CNC machining in subsea manufacturing

• 1
  Why CNC machining in subsea manufacturing is a tolerance problem first
• 2
  Alloys that survive salt water, and how they cut
• 3
  How 5-axis work changes the fixture and the flatness you get
• 4
  How you prove a subsea part is good before it ships
• 5
  When a machined part is the wrong choice
• 6
  Alloy and process fit for subsea parts

Seawater at 300 m depth sits near 30 bar. At 3,000 m it is closer to 300 bar. Every flat cover, connector body and valve housing has to hold that load without deflecting into the seal. CNC machining matters here because the seal groove depth, the face flatness and the bore roundness are cut features, not cast ones. A groove that runs 0.03 mm deep on one side leaks.

Subsea work is less about exotic geometry than about repeatability at low volume. A run might be four valve bodies or two hundred sensor housings. The drawing usually calls out tight tolerances on a few critical features and leaves the rest loose. That is the right way to write it, because grinding every surface triples cost for no gain.

Most of the value a machine shop adds is in the setup: how the part is held so the critical features are cut in one fixturing, and how the datum scheme matches the way the part is later inspected. Get that right and the tolerances fall out. Get it wrong and the shop spends a week chasing a stack-up that was never going to close.

• 1
  Sealing facesFace flatness and groove depth drive leak rate more than surface roughness alone.
• 2
  Bores and journalsRoundness and straightness matter; a few microns of lobing wears seals fast.
• 3
  Wall thicknessSet by pressure class, corrosion allowance and the minimum the tool can hold.

Duplex and super duplex stainless are the default for wetted parts: 2205 for general service, 2507 where chloride stress corrosion is a real risk. Both work-harden. A roughing pass that rubs instead of cuts will harden the surface to the point where the finishing insert wears out in minutes. Sharp tooling, positive rake and a constant feed that stays under the hardened layer solve it.

316 and 316L remain the practical choice for brackets, covers and non-wetted hardware, and they machine far more predictably. 17-4PH gives you strength close to a low-alloy steel with much better corrosion behavior, and it heat-treats to H900 or H1075 after machining with predictable shrink. Titanium Ti-6Al-4V shows up on lightweight frames, but it is a poor thermal conductor, so heat goes into the cutter.

Nickel alloys such as Inconel 625 and 718 appear where temperature and chloride meet, typically on wellhead hardware. They are slow. Expect low surface speeds, heavy flood coolant and a separate roughing strategy. On a 4,000 mm maximum processing size machine, a large Inconel flange can take a full shift of roughing before it sees a finishing pass.

• 1
  2205 / 2507 duplexHigh chloride resistance; control feed to avoid work-hardening.
• 2
  316 / 316LForgiving to machine, good general seawater performance.
• 3
  17-4PH (SUS630)Machined in condition A, then aged; plan for shrink.
• 4
  Ti-6Al-4V, InconelLow speeds, high coolant pressure, short tool life.

A sealing face cut on a 3-axis machine in two setups almost always shows a step at the joint. It might be 5 μm, and it might still leak. Turning the part once on a 5-axis center so the bore, the flange face and the O-ring groove come from the same datum removes that step entirely. That is the main reason subsea housings get quoted on 5-axis.

Thin walls are the other reason. A pressure housing with a 6 mm wall will move when you unclamp it if the roughing and finishing passes both run at full load. Rough leaving 0.5 mm, stress-relieve or simply let the part sit, then finish with light radial cuts. On a Ø400 mm rotary table a round housing can be finished in one continuous pass, which keeps wall thickness even all the way around.

Sealing faces usually land at Ra 0.2–0.8 μm. General mating surfaces are fine at Ra 0.8–1.6 μm, and non-critical brackets at Ra 1.6–3.2 μm. Specify the roughness only where a seal, a bearing or a gasket sits. A drawing that calls Ra 0.4 μm on every face will be quoted high and machined slowly for no functional reason.

• 1
  One-setup critical featuresBore, face and groove from a single datum on a 5-axis center.
• 2
  Rough, relax, finishLeave 0.5 mm, release stress, then take light finishing cuts.
• 3
  Roughness where it mattersRa 0.2–0.8 μm at seals only, not across the whole part.

Subsea parts are hard to service, so inspection has to substitute for access. A workable scheme is a raw material check when the bar or forging arrives, in-process monitoring of the critical dimensions as the part is cut, and a final inspection before shipment. That final pass is worth more than any certificate on the wall, because it catches the one dimension that drifted on the last part of the run.

CMM reports on the sealing faces, the bore and the bolt pattern are the usual deliverable. If a drawing has a true position callout on a bolt circle, that is the number to verify, not the individual hole diameters. On fittings and manifolds, a pressure test at the specified class is often required before the part is released.

Traceability is part of the same conversation. Mill certificates for the alloy heat, plus a report that ties the machined part to that heat, are standard for subsea hardware. If a customer needs that mapping, say so at the quote stage. Reconstructing material traceability after the chips are gone is not possible.

• 1
  Raw material checkChemical and hardness verification against the mill certificate.
• 2
  In-process monitoringCritical features measured while the part is still in the machine.
• 3
  Final inspectionFull dimensional report before shipment; reports on request.

CNC machining wins on low volume, tight tolerance and complex internal features. It loses on hollow shapes with thin uniform walls in large quantity. A cast or forged near-net shape that only needs finishing passes will beat a part hogged out of solid plate on both cost and grain structure, especially on a heavy 4,000 mm housing where most of the plate becomes chips.

Very large pressure vessels are another boundary. If the finished part is bigger than the available travel, you are looking at fabrication, welding and then machining of the critical faces on a mill-turn or a large boring setup. That is a valid route, but it changes the drawing: you need machining allowance on the weldment and a datum you can trust after welding distortion.

There is also a corrosion boundary that machining cannot fix. If the alloy choice is wrong for the chloride level and the temperature, a perfect surface finish just delays the failure. Engineering the material and the machining process together, at the drawing stage, is what keeps a subsea assembly in service.

• 1
  Thin uniform walls, high volumeNear-net casting or forging plus finishing usually wins.
• 2
  Beyond machine travelFabricate and weld first, then machine the critical faces.
• 3
  Wrong alloyBetter machining cannot rescue a corrosion design error.