Start With the Cost Stack, Not the Machine
Automated welding in aerospace and nuclear work is not a machine purchase decision. It is a qualification, tooling, calibration, inspection, and rework decision. The machine has a known hourly rate; that is the easy number. The money that surprises fabrication managers sits in procedure qualification and heat-input mapping, operator certification, fixture repeatability, non-destructive testing, and the rework that follows a first-layer dilution miss.
If the business case is built only on arc-on time, it will be wrong.
The useful way to evaluate a cell is to treat every weld as a cost stack. At the top is the weld itself. Below it are the procedure qualification that proves the weld, the calibration that keeps the power source honest, the fixture that holds the joint, the preheat and interpass control that protects the metallurgy, the NDT that releases the part, and the rework loop that opens when any one of those items drifts.
Automation can reduce the top-line labour, but it only does so if the lower lines are funded and controlled.
Heat input is a production number
Heat input is often discussed as a metallurgical footnote. On aerospace and nuclear work, it is a line item. Use the standard calculation: heat input in kJ/mm equals voltage times amperage times 60, divided by travel speed in mm/min times 1000 times process efficiency.
Use 0.75 for GTAW, 0.8 for GMAW, and 0.9 for SAW. A 28 V, 450 A SAW pass at 450 mm/min gives (28 450 60) / (450 1000 0.9) = 1.87 kJ/mm. That sits inside a 1.22.0 kJ/mm window for a pressure-vessel steel such as SA-508. If travel speed falls to 350 mm/min, the same volts and amps produce about 2.4 kJ/mm, outside the window.
The arc log may still look stable. The procedure has already moved.
That is why closed-loop control is not a luxury on automated nuclear work. It is how the shop knows the heat input stayed in the qualified range when the positioner speed changed, the part heated up, or the operator adjusted a trim control. The cost appears in calibration and feedback maintenance, and it is cheaper than the rework loop.
A heat-input miss can trigger rework costing ten to twenty times the original weld value once cut-out, re-prep, re-purge, re-weld, repeat NDT, and documentation are counted.
The five cost buckets that decide the business case
For a typical automated aerospace or nuclear job, the budget divides into five buckets. The first is procedure qualification and heat-input mapping. The second is calibration and closed-loop feedback. The third is preheat and interpass control. The fourth is inspection and expected rework.
The fifth is deposition rate and the energy trade-off that comes with it. A 200-tube-joint job and a 12 m cladding job will have different unit rates, but the same five buckets. Build the quote around them and the number is defensible.
Procedure qualification is a fixed toll followed by a variable toll. The initial WPS and PQR prove mechanical properties, chemical composition, and soundness for a joint family. Every new thickness, consumable, joint design, or heat-input window adds work.
Calibration is a recurring line. Preheat and interpass control adds floor time and temperature records. Inspection and rework are the invoice that arrives late if they are not planned. Deposition rate is where automation earns its keep, but only within the qualified window.
Push wire feed speed or travel speed to gain output and the heat-input model changes with it.
Two Routes to the Same Acceptance Criteria
The useful comparison is not automated versus manual in the abstract. It is two production routes judged against the same acceptance criteria. Approach A is manual-assisted GTAW with manual cladding. Approach B is automated orbital GTAW with hot-wire cladding. Both can pass aerospace and nuclear NDT. They fail differently, cost differently, and make sense at different production volumes.
Where manual-assisted and automated orbital diverge
Manual-assisted GTAW puts a skilled welder on conventional positioners, using simple orbital equipment only where the joint family repeats. Its strength is flexibility. It handles mixed diameters, tight access, one-off repairs, and fit-up variation without a programming cycle.
Its weakness is that heat input, travel speed, and dilution follow the operator. Two welders can meet the same WPS and still produce different dilution profiles on a hot-wire overlay. That variability is not necessarily a defect, but it is a cost in NDT and release time.
Automated orbital GTAW moves the repeatability into the fixture and program. It is strongest on tube-to-tube and tube-to-tubesheet joints with stable end prep, stable fit-up, and a repeated diameter. The floor change is real: purge control becomes a fixture function, the weld head needs clearance and repeatable standoff, the positioner must hold angle, and the data log must prove the heat-input window.
