7m Aluminium Autonomous USV for Southern Ocean — Desktop Engineering Review Request

Hello everyone,

I am working on a concept design for a 7.4m autonomous USV intended for long-duration patrol in the New Zealand Southern Ocean (latitudes 35S to 52S). I have no formal naval architecture background — this has been a one-person effort using AI tools to assist with the engineering calculations. I am here because I need experienced eyes on this work before it goes any further.

I would be grateful for brutal, honest feedback. If something is wrong, I need to know now.

---

The Vessel: Kereru

A solar/wind/battery-powered autonomous monohull, loosely inspired by the Ocius Bluebottle class but adapted for the much harsher conditions of the Southern Ocean (Sea State 5-6 operational requirement).

Principal Dimensions

LOA: 7.40 m
LWL: 6.80 m
Beam: 1.80 m
BWL: 1.60 m
Draft (hull): 0.38 m
Draft (with keel): 1.80 m
Depth: 1.05 m
Displacement (full load): 1,200 kg
Freeboard (min): 0.67 m

Hull Form

- Marine aluminium 5083-H321
- Monohull, round bilge, 30 deg deadrise
- Cb = 0.28, Cp = 0.58, Cm = 0.72, Cwp = 0.68
- L/B = 4.25, B/T = 4.21
- Wetted surface approx 9.3 m2

Stability (IMO Criteria)

GM: 0.796 m (req >= 0.50 m)
GZ at 30 deg: 0.493 m
GZ max: 1.755 m at 60 deg
Vanishing angle: 150 deg
Self-righting: Yes (150 kg lead keel bulb)
Wind heel ratio (25 kn): 3.3x (req > 1.5x)

All 8 IMO intact stability criteria are met.

Propulsion and Energy

- Primary: Retractable rigid wing-sail, 4.5 m2 (NACA 0012, AR 2.0)
- Secondary: 1.5 kW electric thruster (pod mount, Wageningen B3-50 propeller, D=400mm)
- Tertiary: Flipper-rudder (wave energy, included in full design only)
- Solar: 3.0 m2 total (deck + wing-sail surfaces), 22% efficiency
- Wind generator: VAWT, 160W at 40 kn
- Battery: 25 kWh LiFePO4 (520 Ah at 48V), 80% DoD = 20 kWh usable

Energy Budget at 3 kn Cruise

Total electrical demand: 271 W (150 W sensors + 121 W thruster at eta=0.456 for pod mount).

The wing-sail reduces thruster demand mechanically but does not charge the battery. Monte Carlo endurance simulations (10,000 runs per scenario, daily weather variability from ERA5/NIWA data, no flipper) give:

35S (Auckland) Summer — MC P50 Endurance: 15 days
35S Winter — 7 days
46S (Dunedin) Summer — 17 days
46S Winter — 8 days
52S (Campbell Is.) Summer — 19 days
52S Winter — 8 days

These are honest numbers — the Monte Carlo is significantly more conservative than the deterministic calculation because real weather has multi-day calm spells that drain the battery faster than averages suggest.

Structural Scantlings (ISO 12215-5)

Bottom: 3.5 mm plate, 400 mm frame spacing
Side: 2.6 mm plate, 500 mm frame spacing
Deck: 2.6 mm plate, 600 mm frame spacing
Keel box: 10 mm (140x200 mm section), butt weld

Frames: web 60x4 mm, flange 40x5 mm, 18 frames total.

Keel Root (DNV-GL)

Design BM: 3,380 Nm (dynamic, 2g + hydro, SF=2.0)
Root bending stress: 20.5 MPa (utilisation 0.14 of welded yield)
8x M16 SS316L bolts, utilisation 0.34
Fatigue damage ratio: 0.631 (DNV detail cat D, 10-year life)

Resistance

ITTC 1957 friction + empirical residuary (Holtrop-Mennen is out of range for Cb=0.28 at this size). Froude number at 3 kn cruise = 0.189; hull speed approx 6.4 kn (Fn=0.4).

