Direct answer: Mobile platform work-position localization is the real specification question. A mobile welding platform removes the need to move the workpiece; it does not remove the need to know where the workpiece is. Every parking position is a new setup in which the robot-to-part relationship has to be re-established before welding, and floor flatness, aisle width, cable routing, and power provision decide how repeatable that is. Verify the localization method and the site conditions before selecting the platform.
Who this is for: This guide is written for fabrication and project engineers evaluating a mobile welding platform for large or immovable workpieces, and for the people who will have to make it work on an existing shop floor.
Scope: It covers what changes when the equipment moves to the part, what every parking position has to re-establish, and which site conditions decide whether that is repeatable. It does not cover welding procedure development, seam tracking selection, or the structural design of the workpiece.
What the platform changes, and what it does not
The premise of a mobile welding platform is simple and often correct: for a large or awkward workpiece, moving a few hundred kilograms of equipment is easier than moving several tonnes of fabrication into a fixed cell. In the reference footage for this guide the platform drives itself across the shop floor, carrying the manipulator and its onboard equipment, and repositions beside a workpiece support.
What that changes is material flow. What it does not change is the fundamental requirement of any robotic welding: the controller has to know where the joint is relative to the tool. In a fixed cell, a fixture answers that question once and keeps answering it. On a mobile platform, the question is asked again at every stop.
EVST reviews mobile welding proposals on that basis. Mobile platform work-position localization is the whole design problem in one sentence, and it is why the interesting specification questions are about localization and site conditions rather than about the arm. It is also why comparing this approach with flexible welding for large structures and datum recovery is more informative than comparing arm payloads.
Mobile platform work-position localization at every stop
When the platform parks, the relationship between the robot base and the workpiece is whatever the parking accuracy delivered. Unless that relationship is measured, every taught position is wrong by the parking error. So the platform has to re-establish the frame at each station, using touch sensing, laser or vision measurement of known features, a physical docking arrangement, or a combination.
Each method has a different cost in cycle time and a different sensitivity. Touch sensing on a machined or otherwise well-defined feature is robust but slow, and it needs a feature that actually exists on the fabrication. Optical measurement is faster but depends on surface condition and lighting. Mechanical docking is fast and repeatable but requires prepared datums at every working position, which turns part of the shop floor into fixed infrastructure.
EVST records the re-localization result as an acceptance item. Whichever mobile platform work-position localization method is chosen, the acceptance question is the same: after re-localization, how close is the taught path to the real joint, and what does the process do with the remainder? That remainder is what seam tracking or a tolerant parameter window has to absorb, and it should be stated rather than assumed.
The floor is a specification, not a detail
A mobile platform inherits the shop floor as part of its mechanical system. Flatness and level affect how the platform sits and therefore how the arm’s base is oriented at each stop. Surface condition affects traction and wear. Slopes, drainage channels, expansion joints, and cable ducts affect where the platform can go at all.
Aisle width matters twice: once for the platform to travel and once for it to turn and park with the arm in a usable orientation. A route that a platform can drive through is not necessarily a route in which it can work, and the working footprint including the extended arm is the number that belongs on the layout drawing.
EVST asks for a route survey before recommending a platform. None of this is exotic, but it is site-specific, and it is the most common reason a technically capable platform underperforms. A survey of the actual floor along the actual routes is worth more than any datasheet comparison.
Decision table: site and part evidence mapped to platform configuration
The second row is included deliberately. A mobile platform is not automatically the right answer for a large workpiece, and the comparison against a fixed arrangement should be made explicitly rather than skipped.
| Evidence from the site and the workpiece | Configuration it justifies | Evidence you still owe |
|---|---|---|
| Workpiece cannot be moved, joints spread over its length | Mobile platform with multiple parking positions | Re-localization method and its result at each stop, segment tie-in quality, total setup time per stop |
| Workpiece can be moved and the family is stable | Fixed cell with a positioner or rail instead | Handling method and time, fixture design, whether the family really is stable |
| Good flat floor, clear wide aisles | Self-propelled platform with free routing | Route survey, obstruction clearances, working footprint with the arm extended |
| Uneven floor, ducts or slopes on the route | Prepared working positions or docking datums | Cost of the prepared positions, what happens if the workpiece is placed off-datum |
| Long trailing cable and gas package required | Cable management and a defined drag path | Cable behaviour at full travel, snag points, whether the path stays clear during production |
Cables, gas, power, and the drag path
In the reference footage a cable trails behind the platform across the floor. That is normal for a tethered mobile welding system, and it is also a real constraint. The tether sets the maximum distance from the supply point, it defines a drag path that has to stay clear of traffic and of the workpiece, and it adds a snag risk at every turn.
