Here's a complete HTML document covering all your questions. Prices are ballpark USD (verify with vendors), and I've included a practical "recommended path" section at the end.

```html
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>NDT for Aluminum Catamaran Welds — A Buyer's & Planning Guide</title>
<style>
  body { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Helvetica, Arial, sans-serif;
         line-height: 1.6; color: #222; max-width: 1000px; margin: 0 auto; padding: 20px; }
  h1 { border-bottom: 3px solid #0a5; padding-bottom: 8px; }
  h2 { margin-top: 2em; border-bottom: 2px solid #ccc; padding-bottom: 4px; }
  h3 { margin-bottom: 0.3em; }
  table { border-collapse: collapse; width: 100%; margin: 1em 0; font-size: 0.95em; }
  th, td { border: 1px solid #bbb; padding: 8px 10px; text-align: left; vertical-align: top; }
  th { background: #0a5; color: #fff; }
  tr:nth-child(even) { background: #f4f8f6; }
  .note { background: #fff8e1; border-left: 4px solid #f0b400; padding: 10px 14px; margin: 1em 0; }
  .key  { background: #e8f4fd; border-left: 4px solid #1976d2; padding: 10px 14px; margin: 1em 0; }
  ul { margin-top: 0.3em; }
  .small { font-size: 0.85em; color: #555; }
</style>
</head>
<body>

<h1>Non-Destructive Testing (NDT) for Aluminum Catamaran Welds</h1>
<p><em>A practical guide for an owner/design team that wants to independently verify a shipyard's welding.
All prices are approximate USD ranges and will vary by vendor, region, and specification. Training costs are per person.</em></p>

<div class="note">
<strong>Aluminum-specific note:</strong> Because aluminum is non-magnetic, Magnetic Particle Testing (MT) — the cheap,
common surface method used on steel — is <strong>not possible</strong>. Dye Penetrant (PT) is your primary surface-crack method.
Ultrasonic methods work very well on marine-grade aluminum (5083, 5086, 6061, etc.), which has low acoustic attenuation.
</div>

<h2>Quick Comparison Table</h2>
<table>
  <tr>
    <th>Method</th><th>Equipment Cost</th><th>Training</th><th>Best At Detecting</th>
  </tr>
  <tr><td><strong>VT</strong> — Visual</td><td>$100–$2,000 (up to $40k with videoscope)</td><td>1–3 days ($500–$1,500); AWS CWI ≈ 1 week, $2,500–$4,000</td><td>Surface defects, weld size/profile, fit-up</td></tr>
  <tr><td><strong>LT</strong> — Leak test</td><td>$100–$1,000 (helium systems $20k–$50k)</td><td>Hours to 2 days ($0–$1,500)</td><td>Through-wall leaks only</td></tr>
  <tr><td><strong>UT</strong> — Ultrasonic</td><td>$10k–$25k (kit)</td><td>≈2 weeks courses ($3k–$5k) + months of experience</td><td>Internal cracks, lack of fusion, porosity, thickness</td></tr>
  <tr><td><strong>PAUT</strong> — Phased Array</td><td>$40k–$80k</td><td>UT Level II first + 1–2 weeks ($2.5k–$5k) + experience</td><td>Same as UT, with images & permanent records</td></tr>
  <tr><td><strong>PT</strong> — Dye Penetrant</td><td>$50–$300 (aerosol kits)</td><td>2–3 days ($800–$1,500)</td><td>Surface-breaking cracks & porosity</td></tr>
  <tr><td><strong>ET</strong> — Eddy Current</td><td>$10k–$30k</td><td>≈2 weeks ($3k–$6k) + experience</td><td>Surface/near-surface fatigue cracks, through paint</td></tr>
  <tr><td><strong>RT</strong> — Radiography</td><td>$30k–$150k+ plus licensing & safety program</td><td>≈3 weeks ($4k–$8k) + government licensing</td><td>Volumetric defects (porosity, inclusions), film record</td></tr>
</table>

<!-- ============================ PER-METHOD DETAIL ============================ -->

<h2>Method Details</h2>

<h3>1. Visual Testing (VT)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $100–$2,000 for a complete kit: fillet weld gauges, hi-lo (misalignment) gauge, dial calipers,
      bright LED flashlight, inspection mirror, magnifier, camera. Add $10k–$40k only if you want a professional videoscope for
      inaccessible areas (inside tanks, frames).</li>
  <li><strong>Training:</strong> A formal VT Level II course is 16–24 hours (2–3 days, $500–$1,500). The gold-standard credential for weld
      inspection is the <strong>AWS CWI (Certified Welding Inspector)</strong>: a 5–6 day seminar plus exam, roughly $2,500–$4,000 all-in.
