← Back to Structural Safety Report

Purpose

This page maps RP2’s existing structural safety factors (computed by the analytical beam-theory suite in src/structural/) against the BS EN ISO 12215 small-craft standard series, and reports where new ISO-anchored checks agree, disagree, or cannot yet be computed for this vessel’s asymmetric proa layout.

Full clause/table citations are recorded in constant/standards/iso12215.json in the repository. Per copyright terms, the ISO/BSI source PDFs themselves are held locally by the design team and are not redistributed through this repository or site — only bibliographic citations (standard, edition, clause/table number) are recorded.

Numbers on this page were computed by running make validate-structure BOAT=rp2 CONFIGURATION=beaching, which uses RP2’s real modeled mass/geometry — they are reproducible from the repository, not illustrative placeholders.

The multihull gap is largely resolved

RP2’s design team previously flagged that ISO 12215-5 (design pressures) is written for monohulls, while RP2 is an asymmetric proa (one main hull + one outrigger float) — a mismatch with no obvious governing standard on hand.

ISO 12215-7:2020 (“Determination of loads for multihulls and of their local scantlings using ISO 12215-5”) is, in fact, on file with the design team. It gives global load cases (crossbeam torsion, longitudinal shock, rig-load combination) and float-specific local pressures that ISO 12215-5 alone does not cover.

It is still not a perfect fit: ISO 12215-7’s worked formulas assume a symmetric catamaran or trimaran (two equal hulls, or one hull + two equal floats). RP2 has only one, asymmetric outrigger. Where a Part 7 formula is applied below, the adaptation used and its confidence level is stated explicitly — nothing is silently forced onto the proa geometry.

Cross-check summary

# Load case Existing SF (vs. raw yield, required ≥2.0) ISO-anchored cross-check Status
1 Suspended ama (aka bending) 3.27 GLC5 (ISO 12215-7 §12.10) dynamic longitudinal-shock case on the same aka: SF = 1.35 vs. ISO’s own design stress ⚠️ See below — different load case, much tighter margin
2 Aka point load (crew standing) 5.74 Not separately re-derived (static, non-ISO-specific load case) —
3 One end supported (spine bending) 16.50 Not separately re-derived —
4 Mast wind loading (25 kn) 6.18 3.09 (corrected) ISO 12215-10:2020 Table 5 formula itself ✅ Bug fixed — see below
5 Diagonal braces (lateral) 33.60 Reviewed against ISO 12215-7 Table 12 aluminium design stress convention ✅ Confirmed genuine margin, not a modeling artifact — see below
6–8 Wave slam (vertical/frontal/sideways) 4.46 / 5.23 / 4.46 ISO 12215-7 §9.4 float bottom pressure: 0.44× the ad hoc model’s dynamic pressure ℹ️ Ad hoc model is more conservative — see below
9 Lifting sling 6.16 Not an ISO-covered load case (crane/rigging, not a small-craft design load) —
10 Gunwale loads 12.24 Not separately re-derived —
11 Ama lift wind speed 22 kn Informational, unchanged —
— Aka-to-vaka joint (not previously checked at all) ISO 12215-6 good practice + first-pass weld sizing: passes a web-area rule of thumb ⚠️ Real gap confirmed — see below
— Aluminium alloy spec (not finalized) ISO 12215-3 §4.7: extruded-profile family (6000-series) is the correct choice ⚠️ Documentation gap confirmed — see below

1. Mast wind loading: a real formula bug, now fixed (6.18 → 3.09)

mast_analysis.py computed wind force as F = 0.72 × V² × A / 2, citing ISO 12215-10 Clause 7.2, Table 5. Checking the ISO 12215-10:2020 source directly (Table 5, Note 3):

“the wind force is F_A = N × 0.5 × ρ × A × V² (N) where N=1.2 … ρ=1.2 kg/m³ … hence F_A = 0.720 × A × V² (N)”

The 0.72 coefficient already includes the ½ dynamic-pressure term (1.2 × 0.5 × 1.2 = 0.72). The code’s extra /2 halved the applied wind force, which approximately doubled the reported safety factor. This has been corrected — mast_analysis.py now computes F = 0.72 × V² × A with no extra division.

Result on the real design: mast wind-loading safety factor moves from 6.18 → 3.09. The mast still comfortably passes the required minimum of 2.0 — this is a correction to the reported number, not a design deficiency.

2. Diagonal braces (33.60): confirmed genuine margin, not a boundary-condition artifact

The design team’s original concern was that the diagonal braces’ very high safety factor (33.60) might be a “fixed-support artifact” — i.e. that brace_analysis.py’s Euler buckling calculation assumed unrealistically rigid end conditions.

Checking the actual computation: the code uses K = 1.0 (pinned-pinned), which is the most conservative standard idealization for buckling (it gives the lowest buckling capacity among common end-fixity assumptions — any real end fixity, K < 1.0, would only increase the buckling load and therefore increase the reported safety factor further). So K=1.0 is not an overly-rigid assumption; if anything it already understates the brace’s true capacity.

The real driver of the 33.60 figure: the brace’s actual compressive load under the “boat on its side” load case (698.5 N) is simply small relative to its Euler buckling capacity (23,468 N) for this slenderness (88.3). Checked against ISO 12215-7 Table 12’s aluminium buckling design-stress convention as well, nothing in the governing standard changes this conclusion. The 33.60 figure is genuine margin, not a modeling artifact.

3. Aka bending: existing 3.27 covers a different (lighter) load case than ISO’s own dynamic global load

aka_analysis.py’s 3.27 safety factor is for the suspended ama case: static weight of the outrigger (1g), no dynamic amplification, compared against raw aluminium yield strength (240 MPa) with a required SF of 2.0.

