The FE result for a lug is early access and under verification. It is a local check of the shell and attachment junction under one lift step. It is not a replacement for the hand checks in the other articles, and the hand checks do not replace it either: they answer different questions.
What is in the model
- A band of cylindrical shell on its mid-surface, quad shell elements (the solver element is MITC4), at the shell's corroded thickness. Element size follows the shell decay length λ = √(R·t): first element at a junction min(t, 0.2·λ), refined zone 2.5·λ.
- The lug as plate or plates standing on the shell. A plain lug is one vertical plate of the lug thickness and the pin height. With end gussets the lug is meshed as an I-section standing its full height, which is stiffer than tapered gussets.
- An optional pad under the lug (width, height, thickness) and, for a trunnion, a pipe on the shell with no hole cut, with a round pad if chosen.
- Two lugs for the twin case; the second is a further attachment with the same plate, the same pad and its out-of-plane load mirrored.
- The lift as a primary (weight) load in global axes, with pressure and temperature off: the vessel is empty and unpressurised for the lift.
How it is loaded and held
The three force components of the governing lug load are applied at the lug's end, at the pin line, and spread over the plates by the beam analogy (axial force and bending as linear stress over the section, shear by the plates parallel to it). For a trunnion the end is a plane, circular pipe section loaded at its centre. The shell band is held at its bottom cut along the meridian, around the circumference and against meridional rotation, free to move normal to the shell; its top is free. The band length is set automatically: each cut end lies beyond the outermost attachment by the largest of 4·√(R_m·t), the attachment's half-size doubled, and one shell radius.
What is left out
- The pin hole, the pin, the shackle and any contact: the plate ends at the load line. Hole stress and bearing are the hand checks.
- The skirt base ring and the free base edge. The tail lug sits on a plain shell band; the tail pin elevation (z) enters the rigging statics only and is not used by the model. The base ring is checked by hand (Moss 10-5, Table 10-7, one point) and is not in the mesh.
- The tail beam and the bolts under the skirt: hand check only.
- The vessel's global bending in the lift (the shell bending formula covers it) and the pick-up reaction: the support at the band's bottom cut reacts the lug load, so the shell far from the lug is not the vessel in the lift.
- Weld geometry as such. The welds are checked by hand on the base lines; the FE model joins the plates to the shell by shared nodes.
[UNVERIFIED] Which of the FE assessment categories (membrane, membrane plus bending, range) the tool reports for a lug job, and the allowables it compares them with, were not read in this pass. Check the results page and the validation record before quoting any FE margin for a lug.
Reading the result with the hand checks
The hand checks answer for the plate: hole, bearing, shear, the base and the welds, for a load applied at the pin. The FE model answers for the shell: how the load spreads into the shell and pad, and where the local peak is. Report them side by side. Where they disagree, trust neither automatically: look for what the model leaves out (the base ring and edge, pin contact) and for the load sense that was worst in the hand check.
Default tail lug model, 2400 × 20 shell, lug 400 × 40, pin height 250, pad 700 × 600 × 20
| Decay length | R_m = 1210, t = 20: λ = √(24200) = 155.6 mm |
| Refined zone | 2.5·λ = 389 mm around the lug and pad |
| Band cut distance | the largest of 4·λ = 622 mm, twice the attachment half-length along the axis, and R_m = 1210 mm. Here R_m is the largest of the three unless the footprint half-length exceeds 605 mm, so each cut end lies at least 1210 mm beyond the outermost attachment |
| Pin elevation | z = 300 mm is not used; the code note says a lower weld would fall inside the 2·√(R_m·t) cut exclusion and vanish from the report if the band were flipped |
The model note on the geometry page states the automatic length actually used; read it there.