The failure mode also changes. Automated orbital welds tend to fail on lack of fusion when the tooling drifts, the gap opens, or the arc standoff changes. Manual welds tend to fail on operator-dependent heat input and bead profile. The physics is the same; the exposure is different.
Cladding economics: dilution is a cost line
Hot-wire cladding makes the economics visible. A first-layer dilution target of 12% is not just a metallurgical target; it is a filler-metal and rework target. A batch that runs at 15% dilution may look acceptable to the arc eye and still fail the lab sheet.
The excess dilution wastes filler, can force a second layer, and increases the chance of chemical-composition or corrosion-property nonconformance. On a 12 m overlay job, that is not a small variance. It changes powder or wire consumption, deposition time, NDT coverage, and the probability of a repair loop.
The cladding parameters are distortion controls with a paper trail. Preheat and interpass bands, often in the 150250 C range for many pressure-vessel steels but always material-specific, must be recorded and held. Travel speed and wire feed speed must be tied to the qualified dilution window. If the fixture allows the torch standoff to move, the dilution moves with it. The shop then argues about the power source when the real cause is a positioner faceplate running out by 0.2 mm.
Defect exposure differs; the physics does not
Automation does not remove defects; it changes which defects are likely and how quickly they repeat. A drifting fixture can make the same bad weld on every part. A manual route can correct a one-off fit-up problem and then carry that correction into the next part as an uncontrolled variable. The comparison table below is the practical decision aid. It is not a process recommendation by itself; it shows where each route fits and what the floor must control.
| Route | Best fit | What changes on the floor | Dominant failure mode | Economic trigger |
|---|---|---|---|---|
| Manual-assisted GTAW with manual cladding | Mixed diameters, tight access, one-off repairs, unstable fit-up | Skilled welder, simple positioners, manual purge and parameter adjustments | Heat-input and dilution scatter between operators; inconsistent bead profile | Fewer than 30 identical joints, or fit-up varies more than 0.5 mm across the lot |
| Automated orbital GTAW | Repeated tube-to-tube and tube-to-tubesheet joints with stable end prep | Fixture-based purge, repeatable standoff, positioner angle control, heat-input data logs | Lack of fusion from drift, gap variation, or standoff change | More than 100200 joints per month, stable fit-up, and tooling capable of 0.2 mm runout |
| Hot-wire cladding | Corrosion or erosion overlay on pressure vessels and piping | Preheat and interpass control, dilution monitoring, filler delivery, layered NDT planning | Excessive first-layer dilution, such as 15% against a 12% target | 12 m or more of overlay, or a repair rate that justifies closed-loop dilution control |
| Automated GMAW | Long, repeated runs on thicker stainless or structural joints | Clamping, gas coverage, travel-speed control, spatter and post-weld cleaning | Lack of fusion and inconsistent penetration when fit-up or clamping relaxes | Long straight runs with repeatable joint geometry and enough volume to amortize tooling |
The Mechanical Foundation: Fixtures, Positioners, and Fit-Up
Automated welding is a following process. The torch does exactly what the fixture and positioner allow it to do. Put a good torch on a fixture that drifts 0.2 mm and the arc will make a consistent bad weld. That sentence explains most of the reject-rate arguments in aerospace and nuclear cells.
The power source log can show identical current and identical wire feed speed while the lab sheet shows first-layer dilution at 15% against a 12% target. The WPS never changed. The workpiece moved, or it was never running true to begin with.
Measure the tooling before changing the schedule
Before anyone rewrites parameters, put a dial indicator on the positioner. Sweep the faceplate or chuck jaw faces through a full rotation and record the runout. Check clamping force and repeatability. Jaws are consumables, not furniture. A clamp that relaxes after a few thermal cycles will change arc standoff, dilution, and clad thickness.
On a high-mix cell running thin nickel-base alloy 718 and 316L, a 4.6% reject rate was traced not to the welder or the program but to the fixture moving, clamps relaxing, and the positioner failing to hold angle. The target was the 0.2% defect rate that matched tooling can deliver in that kind of work.