      One CWI on your team covers VT competence credibly.</li>
  <li><strong>Defects detected:</strong> Undercut, overlap, surface porosity, surface cracks, incorrect weld size/profile, excessive
      reinforcement, misalignment, arc strikes, spatter, incomplete groove fill, visible root problems. Studies consistently show most
      defective welds are caught visually — VT should be 100% coverage on every weld.</li>
  <li><strong>Top sources:</strong> G.A.L. Gage Co. (weld gauges); Evident/Olympus, Waygate Technologies, Vividia (borescopes);
      American Welding Society (CWI program, D1.2 code books).</li>
</ul>

<h3>2. Leak Testing (LT)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $100–$1,000 for practical boat work: low-pressure air regulator, calibrated gauge, soap/bubble
      solution. Helium mass-spectrometer leak detectors ($20k–$50k) are overkill for hulls — useful only for fuel tank or sealed
      compartment production QA if you want quantified leak rates.</li>
  <li><strong>Training:</strong> Bubble testing: a few hours of instruction. Formal LT certification courses run 2 days ($1,000–$1,500).</li>
  <li><strong>Defects detected:</strong> Only <em>through-thickness</em> defects: cracks that penetrate the plate, pinhole porosity
      chains, incomplete penetration that breaches the joint. It will not find embedded defects. Ideal for tanks, watertight
      compartments, plumbing, and hull envelopes.
      <span class="small">Caution: pressurize large flat compartments to only 1–2 psi — forces on big panels multiply fast.</span></li>
  <li><strong>Top sources:</strong> Swagelok "Snoop" (bubble fluid); INFICON, Pfeiffer Vacuum, Leybold (helium systems, if ever needed).</li>
</ul>

<h3>3. Ultrasonic Testing (UT, conventional)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $5k–$15k for a quality flaw detector (Olympus/Evident EPOCH 650, Waygate/Krautkramer USM Go+,
      Sonatest WAVE). Add probes ($300–$1,500 each), cables, calibration blocks (IIW/DSC blocks, $500–$2,000 each), couplant.
      Realistic working kit: <strong>$10k–$25k</strong>. Budget Chinese units ($2k–$4k) exist but support and code compliance are weaker.</li>
  <li><strong>Training:</strong> UT Level I = 40 hours (1 week, ~$1,500–$2,500); UT Level II = another 40 hours (~$1,500–$2,500),
      <em>plus</em> documented on-the-job experience (typically several months) before certification under ASNT SNT-TC-1A or ISO 9712.