Seakeeping

Roll natural period 1.65 s (well above typical Southern Ocean wave periods — low resonance risk). Operational to SS4, survival to SS6+. Capsize probability: 0.000% through SS8 in Monte Carlo simulation (self-righting to 150 deg).

---

What Has Been Calculated

42 desktop calculations have been completed, each as a standalone Python script with text reports and SVG/PNG output. Summary: 38 PASS, 2 FAIL, 1 WARNING, 1 CONDITIONAL.

The calculations cover:
1. Hull geometry and hydrostatics (with independent Simpson's 1/3 cross-check)
2. Intact stability and GZ curve (8 IMO criteria)
3. Structural scantlings (ISO 12215-5)
4. Resistance and power (ITTC 1957)
5. Energy budget by latitude and season
6. Propeller sizing (Wageningen B-series)
7. Wing-sail aerodynamics (NACA 0012 finite wing)
8. Monte Carlo endurance (10K simulations per scenario)
9. Seakeeping (strip theory, 1-DOF RAO, Bretschneider spectrum)
10. Keel root structural analysis and fatigue (DNV-GL, Palmgren-Miner)
11. Longitudinal hull girder strength
12. Iridium/AIS communication link budgets (ITU-R P.676/P.618)
13. Thermal balance (hull-as-heatsink model)
14. Capsize probability (Monte Carlo, nonlinear roll)
15. Sensitivity analysis (14 parameters)
16. Battery and solar degradation (10-year projections)
17. Biofouling drag penalty
18. Sea anchor loads
19. Rudder sizing (NACA 0015, DNV-GL)
20. Mast structural analysis (cantilever, VIV, fatigue)
21. Galvanic corrosion and cathodic protection
22. Electrical system (48V DC bus, cable sizing)
23. Reliability MTBF (MIL-HDBK-217F)
24. FMEA (MIL-STD-1629A)
25. Green water loading
26. Watertight integrity (damage stability)
27. Weight budget reconciliation
28. Emergency operations and degraded modes
29. Plus several operational analyses (patrol coverage, collision avoidance, navigation, etc.)

Known Failures and Weaknesses

I want to be upfront about the problems:

- Reliability (baseline): The original architecture only achieves R(14-day) = 55.3%. A redundancy redesign (dual actuators, N+1 power electronics, +29 kg, +NZ$18.5K) brings this to 95.3%. The baseline architecture is not adequate.
- Damage stability: FAIL. Negative GM after any single-compartment flooding. This is expected for a 7m vessel without subdivision — accepted as a design limitation.
- FMEA: CONDITIONAL. 48 failure modes identified, 22 critical, top RPN = 480. The top 10 RPNs must be reduced below 200 before production.
- Green water: 33% deck wetness at SS3. All deck equipment needs IP67/IP68.
- Keel fatigue: D = 0.631 at 10 years. This passes (D < 1.0) but the margin is not large.
- Propeller efficiency: Corrected from eta=0.511 (saildrive) to eta=0.456 (pod mount, w=0.08). This makes the energy budget tighter than originally designed.

---

What I Would Like Reviewed

I am specifically asking for feedback on:

1. Hull form — Is a Cb of 0.28 with 30 deg deadrise reasonable for a 7m monohull at Fn=0.19? Does the L/B of 4.25 make sense for Southern Ocean conditions? Any concerns about the round bilge form at this size?

2. Stability — The GZ curve shows max GZ of 1.755 m at 60 deg with a vanishing angle of 150 deg. The 150 kg lead keel bulb provides self-righting. Does this look right for this displacement and beam? Is the GM of 0.796 m appropriate?

3. Structural scantlings — 3.5 mm bottom plate in Al 5083-H321 at 400 mm frame spacing for a 1,200 kg vessel. Is this adequate for Southern Ocean slamming? The longitudinal FoS of 78 seems extremely high — is that normal for a vessel this small, or did I make an error?

