of any repaired area. Respect the product's temperature limits (roughly 5–50 °C).
6. Eddy Current (ET)
A niche but genuinely valuable method for a boat after it's painted — and for a fleet operator that's you, long-term.
| Equipment cost | Instrument (Evident/Olympus Nortec 600, Eddyfi, Zetec, Foerster, UniWest) $8k–15k · weld-specific and paint-through probes extra · eddy current arrays (ECA) push the all-in figure toward $15k–20k. |
|---|---|
| Training | Level I ~3–5 days, Level II ~3–5 days plus exam; $1,500–3,500 per level. Interpretation is a real skill; add vendor product training for weld-through-paint applications. |
| Defects detected | Surface and near-surface cracks — including through several millimeters of paint, which is ET's superpower. Also conductivity changes, which can flag heat damage or over-tempered heat-affected zones. Cannot find buried volumetric defects; penetration is shallow. |
| Top sources | Evident (Olympus Nortec), Eddyfi Technologies, Zetec, Foerster, UniWest · Related technique worth a look: ACFM (alternating current field measurement) for sizing cracks through coatings. |
Fit: during a new build everything is bare metal, so cheap PT usually wins. ET earns its keep later in the boat's life: annual crack sweeps of hot spots without stripping paint, which matters for a vessel living in warm, saltwater Caribbean conditions.
7. Radiographic Testing (RT) — X-ray and gamma
| Equipment cost | Portable X-ray generators (COMET/Yxlon Eresco, etc.) $30k–90k · compact pulsed units (Golden Engineering XRS-3) ~$20k, limited to small parts · gamma-ray exposure devices (Ir-192/Se-75) $10k–25k plus isotope purchase, decay replacement, licensing, security storage · imaging: film + darkroom $5k–15k, computed radiography scanners (DÜRR NDT, Carestream, Fujifilm) $20k–50k, digital panels $25k–80k. A genuinely capable in-house RT setup is $50k–150k+ before recurring regulatory costs. |
|---|---|
| Training | Level I ~40 h + Level II ~40 h + a ~40 h radiation safety course (and a Radiation Safety Officer if you own sources): $5k–10k in classroom. The real barrier is regulatory: radioactive material licenses, dosimetry programs, exclusion zones, secure storage, and periodic authority audits — weeks to months of setup and permanent compliance overhead. |
| Defects detected | Volumetric defects shown as a 2D projection: porosity, inclusions, lack of penetration aligned with the beam. Weak on tight planar lack-of-fusion oriented across the beam — exactly the defect that matters in aluminum — and fillet welds (most of a boat) are practically impossible to interpret radiographically. |
| Top sources | X-ray: COMET/Yxlon, Golden Engineering · Imaging: DÜRR NDT, Carestream, Fujifilm · Training: Hellier, Lavender International, TWI · Regulators: your national radiation authority. |
Can we avoid radiography altogether? Yes — and for aluminum it's arguably an upgrade
Short answer: you can build a sound, certifiable aluminum catamaran without owning or operating a single X-ray or gamma source, and without any loss of quality — provided you cover the internal-defect question with PAUT/UT on critical welds. Here's why:
- The critical aluminum defect is lack of fusion — a planar defect. Radiography (a shadowgraph) is weak on tight planar flaws; UT and PAUT are strong on them. For this material, PAUT isn't a consolation prize for skipping RT; it's better at the defect that actually sinks boats.
- Fillet welds can't be meaningfully radiographed anyway. A hull is mostly fillets and T-joints. Those are covered by VT, by functional leak testing, and — where structurally critical — by UT/PAUT.
- Classification societies accept UT/PAUT in lieu of RT for hull structural welds under modern rules (ABS, DNV, Lloyd's Register). If you are building to class, get the substitution agreed in writing with your surveyor before signing the build contract. If your boats are recreational and unclassed, there is no statutory RT requirement at all — your NDT scope is whatever your spec says.
- The safety and regulatory burden is disproportionate for a boatyard: exclusion zones that shut down nearby work, licensed sources, dosimetry, security. In a port-area production yard this is a poor fit.
- Where RT still has a legitimate small role: welding procedure qualification (PQR) coupons. You don't need to own equipment for that — send coupons to a third-party lab, or use macro-etch/bend tests, which class rules generally accept as an alternative anyway.
One caveat: if any target market requires class notation or flag approval for passenger service, confirm the NDT extent with the class society and flag administration early — there can be specific spot-check requirements, but volumetric spot checks are almost always satisfiable by UT/PAUT.
A practical NDT program for a production aluminum catamaran
| When | What |
|---|---|
| Before production | Qualified WPS + PQR for each alloy/thickness/process combination (AWS D1.2 or ISO 15614-2); welder qualification (ISO 9606-1 or AWS D1.2); written acceptance criteria (ISO 10042 quality levels or AWS D1.2) in the build spec; a weld map plan. |
| During fabrication | VT: 100% of welds, by the yard's QC and independently by your inspector. LT: every tank and watertight compartment before closing-up and fairing. PT: all repaired welds plus sampled hot spots (cross-deck/hull connections, engine beds, rudder areas, keel shoe, baffle corners). |
| At defined milestones | PAUT: e.g., 100% of hull shell butt welds at/below the waterline and at chines, plus a sampling percentage of critical frame-to-shell joints — scope agreed with your designer and class surveyor. Records archived as the boat's baseline. |
| After any repair | Re-perform the applicable NDT on the repair and adjacent material before re-fairing. |
| Documentation | Weld map, NDT reports, welder certs, WPS/PQR, material mill certificates, instrument calibration records — the deliverable set in your contract. |
In-house team strategy: certifying your own inspectors under ASNT SNT-TC-1A requires a written practice and a Level III overseeing it — you can retain a consultant Level III (common and affordable) while your own staff do Level I/II work. Train at least two people in VT and PT (cheap, fast), and for UT/PAUT either send one person on the full multi-month path or — more sensibly at your volume — hire a certified contractor for the PAUT campaign on each boat.