      Interpretation skill is the long pole — budget real mentored practice time.</li>
  <li><strong>Defects detected:</strong> Internal cracks, lack of fusion, incomplete penetration, porosity, laminations in plate, and
      precise thickness measurement (corrosion monitoring). Works well on aluminum welds.</li>
  <li><strong>Top sources:</strong> Evident (Olympus), Waygate Technologies (Baker Hughes), Sonatest, Eddyfi, Dakota Ultrasonics
      (excellent value for thickness gauges, $2k–$5k).</li>
</ul>

<h3>4. Phased Array UT (PAUT)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $40k–$80k for instrument + probes + wedges + encoder scanner (Evident OmniScan X3/X4,
      Eddyfi M2M Mantis/Gekko, Sonatest Veo3, Waygate Mentor UT). Refurbished units can be found at $20k–$35k.</li>
  <li><strong>Training:</strong> Requires conventional UT Level II as a foundation, then a 40–80 hour PAUT course ($2,500–$5,000),
      plus mentored field time. Realistic path to competence: 3–6 months total.</li>
  <li><strong>Defects detected:</strong> Same as UT but with cross-sectional imaging (S-scans), encoded scan positions, and a
      <strong>permanent digital record</strong> of every weld inspected — extremely valuable in a production setting and in any
      dispute with a yard. Higher probability of detection than manual UT, and widely accepted by codes as an RT substitute.</li>
  <li><strong>Top sources:</strong> Evident (OmniScan), Eddyfi Technologies, Sonatest, Waygate, Zetec.</li>
</ul>

<h3>5. Dye Penetrant Testing (PT)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $50–$300 for aerosol kits (cleaner, penetrant, developer). Fluorescent PT with UV-A lamps:
      still under ~$3,000 for a full setup. This is the best cost-to-value ratio in all of NDT.</li>
  <li><strong>Training:</strong> Level II course: 16–20 hours (2–3 days, $800–$1,500), plus modest experience hours. Technique
      discipline (cleaning, 10–30 min dwell time, proper removal) is the main skill.</li>
  <li><strong>Defects detected:</strong> Surface-breaking defects only: fatigue cracks, solidification cracks, crater cracks,
      surface porosity, cold laps/lack of fusion open to the surface, grinding cracks. <strong>This is your primary crack-finding
      method for aluminum</strong> (since MT is impossible).</li>
  <li><strong>Top sources:</strong> Magnaflux (Spotcheck kits), Sherwin Inc., Met-L-Chek, Chemetall/Ardrox.</li>
</ul>

<h3>6. Eddy Current Testing (ET)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $10k–$20k for portable instruments (Evident NORTEC 600, Rohmann, Foerster); multi-technology
      units (Eddyfi Ectane) run $30k+. Probes $500–$3,000 each.</li>
  <li><strong>Training:</strong> Level I + Level II courses ≈ 80 hours total ($3,000–$6,000) plus experience. Impedance-plane
      interpretation takes real practice.</li>
  <li><strong>Defects detected:</strong> Surface and near-surface (to ~2–3 mm) cracks — the aerospace-standard method for finding
      fatigue cracks in aluminum, and it can work <em>through paint</em>. Also verifies alloy/heat treatment via conductivity, and
      measures coating thickness. Not a volumetric weld-inspection method; it complements PT for in-service crack hunting.</li>
  <li><strong>Top sources:</strong> Eddyfi, Evident (NORTEC), Rohmann, Foerster, UniWest, Zetec.</li>
</ul>

<h3>7. Radiographic Testing (RT)</h3>
<ul>
  <li><strong>Equipment cost:</strong> $30k–$150k+ depending on approach (digital detector arrays &gt; computed radiography &gt; film),
      plus survey meters ($500–$2,000), dosimetry service, shielded storage, and — for gamma sources (Ir-192/Se-75) — source
      licensing, transport, and leak-test obligations.</li>
  <li><strong>Training:</strong> RT Level I + II courses (80 hours, ~$3,000–$4,000) <em>plus</em> a 40-hour radiation safety course
      and government licensing/registration, which varies by jurisdiction and can be genuinely painful on small islands
      (importing and storing a radiation source).</li>
  <li><strong>Defects detected:</strong> Volumetric defects (porosity, slag, inclusions) with a permanent film/image record.