4. Energy budget — The balance is tight. At 52S summer the system is near break-even; in winter everywhere there is a net electrical deficit. The Monte Carlo P50 of 7-19 days is the honest number. Are there energy sources or efficiencies I am missing? Is the eta=0.456 for a pod-mounted thruster reasonable?

5. Fatigue life — Keel root D=0.631, mast D=0.594, both at 10 years. These pass but are not generous. Are the S-N categories (DNV detail cat D for the keel, cat C for the mast) appropriate for these joint types?

Any other issues you see — please flag them. I would rather hear bad news now.

---

Calculation Scripts

All 42 Python calculation scripts are available. They require only Python 3.10+, numpy, and matplotlib (no proprietary software). If anyone wants to check the math, I am happy to share the complete package. Each script is self-contained and produces both a text report and drawings.

Standards referenced: ISO 12215-5, DNV-GL Pt.3 Ch.13/15, IMO intact stability criteria, ITTC 1957, ITU-R P.676/P.618, MIL-HDBK-217F, MIL-STD-1629A, ABYC E-11/TE-4, IEC 60092/62619.

Attached drawings:
1. Lines plan — profile, body plan, waterplane views
2. GZ stability curve — with dynamic stability areas and IMO criteria
3. General arrangement — side and top views
4. Energy budget chart — generation vs demand by latitude and season
5. Longitudinal strength — shear force and bending moment diagrams

Additional drawings (resistance curves, seakeeping RAO, keel structural analysis, Monte Carlo endurance distributions, propeller efficiency maps, etc.) are available on request.

Thank you for taking the time to look at this. I know it is a lot of material for a forum post. I tried to keep it to the essentials while being honest about the limitations.

Any feedback — from "this looks reasonable" to "you have fundamentally misunderstood X" — is welcome and appreciated.

 

A retractable wing sail is a really complicated mechanical system. Even in crewed vessels they are problematic. Also, your tertiary propulsion, flipper-rudder is complicated and inefficient. Keep it simple. Solar and wind are fine. However, propellers are more efficient than flippers. For wind power, a furling sail would be simpler. Keep in mind that all the electrical and electronics will be operating in a wet environment. The power requirement is grossly underrated. The vessel will need more power than that just to stay in place in Sea State 5. Also, ind those conditions, the sail will get no wind when the vessel is in the trough of the waves, so zero power. Consider a reasonable worst case scenario. For example, a week of Sea State 7 with overcast skies.

 

Thank you gonzo, this is exactly the kind of feedback I was hoping for. Some context on how we got here: the design was loosely inspired by the Ocius Bluebottle — they use a rigid wing sail and flipper-rudder on their USVs in Australian waters. We tried to adapt that concept for the Southern Ocean, which is obviously a much harsher environment. Your feedback confirms what I suspected — what works in moderate Australian conditions doesn't necessarily translate to 50S latitudes. New Zealand recently acquired two Bluebottle USVs from Ocius. Our project started by asking: what would it take to build something similar but designed specifically for Southern Ocean conditions from the ground up? That's how the wing sail and flipper ended up in the initial design — and why we're now reconsidering them based on feedback like yours. Wing sail → furling sail. I agree — the retractable rigid wing is over-engineered for an unmanned vessel, especially one that needs to survive SS6+. A furling soft sail (jib on a Harken furler) would be mechanically simpler, fewer failure modes, and easier to maintain. We'll switch to this. Any recommendation on sail area for a 1,200 kg vessel in these conditions? Flipper-rudder — removed. Agreed, it added complexity for marginal gain. We didn't include it in the Monte Carlo endurance numbers anyway. It made sense for Bluebottle in calmer seas, but down here reliability matters more than squeezing out extra watts from wave energy. Power budget — your most important point. To clarify our concept of operations: this vessel is not designed to hold station in SS5. In heavy weather (SS5+), she deploys a sea anchor, retracts the mast, and drifts in survival mode at ~80W (sensors + comms only). Station-keeping is only attempted in SS3 and below. The 271W budget is cruise power at 3 knots in moderate conditions. That said, your worst-case scenario (a week of SS7, overcast) is real for the Southern Ocean. Our 25 kWh battery at 80W survival gives ~10 days — but that assumes no solar at all. The question is whether the vessel needs to do anything useful during that week, or just survive and resume patrol when conditions ease. Wind shadow in wave troughs — good point, we haven't modelled this. For SS5 (Hs 3.25m) our sail is 4.5m above waterline, so partial shadowing is likely in the troughs. This would reduce effective sail power. Do you have a rule of thumb for effective wind reduction in heavy seas for small vessels? What power-to-displacement ratio would you consider adequate for a vessel like this in the Southern Ocean?