Fixture material also changes the preheat story. A chill or heat sink can pull interpass temperature down; an insulated fixture can hold heat in the part. Neither is automatically wrong, but the WPS and the temperature records must match what the fixture actually does. The parameter window only exists if the mechanics hold. If the positioner cannot repeat angle and standoff, the closed-loop heat-input control is controlling a moving target.
Tolerances that belong in the router
Fit-up tolerance starts at the end prep, not at the torch. Write the tolerances into the router and inspect them before the part reaches the cell. For a 60 mm OD, 3 mm wall stainless tube, a cut-out and re-weld is not 40 minutes. It is cut-out, re-prep, re-purge, re-weld, re-radiograph, and a new traveller sheet: budget 3.55 hours when the joint fails dimensional check. That is the real cost of allowing a fit-up variable to enter the cell.
- Specify end-prep squareness, land thickness, and gap range for each joint family, and check them before clamping.
- Specify positioner runout and repeatability, for example 0.2 mm or better where the heat-input window is tight.
- Specify purge quality and pre-purge time, and record it as part of the weld data package.
- Specify clamp force and clamp maintenance intervals; treat worn jaws as a controlled consumable.
- Specify fit-up hold points before the arc starts, not after the first NDT rejection.
Audit records begin at the fixture
A quality auditor will ask what held the part, how it was checked, and what changed between the qualified procedure and the production run. The fixture drawing, positioner calibration, jaw replacement log, and fit-up inspection records are part of the answer. When a flange face moves 0.1 mm and the customer rejects the weld, the argument is rarely about welding skill. It is about whether the job file captured the fixture, the fit-up, and the thermal sequence that produced the movement.
Qualification, Calibration, and NDT: The Recurring Invoice
Most shops price automated aerospace and nuclear welding as if the machine is the cost center. It is not. The machine has a known hourly rate. The money that surprises fabricators sits in procedure qualification, operator certification, non-destructive testing, and the rework that follows a first-layer dilution miss. Build the quote around those four items and the number can be defended.
Procedure qualification and heat-input mapping
Both orbital fusion welding and hot-wire cladding need a qualified WPS and a PQR that proves mechanical properties, chemical composition, and soundness. The PQR is not a formality. It defines the heat-input window, the preheat and interpass band, the dilution target, and the travel-speed range that the cell must hold.
Map the heat input at the top, middle, and bottom of the qualified travel-speed range. If the procedure allows 1.22.0 kJ/mm, the cell needs to know what happens at 1.25 and 1.95, not just at the nominal midpoint.
Calibration is the recurring line that keeps the map honest. Voltage, current, wire feed speed, travel speed, and temperature instruments need a calibration schedule and traceable records. Closed-loop feedback is only as good as its sensors. A non-calibrated travel-speed signal can move the calculated heat input outside the window while the display still shows green.
Operator certification does not disappear
Certified operators are still required even when the weld head moves automatically. The operator sets up the joint, verifies fit-up and purge, loads the program, monitors the arc, and responds to faults. In nuclear work, the operators certification and the procedure qualification are separate lines. The operator certification is a recurring cost, not a one-time purchase. Plan for requalification, refresher training, and audit time.
NDT and rework must be priced before the first arc
NDT is the invoice that does not appear until late. Radiography, ultrasonic testing, penetrant testing, and destructive testing all need slots, qualified personnel, and documentation. A cut-out and re-weld consumes a fixture, a certified welder, an NDT slot, and a quality engineers afternoon. None of that appears on the original quote unless it is planned as an expected cost.
Use a simple cost-per-metre check. At a $95/h shop rate and a travel speed of 0.35 m/min, one pass costs about $4.50 per metre in arc labour alone: $95 divided by 60, divided by 0.35. Add setup, purge, qualification amortization, NDT, and expected rework, and the true cost can land at $120$220 per metre for critical aerospace and nuclear work.
That range is not a quote; it is a warning that arc labour is a small fraction of the released-part cost. If the automation proposal only reduces the $4.50 line, it has not addressed the money.
Where Automation Pays and Where It Stops Paying
Automation pays when the joint family repeats, the fit-up is stable, the fixture can hold runout, the heat-input window is mapped, and the NDT plan is funded. A 200-tube-joint-per-month job with stable end prep can justify an orbital cell. A 12 m cladding job with a corrosion-resistant overlay can justify hot-wire cladding and closed-loop dilution control.