      Weak on planar defects (cracks, lack of fusion) unless the beam happens to align with the flaw.</li>
  <li><strong>Top sources:</strong> Waygate (Seifert X-ray), Comet/Yxlon, Teledyne ICM, DÜRR NDT (CR), Carestream NDT,
      QSA Global (gamma projectors).</li>
</ul>

<!-- ============================ QUESTIONS ============================ -->

<h2>Can We Skip Radiography Entirely?</h2>
<div class="key">
<strong>Yes — for welded aluminum boat structures, you can achieve equal or better inspection coverage without RT.</strong>
</div>
<ul>
  <li>RT's strength is volumetric flaws (porosity). UT/PAUT finds those <em>and</em> is far more sensitive to the dangerous planar
      flaws — cracks and lack of fusion — that RT often misses.</li>
  <li><strong>AWS D1.2 (Structural Welding Code — Aluminum)</strong> and the major classification societies (DNV, ABS, Lloyd's, BV)
      all accept UT/PAUT in lieu of radiography for weld examination.</li>
  <li>Sensible combined coverage: <strong>100% VT + PT on critical joints + UT/PAUT on butt welds and critical structure</strong>.
      Add leak testing for tanks and watertight compartments. This covers the full defect spectrum RT would find, plus more.</li>
  <li>Keep a relationship with a third-party NDT services company for the rare case a spec or customer demands film —
      subcontracting one day of RT is far cheaper than owning the safety/licensing burden.</li>
</ul>

<h2>Will Caribbean Shipyards Do Their Own Inspection?</h2>
<ul>
  <li><strong>Varies enormously — verify, don't assume.</strong> Larger commercial yards (fast-ferry and crew-boat builders) typically
      have in-house VT and PT, certified welders, and subcontract UT/RT. Smaller yards may have little beyond "the foreman looks at it."</li>
  <li>Trinidad &amp; Tobago and Puerto Rico have mature industrial NDT service sectors (from oil &amp; gas), so third-party NDT
      contractors are available in the region even if the yard itself has none.</li>
  <li><strong>Your instinct is correct and matches standard practice:</strong> responsibility should be shared —
    <ul>
      <li><strong>Shipyard (QC):</strong> qualifies welders and procedures (AWS D1.2 welder certs or ISO 9606-2; WPS/PQR per D1.2 or
          ISO 15614), performs 100% VT and process control, warrants workmanship.</li>
      <li><strong>Owner/design team (QA):</strong> independently verifies with your own inspectors — you are checking <em>their</em> work,
          and you carry the long-term consequence of latent defects. Never let the yard's inspector be the only eyes.</li>
      <li><strong>Class society (if classed):</strong> the surveyor enforces NDT per class rules as a third layer.</li>
    </ul>
  </li>
  <li>Put it in the contract: an <strong>Inspection &amp; Test Plan (ITP)</strong> with hold points, defined NDT coverage by method
      (e.g., 100% VT, PT on all critical welds, UT/PAUT on a defined % of butt welds), acceptance criteria (AWS D1.2 or
      ISO 10042 quality level), and a documentation/weld-traceability package delivered with the boat.</li>
</ul>

<h2>When Will a Humanoid Robot Do This?</h2>
<ul>
  <li><strong>Robotic weld inspection already exists — just not humanoid.</strong> Magnetic crawlers with PAUT/UT arrays
      (Eddyfi/Inuktun, Gecko Robotics), drone-based visual and thickness inspection, and automated PAUT scanners are in
      commercial use today. They need a human operator and a certified human to sign the report.</li>
  <li>AI <em>interpretation</em> assist (auto-flagging of PAUT and RT indications) is being deployed now and will mature over
      the next 3–5 years.</li>
  <li><strong>Humanoids specifically:</strong> the humanoid form factor (Optimus, Figure, Atlas, etc.) offers no real advantage for
      inspection over crawlers/drones. A demo of a humanoid carrying a PT kit or UT probe around a hull is plausible within
      ~3–5 years; trustworthy, autonomous field use is more like <strong>8–15 years</strong>.</li>
  <li>The hard bottleneck isn't the robot — it's <strong>certification and liability</strong>. Standards (ASNT SNT-TC-1A, ISO 9712)
      require inspection results to be evaluated and signed by certified humans. Expect "robot collects, AI screens, human
      certifies" to be the model for the foreseeable future.</li>
</ul>

<h2>One-Time at Build, or Recurring?</h2>
<div class="key">
<strong>Recurring — and this matters more for aluminum than for any other common boatbuilding material.</strong>
Aluminum has no fatigue limit: under cyclic loading (and a catamaran's cross-structure is heavily cycled), microscopic
flaws at weld toes can grow into cracks over years of service. Welds that pass at build can and do develop fatigue cracks later.