 

Hello @redice, welcome to the forum.
The stability curves are really strange. The initial stability is very low, and the shape of the GZ curve can't be right. I think I'm not mistaken, but just in case, I suggest you ask an expert to review your calculations, at least the stability ones.
Have you calculated the transverse stability curves? How did you arrive at the GZ curve?

Good luck with the project.

 

An autonomous vessel, especially with a sail probably has to be fully self-righting through 360 degrees. A 7 m boat will readily encounter a wave environment high enough to capsize it and so far the Coast Guard hasn't established a program at the National Motor Lifeboat School to qualify an AI as a surfman.

 

1. I use AI in engineering like tasks and I sucks.(both chatgpt and Claude paid versions, had paid Google for a whole but that was quite bad).
It will cling to details (that can be also wrong) while the bigger picture has major issues.

My 1st questions are:
Why aluminum and why so heavy. Battery is ~250kg solar not that much.

The wing sail is nice free propulsion but anything sticking out will be a big failure point when tumbling down from 9m wave crest.
Aluminum is good boat material but not a good small boat material. Fiberglass would give more shape freedom, lower resistance (not massive factor at low speeds but at low watts still a big deal).
The boat lines look quite odd. I would think plywood rowboat shape would be a good starting boat if sticking with alu. At least 4 plates (2 shines per side, my English might be failing here). Naturally fully self righting but you don't need 150kgvkeel for that.
The keel seems excessive imo - even if you keep the sail, the sails should be small to not have as much lever arm when crashing down on it.
You are already counting on it being retracted in bad weather (I assume) so you are down to solar in winter storms.
I think your solar estimates are bit pessimistic. You do get some even in the winter and overcast. What was your electronics wattage budget?

 

Interesting project.

I am only a casual observer, not an expert and not a sailor. I read your initial post and my first thought was I would not want to be on that boat; good thing it is autonomous.

I looked up the Ocius and it is an underwater vessel, so that bothered me when used as inspiration. The underwater vessel has built in protections from rough seas and somewhat from piracy. I don’t think an autonomous vessel lines up well with sailing due to complexity.

How did you decide 7m was the best size? For me, the 7m size limits deckspace that could be used for solar array. In the simplest context, a solar array 2m wide at midships would mean a ship length of 14m for 7:1 l/b. The vessel would not require a great deal of freeboard unless the use case demanded.

Anyhow, I won’t speak to many of the details, but I don’t understand how you arrived at the size.

Also, just as a concept, I could see how a sail could be used autonomously in conjunction with propellor generators. Basically, rather than used all the time, the sail comes up when the system is low on power and conditions allow and the batteries get charged and the sails remain usually unused.

I think the danger of AI is it is generally overly positive and often does not tell the user no enough. And here, your design is overly complex, and the size seems odd to a casual observer.

Maybe justify the size here.

I’ll bow out and let experts speak if they find it interesting.

 

If you can find any mention of piracy on the Southern Ocean by all means fit a burglar alarm or make the boat submersible.

 

I wouldn’t laugh too hard, an autonomous vessel in the US would not last long before someone went onboard and stripped it and we don’t call it piracy, but theft.

 

These have been around for a few years and seem to be successful Autonomous sailing drones deployed to protect marine life https://interestingengineering.com/innovation/autonomous-sailing-drones-deployed-to-protect-marine-life .While I don't advocate plagiarism,it might be useful to compare and contrast your design decisions with those of the people who have a successful product.