USS Shuba

The situation

USS Shuba is operating in contested waters. The mission requires her deck crane to work with two unmanned vehicles: an unmanned underwater vehicle, launched and recovered over the side, and an unmanned aerial vehicle handled on deck. The crane was built for neither.

Returning to port to have a fitting made is the expensive answer. The alternative is to design the fitting where the problem is, send the design and the instructions to a shore base that can make it, and collect the finished part at a rendezvous.

That is only credible if the design survives the trip as engineering data rather than as a drawing someone re-interprets. This page is the record of doing it that way, and of what the data said when it was actually opened.

What was already there

The ship exists as a real product structure, not a placeholder. Read directly out of Teamcenter:

The crane is modelled to component level, including 033189 Crane_Winch_Cable_Hook, a real hook with a shank and a safety latch. That mattered later.

The UUV is modelled with a carrier assembly and a cradle alongside the vehicle, which suggests it is handled in its carrier rather than slung bare. That changes what an adapter has to grab, so it became a requirement rather than an assumption.

The UAV is the gap. Searching the tier returns UAV objects, but they belong to other programmes: an observation UAV requirement set, an "Angel Owl" airframe, outer-mould-line designs and radar-cross-section analyses, none tied to this ship and none built to be lifted by a crane.

Design the interface, not the vehicle

With no UAV to measure, the choice was to specify the interface the vehicle must conform to, and let the vehicle follow. That keeps the adapter moving and turns the unknown into a written requirement instead of an accident of whichever model was to hand.

The result is a specification in Teamcenter, 034666, holding 35 numbered requirements as real content, not as loose objects beside it.

Group	Covers	Count
CS	Crane side, hook to adapter	8
UUV	Underwater vehicle interface	8
UAV	Aerial vehicle interface	9
LD	Load and dynamics	5
EN	Environmental	5

Every value says where it came from

A proposed number and a measured one look identical once written down, so each requirement carries its provenance in its own description, visible to anyone opening it:

EXERCISED

read back from a live call

DERIVED

reasoned from something exercised, with the reasoning stated

PROPOSED

a recommendation, defensible but not derived from this programme's data

OPEN

unknown, and it blocks detail design of whatever depends on it

The adapter

One below-the-hook lifting device carrying two vehicle interfaces. A bored eye plate on the vertical centreline takes the crane side. A longitudinal spine carries two lateral cross beams, and the four beam ends carry bored lift lugs in a rectangular pattern about the vehicle's centre of gravity. Underneath, a cradle plate and two rails form a channel that engages the UUV carrier.

It is driven by 22 named expressions, so every placeholder can be replaced with a real value and the part rebuilt from the script rather than re-modelled. Three interface checks run before any geometry is created, and the build refuses rather than producing a part that looks finished and does not fit.

Four lift points rather than one, because a UAV's centre of gravity moves with payload and fuel, and a single hardpoint concentrates the whole load into one airframe fitting. Four spread it into existing structure and control attitude while the deck is moving.

The measurement that changed the design

The crane hook was pulled out of Teamcenter and opened in NX rather than assumed. Two things came back that no amount of reasoning would have produced.

The second finding was dimensional. The hook's inner opening measures roughly 47 mm. The adapter's eye plate, as first drawn, was 40 mm thick. Seven millimetres is not a working fit once paint, tolerance and a moving deck are allowed for.

Thinning the plate would have been the wrong fix: it weakens the part to suit a fitting it was never meant to enter. Instead a shackle now goes between the hook and the adapter. The hook takes the round pin it was designed for, the adapter takes a clean pinned eye, and the shackle is a consumable that can be replaced without touching either.

That decision added two requirements, and the geometry was rebuilt from the script to suit: the eye plate went from 40 mm to 20 mm and its bore from 80 mm to 22 mm, sized to the shackle pin.

Checking it fits

The hook and the adapter were brought into one assembly and tested for interference at solid level.

The control is the point. A clean result from a test that cannot fail is worth nothing, so the adapter was deliberately driven into the hook to confirm the check detects interference before the passing result was believed.

The thread, end to end

How it was built

No interactive sessions. Teamcenter was driven over its JSON service layer, and NX ran headless, scripted through its open API. Every object named on this page came back from a call; none was transcribed from an intention.

• The adapter is generated by a build script, so the model and its parameters cannot drift apart. Change a value, re-run, get the part.
• The requirements were placed into the specification one at a time, because placing several in one call has been measured to write duplicates.
• Every write was verified from a separate process with its own login. A re-read inside the writing session passes whether or not anything was committed.

The operation, in motion

Every frame is rendered from the real ship model. The crane's luff, the ram's stroke, the hook's plumb hang and the rope's length all come from one definition shared by the renderer and the clearance checks, so the animation cannot show a pose the checks have not passed.

The UAV interface

The vehicle is the QX-250: a 250 mm wheelbase quadcopter, 0.50 kg all up, four five-inch propellers. A 0.50 kg aircraft is not craned off a ship on a 1500 kg A-frame, so the adapter's UAV side is a capture nest rather than a sling: somewhere to hold the aircraft while the ship moves.

The nest mounts on the after working deck at frame X 1178, 46 inches to port, on a pedestal clear of the roller rails, which occupy Z 142.6 to 147.1. There is no video of it in place: the nest is 10.2 inches across on a 127 foot ship, and the renderer will not draw an object that small inside a scene that large. Fitting the view tightly onto it returns a blank frame, and at a workable zoom it falls below the threshold entirely. The position is measured; the shot is not available.

Made ashore

Both new parts print. Neither video is a process simulation: they show material appearing in layer order and say nothing about scan strategy, thermal history, distortion, supports or build orientation.

Honest scope

What is real, what is provisional, and what has not been done.

Item	Status
Ship, crane and UUV geometry	real existing models, read not authored
Interface specification, 35 requirements	real native objects, placed as spec content
Hook units and envelope	measured from the hook's own part file
Adapter geometry	real model, checked in and trace linked
Adapter dimensions	placeholders every value tied to an open requirement
Hook inner opening, 47 mm	approximate scaled from a render, not a CAD measure
Shackle size	proposed follows a rated load that is still open
UAV vehicle	representative QX-250 chosen; wheelbase, props, arms and mass sourced, airframe geometry constructed
Stress and load analysis	not done
Interference against the full crane and ship	not done hook only
Manufacture ashore, work instruction	not done
Deployment and recovery videos	done rendered from the ship model, checks run before every frame
Ram bracket to reach the luff	real model, verified against the measured transom and rail
Ram bracket sizing	placeholders cheek and plate thickness; 2.1x pin load is an open stress check
UAV capture nest	real model, propeller clearance measured at 24.6 mm
Nest retention	not done no strap, latch or magnet designed
Additive build videos	geometric layer order only, not a process simulation
Nest shown in place on the ship	not done renderer will not draw a 10 in object in a 127 ft scene; position measured, shot unavailable

The adapter is a concept whose arrangement is a real design choice and whose numbers mostly are not. The model carries that warning as an attribute, so the caveat travels with the geometry rather than living only on this page.