The supply itself needs planning. Welding power, gas, and extraction have to be available at every working position, or the platform’s mobility is limited by the length of the umbilical rather than by its drive system. Distributing supply points around the working area is often cheaper than a longer tether, and it is easier to maintain.
Extraction is the item most often left until commissioning. On a large workpiece the arc moves across a wide area and fixed extraction rarely follows it, so the extraction concept belongs in the same layout review as the drive routes.
Segmenting the seam by parking position
A joint longer than the arm’s working envelope at one stop has to be welded in segments, with the platform repositioning between them. That introduces tie-ins, and tie-ins are a quality question rather than a logistics one. Where the segment boundaries fall, how the arc is restarted, and how the overlap is inspected all belong in the procedure.
The number of stops also drives the honest cycle time. Setup time per stop — drive, park, re-localize, verify — multiplied by the number of stops is frequently comparable to the arc time itself on this kind of work. Quoting arc time alone understates the real duration, and it is worth asking for both.
Acceptance language should cover the segmented result, not just a sample bead. According to ISO 5817:2023, quality levels for imperfections are defined for fusion-welded joints in steel, nickel, titanium and their alloys. According to ISO 13920:1996, general tolerances for welded constructions cover lengths, angles, shape and position. Together they cover both the joint and the assembly it produces.
Safety when the machine itself moves
A moving welding platform combines two hazard classes: the robot application and a moving machine sharing space with people. According to ISO 12100:2010, the general principles for design cover risk assessment and risk reduction, and ISO 10218-2:2025 applies to the industrial robot application and cell, including one whose location changes.
Where the platform is a driverless industrial truck, a second framework applies to the travel phase. According to ISO 3691-4:2023, safety requirements and verification apply to driverless industrial trucks and their systems. A manually positioned platform is a different case, and the assessment should say which one applies rather than borrowing language from the other.
The practical output is a defined behaviour for each phase: what the platform may do while travelling, what it may do while parked and welding, and what stops it in each case. Arc screening for people who are not part of the operation is part of the same picture, because on an open shop floor the arc is not inside a cell.
What to send for a mobile welding review
Send the workpiece drawing with dimensions, mass, and where it sits during welding; the seam map with access directions; a shop layout showing aisle widths, floor condition, and obstructions; your power, gas, and extraction provision at the working positions; and how the workpiece is placed today. That set is enough to say how many parking positions are needed, which localization method fits, and whether the site supports the approach at all.
If the honest answer is that the workpiece could be moved and the family is stable, expect the review to say so. For that comparison, large workpiece pose and reach testing is the more relevant reference than a mobility datasheet.
Frequently asked questions
Does a mobile platform remove the need for fixtures?
It removes the need to bring the part to a fixture. It does not remove the need to know where the part is: each parking position has to re-establish the robot-to-part relationship by sensing, measurement, or a prepared datum.
How accurate is the parking position?
Parking accuracy is a property of the specific platform and floor, and it should be measured on your site rather than taken from a brochure. What matters more is what happens after parking: the re-localization result is the number the process actually sees.
What site condition causes the most trouble?
Floor condition along the actual routes, followed by aisle width at the working footprint with the arm extended. Both are cheap to survey and expensive to discover after delivery.
How should a segmented seam be accepted?
Treat the tie-ins as part of the procedure and inspect them explicitly. ISO 5817:2023 covers quality levels for imperfections and ISO 13920:1996 covers general tolerances for the welded construction that results.
Project inputs for an application review
Send the following and the mobile welding question can be reviewed against your shop rather than a brochure:
- the workpiece drawing with overall dimensions, mass, and where it sits during welding
- the seam map with joint positions and the access direction for each
- a shop layout with aisle widths, floor condition, and obstructions
- your power, gas, and extraction provision at the working positions
- how the workpiece is placed today and whether that position can be repeated
Send the workpiece drawing with mass and its welding position, the seam map with access directions, a shop layout with aisle widths and floor condition, your power, gas and extraction provision, and how the workpiece is placed today. That set is enough to review parking positions, localization method, and whether a mobile platform or a fixed arrangement suits the work. Related reading: automatic welding system solutions, steel structure welding applications.