A high-mix cell with tight tolerances can justify automated GTAW if the tooling is matched to the target defect rate. The common factor is not the machine; it is the repeatability of the input.
Volume, fit-up stability, and the true cost per metre
Run the rework math before the purchase order. At 200 tube joints per month and a 4.6% reject rate, about 9.2 joints enter rework. At 4 hours each, that is roughly 36.8 hours per month, or about $3,496 in labour at $95/h before repeat NDT and schedule disruption.
If matched tooling and controlled fit-up move the reject rate to 0.2%, the shop recovers roughly 34 hours per month and frees NDT and fixture capacity. That is the business case. It is not the machines travel speed; it is the reduction in the rework loop.
The same logic applies to cladding. If first-layer dilution runs at 15% against a 12% target, the shop pays twice: once for excess filler and once for the higher probability of repair. If the cell can hold the dilution window and document it, the automation is earning money. If it cannot, the cell is just making a consistent out-of-window weld faster.
Do not automate these jobs
Automation is not worth it when the part mix changes every shift, access is restricted, fit-up varies more than the fixture can correct, or the shop cannot fund the qualification and calibration burden. If lot sizes are under 30 identical joints and fit-up varies by more than 0.5 mm across the lot, manual-assisted GTAW is usually the lower-risk route.
If the joint requires frequent one-off repairs, a skilled welder with simple positioners will beat a cell that must be reprogrammed and requalified. If the shop cannot hold 0.2 mm runout on the positioner and cannot maintain a calibrated heat-input log, buying closed-loop control is buying a false sense of control.
Also do not automate a job that has no stable NDT plan. Automation without inspection discipline simply creates a faster route to a late rejection. Nuclear and aerospace acceptance criteria are unforgiving; the cell must be specified around the records, not around the arc.
What to Put in the Router Before You Release the Cell
The router is where the automation decision becomes real. It should name the joint family, the qualified heat-input window, the preheat and interpass band, the dilution target for cladding, the fit-up tolerances, the positioner runout limit, the purge requirements, the calibration interval, the operator certification requirement, and the NDT hold points.
It should also name the rework trigger: what happens when first-layer dilution exceeds the target, when a bend test fails on lack of fusion, or when a flange face moves 0.1 mm.
Specify the fixture as a controlled item. Require a positioner runout check at installation and on a defined interval. Require jaw replacement intervals and clamp-force checks. Require heat-input mapping at the top, middle, and bottom of the qualified travel-speed range. Require the data
Automated welding in other sectors
If you work across more than one sector, these guides use the same engineering logic with the numbers that differ by industry.
Elevating Welding Positioner (EWP Series)
Technical Specifications:
| Model | Rated Load (tons) | Rotation Power (kW) | Tilt Power (kW) | Table Dia (mm) | Max Rotation (mm) | Tilt Angle | Weight (tons) |
|---|---|---|---|---|---|---|---|
| EWP1 | 1 | 1.1 | 1.5 | 900 | 1500 | 0-135 | 1.2 |
| EWP2 | 2 | 1.1 | 1.5 | 1200 | 2000 | 0-135 | 1.6 |
| EWP4 | 4 | 1.5 | 2.2 | 1400 | 2100 | 0-120 | 2.8 |
| EWP6 | 6 | 2.2 | 3.0 | 1500 | 2500 | 0-120 | 5.0 |
| EWP8 | 8 | 2.2 | 4.0 | 1600 | 2800 | 0-120 | 6.5 |
| EWP10 | 10 | 3.0 | 5.5 | 1800 | 3000 | 0-120 | 7.8 |
| EWP15 | 15 | 3.0 | 5.5 | 2000 | 3200 | 0-120 | 12 |
| EWP20 | 20 | 4.0 | 7.5 | 2200 | 3400 | 0-120 | 18 |
Note: Variable rotation speed. Height adjustable column. Remote pendant control. CE certified options available.
For detailed specifications and pricing, contact us at [email protected] or visit https://biqee.cn.
Conclusion
Choosing the right welding automation solution is key to improving productivity and quality in your fabrication operations.