</div>
<table>
  <tr><th>Timing</th><th>Recommended Inspection</th></tr>
  <tr><td><strong>At build (baseline)</strong></td>
      <td>100% VT of every weld; PT on all critical/high-stress joints; UT or PAUT on critical butt welds (sample % per class or spec);
          leak test all tanks and watertight compartments. Photograph and log everything — this is your baseline.</td></tr>
  <tr><td><strong>Annual haul-out</strong></td>
      <td>VT of the hull and known hot spots; PT anything suspect.</td></tr>
  <tr><td><strong>Every 2–5 years</strong></td>
      <td>UT thickness gauging (corrosion/wastage trending); targeted PT or ET on high-stress zones: cross-beam/cross-deck to hull
          connections, mast step, chainplates, engine beds, shaft struts, transom, chines, appendage attachments.</td></tr>
  <tr><td><strong>After events</strong></td>
      <td>Grounding, collision, rigging failure, or exceptionally heavy weather → targeted VT + PT/UT of affected structure
          before returning to service.</td></tr>
  <tr><td><strong>If classed</strong></td>
      <td>Follow the society's survey cycle (annual / intermediate / special survey ≈ 5-year), which includes NDT and thickness
          measurement requirements.</td></tr>
</table>
<p class="small">Keep all records with consistent measurement locations and photos so cracks and thickness loss can be trended over the vessel's life.</p>

<h2>Recommended Practical Path for Your Team</h2>
<ol>
  <li><strong>Phase 1 (do immediately, ~$3k–$6k):</strong> Full VT kit + PT supplies; send one or two people through
      <strong>AWS CWI</strong> certification and PT Level II. This covers the highest-value inspections at trivial cost.</li>
  <li><strong>Phase 2 (contract):</strong> Write the ITP, NDT coverage, acceptance criteria (AWS D1.2 / ISO 10042), and documentation
      requirements into the build contract. Retain a third-party NDT firm for PAUT at build — roughly a few hundred dollars
      per day-rate mobilization versus $80k+ to own the capability.</li>
  <li><strong>Phase 3 (optional, year 1–2):</strong> If you're building multiple boats, buy a conventional UT flaw detector and train
      a Level II — it pays for itself across a production run and covers in-service thickness gauging for the fleet.</li>
  <li><strong>Skip:</strong> RT (safety/licensing burden, inferior for planar flaws) and ET (buy later only if in-service fatigue
      crack hunting through paint becomes a priority).</li>
</ol>

<p class="small"><em>Disclaimer: Prices and course offerings are approximate and change; obtain current quotes. Have an ASNT NDT
Level III or equivalent review your procedures and acceptance criteria for your specific design and jurisdiction.</em></p>

</body>
</html>
```

**A few key takeaways from the document:**

1. **Start cheap and high-value**: VT (ideally via an AWS CWI on staff) + PT covers most real-world defects for under $6k total.
2. **Skip RT** — UT/PAUT is code-accepted, safer, and better at finding the cracks and lack-of-fusion that actually kill aluminum structures.
3. **Recurring inspection is essential for aluminum** — unlike steel, aluminum has no fatigue limit, so a catamaran's cross-structure welds need periodic PT/UT checks for life.
4. **Robots**: specialized inspection crawlers exist *today*; humanoids are the wrong form factor, and certification rules (human sign-off) will gate autonomy far longer than the hardware will.