Who should inspect: the shared-responsibility model
Your instinct is right, and it matches standard practice: the builder inspects because they own workmanship; the designer/buyer inspects to verify compliance with the design; and — if classed — a classification surveyor independently audits both. Dual inspection isn't distrust; defects found at the weld stage cost a fraction of defects found at sea trial, and your design team's findings feed back into production drawings.
- Put an Inspection & Test Plan (ITP) in the build contract: what gets inspected, by whom, by which method, against which acceptance criteria, with hold points your inspector must sign off before work proceeds.
- Tie NDT to the build sequence: once tanks are closed, foam is poured, and fairing is applied, access disappears forever. Milestone-based inspection is everything.
- Spell out who pays for rework and repeat NDT when defects are found (normally the builder).
- Keep the builder's QC records as contract deliverables — they are the baseline your future in-service surveys will compare against.
- Legally, responsibility for building to spec normally rests with the builder under the construction contract and warranty; your verification role is how you make that enforceable.
What to expect from Caribbean shipyards
Honest answer: expect visual inspection and tank/watertight testing to be done in-house; expect formal NDT (UT/PAUT, and sometimes even PT) to be subcontracted rather than held as an in-house capability. That's the normal pattern at small and mid-size aluminum boatyards almost everywhere, not just the region:
- Welders are usually qualified (ISO 9606-1 or AWS D1.2) and the yard's QC does VT — but a certified UT Level II on staff is rare outside big industrial yards.
- Where the local economy supports it, third-party NDT services exist: Trinidad has a real NDT contractor base thanks to its oil-and-gas sector; Florida/Miami firms service the region; St. Maarten and Antigua are refit hubs with marine surveyors. There is at least one dedicated aluminum catamaran builder in the region (e.g., Alunautica in Curaçao), and the Dutch/French islands connect to European survey networks.
- What this means for you: don't assume — contract. Bring your own trained QC person (VT + PT certified) to the yard, and schedule a regional NDT contractor's PAUT/UT campaign to coincide with production milestones. Make yard cooperation with both a contractual obligation with access rights.
Humanoid robots doing weld inspections: how long?
My honest estimate: 10–15 years before a general-purpose humanoid can reliably set up, couplant, scan, and interpret weld NDT in a cluttered boatyard environment — and 15–20+ years (if ever) before certification and liability regimes allow a robot to sign off an inspection on its own. Human certified technicians will remain in the acceptance-decision loop long after machines do the data collection.
- Purpose-built automation is already here and will reach you first: gantry and scanner-based PAUT in production lines, crawlers for hulls (note: standard magnetic crawlers don't stick to aluminum — suction/track systems are needed), camera drones, and AI-assisted flaw flagging in modern PAUT software. Within ~5 years, drone- or crawler-assisted in-service surveys will be commercially routine.
- The humanoid form factor is actually poorly suited to this job: the hard parts are mobility on/around a hull, access into cramped tank and hull compartments, and consistent probe coupling — problems a small snake robot or tracked crawler solves better than a biped.
- Practical move for a production line: don't wait for robots. Fixture-based automated PAUT of your repetitive joint families is available today, and often cheaper per boat than manual scans.
Is inspection one-time, or recurring?
Recurring — the build inspection is only the first one. Two degradation mechanisms specific to aluminum catamarans make periodic inspection genuinely important: fatigue (catamaran cross-deck bending concentrates load at beam–hull weld connections, and cracks initiate at weld toes and terminations) and corrosion (warm chloride-rich seawater drives pitting and crevice corrosion, and stainless fasteners create galvanic cells). Neither is visible under paint.
| Cadence | Scope |
|---|---|
| Baseline (delivery) | Full NDT record set from the build (weld map, PAUT data, tank test certificates) archived as the reference for all future surveys. |
| Annual (in service) | VT of all accessible welds and structure; PT or ET crack sweeps of fatigue hot spots (cross-deck/hull joints, engine mounts, rudder/shaft areas, keel, tank baffles). In tropical waters, haul out every 1–2 years to check below-waterline welds and plating for pitting/crevice corrosion and to inspect through-hull and fastener sites. |
| Every ~5 years | Deeper survey aligned with the insurance/class cycle: UT thickness gauging of plating, PAUT or UT sampling of critical butt welds compared against the build baseline. |
| Event-driven | After any grounding, collision, overloading, or repair/modification welding: VT + PT/UT of affected areas before and after re-fairing. |
| Commercial vessels | Classification society annual / intermediate / special (5-year) surveys and flag inspections (e.g., USCG annual inspections for passenger vessels) dictate the cadence. |
Because ET finds cracks through paint, it's the most economical annual method once the boat is in service; PT remains the gold standard when a surface can be stripped bare.