ISO 12215-7 §12.10 (GLC5 — longitudinal shock, e.g. hitting a floating object or a steep wave) is a genuinely different, dynamic load case that the existing suite does not check at all. Adapting it to RP2 (treating the single ama as the closest available “trimaran float” analogue, the most defensible fit in Part 7 for this geometry):

An SF of 1.35 against ISO’s own design stress (which already includes ISO’s built-in knockdown from yield) means this case is ISO-compliant but with far less margin than the 3.27 figure suggests for a different loading direction. This is the most actionable finding on this page: the aka has comfortable margin against the static suspended-weight case, but only modest margin against a dynamic longitudinal-shock case that nothing in the existing suite previously checked.

(GLC1 — the diagonal torsional-moment case in a quartering sea — was deliberately not carried through to a per-aka number. ISO 12215-7 Annex D’s distribution method assumes the torsional moment is reacted by crossbeams twisting between two hulls of the same type; RP2 has only one outrigger, so there is no second hull for the akas to react against in the way Annex D models. Per ISO 12215-7 §12.3.3, a non-typical arrangement like this needs the “enhanced method” (FEM), not the simplified method — this is flagged as an open item requiring dedicated engineering analysis, not force-fit into a formula that doesn’t apply.)

4. Aka-to-vaka joint: confirmed real gap, but for a different reason than assumed

The original assumption was that ISO 12215-6 §6.3/7.2 would give bonding-width / k_j-style coefficients for this joint, similar to FRP tabbing calculations.

Checking the source directly: ISO 12215-6 Clause 7.2 “Bonding” is specific to FRP/composite construction. It does not apply to the aluminium aka-to-vaka connection at all. For metal construction, ISO 12215-6 gives only qualitative good-practice principles:

There is no closed-form ISO 12215-6 formula for a metal end-connection’s load capacity. This module derives a first-pass minimum by engineering analogy to Clause 6.3.4 (which addresses a different connection type — floating frames — but states “weld or glue area shall generally not be less than the stiffener web area”): the aka’s web area (833 mm²) exceeds the area required to carry the suspended-ama case’s moment and shear at ISO’s welded design stress (governing requirement: 391.5 mm²) — so this rule-of-thumb passes, but it is not a certified joint design, because no joint design exists yet. This remains the right place to focus engineering effort once a real bracket/weld geometry is proposed.

5. Wave slam pressure: ad hoc model is more conservative than the ISO estimate

wave_slam.py estimates ama slam pressure from an assumed 3 m/s impact velocity, a hand-picked slam coefficient (Cp=1.5), and a 2.5× dynamic factor — none of the three numbers are ISO-derived.

Using ISO 12215-7 §9.4 (“Design pressure for trimaran floats” — again treating the single ama as the closest available float analogue) and ISO 12215-5 §8/§9 for the base formula and k_DC/k_L/k_AR factors:

This is a reasonable outcome (the ad hoc model was deliberately chosen to be conservative) and gives the design team a real ISO-anchored floor to compare future tuning of the ad hoc model against, rather than an unmoored guess.

Caveats on the ISO-anchored number: it uses this codebase’s modeled structural mass rather than full ISO loaded displacement (crew/payload/fuel excluded, so it likely understates the true design pressure); a placeholder Design Category B (unconfirmed); k_AR without a documented stiffener-span reduction for the ama shell; and an assumed mid-length slam location. See iso_design_pressure.py for the full list.

6. Aluminium alloy: family choice confirmed correct, documentation still open

ISO 12215-3 §4.7 classifies aluminium into non-heat-treatable (Mg-based, typically rolled sheet/plate) and heat-treatable (Si-based, typically extruded sections/tubes/closed profiles) families. The aka, braces, mast and gunwale are all extruded closed-profile sections — the heat-treatable (6000-series) family is the correct choice per §4.7.3, consistent with the 6061-T6 placeholder already used in beam_mechanics.py.

What is still genuinely open (per the design team, no alloy/temper has been finalized yet): ISO 12215-3 §7.3 requires a material data sheet stating chemical composition, grade, temper, minimum mechanical properties, and — critically — minimum/expected mechanical properties after welding. This does not exist yet. It is a documentation gap to close before certification, not evidence the current design is undersized: ISO’s own tabulated values for a representative alloy (EN AW-6061 T5/T6 closed profile) are already used throughout this page’s cross-checks as a stand-in.

One consequence worth flagging: whether the aka-to-vaka joint ends up welded or mechanically fastened changes the applicable ISO design stress by ~1.8× (81 MPa welded vs. 147 MPa unwelded/riveted, per ISO 12215-5 Table 17). This is undetermined today because no joint design exists yet (see §4 above) — it is one more reason that joint design is the highest-value open item.

Summary of open items

  1. Aka global dynamic loading (GLC5) — SF 1.35 against ISO’s own design stress. Worth a closer look; the static suspended-ama check (3.27) does not cover this load direction.
  2. Aka-to-vaka joint — no design exists; a first-pass ISO-anchored rule of thumb passes, but this needs a real bracket/weld design once the joint geometry is decided.
  3. Aluminium alloy/temper — needs to be finalized and documented per ISO 12215-3 §7.3 (including post-weld properties), since it affects both the material design stress and the joint design above.
  4. GLC1 (quartering-sea torsion) — order-of-magnitude bound computed (7.6 kNm), but not distributed to individual akas; ISO 12215-7 itself says this asymmetric arrangement needs FEM (“enhanced method”), not the simplified formula.
  5. Design Category — not yet formally declared; Category B assumed as a placeholder throughout this page’s ISO-anchored calculations and must be confirmed per ISO 12217 before certification.