Direct answer: A tank assembly carries two seam families that are solved differently. Fillet seams where stiffeners meet the wall are brought into position by tilting the assembly, so the pool is supported and the bead forms evenly. Long thin-wall seams are usually welded with the part stationary and the torch travelling, where the controlling variable is heat input and the answer is segment order rather than equipment. Root access at stiffeners is settled at assembly, not at programming.
Who this is for: Fabrication engineers specifying robotic welding for thin-wall tank and enclosure assemblies with stiffeners or fin panels.
Scope: This EVST guide is written from two separate filmed stations: one where a positioner tilts a finned assembly to bring fillet seams into position, and one where a thin-wall tank panel stays still while the torch travels along long seams. They are two different stations and two different workpieces, shown together to compare approaches, not as two steps of one line.
Where transformer tank robotic welding splits into two problems
Thin-wall tank structures and enclosures carry long seams between wall panels and fillet seams where stiffeners or fins meet the wall. The two have different failure modes and should not be planned as one.
The enemy of a long seam is distortion. The difficulty of a fillet is whether the torch reaches the root. Treating them alike means one of them is always compromised.
In the reference footage both appear: one station tilts a finned assembly to bring fillets into position; another keeps a tank panel still and travels the torch along its long seams.
Fillets: tilt the assembly into position
In the first station the positioner tilts the whole assembly so that fillets which would otherwise face downward or sideways come into a weldable position.
How far to tilt is not a fixed number; it follows the geometry and the required leg size. Once the angle is set, robot access has to be re-checked, particularly between closely spaced fins.
Positioner capacity is sized on the complete assembly rather than a single panel, and the clamping scheme is verified at the least favourable angle, because the centre of gravity moves through the rotation.
According to ISO 5817:2023, quality levels are defined per imperfection type, so a fillet requirement is stated as a level together with a leg size rather than as an appearance description.
| Seam family | Controlling difficulty | Station approach |
|---|---|---|
| Long seams between thin wall panels | Heat input and distortion | Part stationary, torch travels, segmented sequence |
| Fillets where stiffeners meet the wall | Root access and pool support | Positioner tilts the assembly into position |
| Dense fin or stiffener spacing | Torch neck geometry against the gap | Checked at assembly, before programming |
| Mixed assembly with both families | Transfer between stations | Split the stations and plan the re-datum |
Long seams: keep the part still and travel the torch
On a thin-wall shell, turning the part can cost more than welding it. Those seams suit an arrangement where the workpiece stays put and the robot or gantry carries the torch along the seam.
The footage of that station looks simple; the engineering sits in how heat is distributed along the seam rather than in the motion itself.
Welded end to end in one pass, a thin panel will not stay flat. Segmenting, skip sequences and interpass control are what hold it, and the sequence also has to consider which side is welded first, since the first side pulls the panel and the second only partly compensates.
Root access at the stiffeners
Where a stiffener meets the wall, whether the torch reaches the root depends on stiffener spacing and on the geometry of the torch neck.
That has to be checked at assembly stage, not after the program is written. Where spacing cannot change, a slimmer neck or a modified torch angle can help, but both narrow the usable parameter window.
Lack of root fusion is the characteristic defect of this family and is not always visible from the outside, so it is judged by inspection rather than by appearance.
According to ISO 17637:2016, visual testing is carried out under defined conditions of access, illumination and viewing angle, which is why root-side access matters for inspection as well as for welding.
According to ISO 13920:1996, flatness and straightness on a welded construction are specified as tolerance classes, which is the form a distortion requirement for a thin-wall panel should take.
Choosing between the two approaches
Stiff assemblies with concentrated fillets favour tilting into position. Thin, long-seam dominated parts where turning is expensive favour keeping the part still.
Most real projects need both: the main long seams completed at a fixed station and the fillet-dense subassemblies at a tilting station, with transfer between them.
The decision follows part structure and batch size, not equipment brand. Classify the seams first, then choose the station type – and plan the transfer, because how the part is lifted, set down and re-datumed drives the loading time at the second station.
What the footage cannot establish
Plate thickness, assembly gap, actual heat input and resulting distortion are not visible in any frame; they come from the procedure specification and measurement.
Internal weld quality is the same: a well-formed cap does not establish that the root is fused. ISO 15614-1 covers procedure qualification for those claims.
The two segments used here come from two unrelated stations and two different workpieces, not two operations on one line.
The order EVST works in is the same each time: classify the seams, decide which family is turned into position, then plan the transfer between stations.
Frequently asked questions
Why not tilt the tank panels as well?
Because turning a large thin-wall panel costs handling time and risks distortion in itself. Where the seam is long and the part is flexible, keeping it still and travelling the torch is usually the cheaper and more stable answer.
What controls distortion on long thin-wall seams?
Segment order more than parameters. Skip sequences, alternating sides and interpass control spread the heat; a single continuous pass concentrates it and the panel will not stay flat.
How do I know whether the torch reaches the fillet root?
By checking stiffener spacing against the torch neck geometry at assembly stage. If it does not fit, the options are a slimmer neck, a changed torch angle or a design change – all of which are cheaper before programming.
Is a tilting positioner necessary for fillet seams?
Not always, but it is what allows the pool to be supported. Without it, fillets in awkward positions have to be welded at attitudes that narrow the parameter window and make consistency harder.
Project inputs for an application review
To plan a tank-structure welding station, an EVST application review opens from these inputs:
- tank drawings with wall thickness, stiffener spacing and distribution
- the seam list split into long seams and fillets, with required leg sizes
- flatness and distortion acceptance requirements
- existing positioner capacity, swing diameter and offset capability
Send the tank drawings and the plate thickness, and we will separate the seam families and propose a station split with the welding sequence that goes with it. Related reading: automatic welding system solutions, long sheet cutting automation acceptance.