Arc energy is what the arc delivers: volts times amps divided by travel speed. Heat input is what actually goes into the joint: arc energy multiplied by a thermal efficiency factor for the process. They are not the same number. In MMAW the difference is 20 per cent. If your PQR recorded one and your WPS quotes the other, you have a problem that nobody will notice until an auditor does.
Result
Q = k × (U × I) / (v[mm/s] × 1000)Indicative only. Thermal efficiency factors k per ISO/TR 17671-1:2002 Clause 19 Table 2, equivalent to EN 1011-1. Arc energy measurement per ISO 18491:2026, which supersedes the withdrawn ISO/TR 18491:2015. Confirm against the applicable WPS, PQR and code acceptance limits.
The distinction that catches people out
Arc energy: E = (U x I) / v
Heat input: Q = k x (U x I) / v
Where k is the thermal efficiency factor for the process. Same measurement, same joint, two different numbers.
Worked, so it is concrete. MMAW at 180 A, 24 V, 150 mm/min:
E = (24 x 180 x 60) / (150 x 1000) = 1.73 kJ/mm arc energy
Q = 0.8 x 1.73 = 1.38 kJ/mm heat input
A procedure qualified at 1.73 and a procedure qualified at 1.38 are the same weld. But if the PQR says 1.73 arc energy and the WPS quotes a heat input range built around 1.73, the WPS is now permitting 25 per cent more energy than was actually qualified. On a Q&T steel or a joint with impact requirements, that is not a paperwork problem.
ISO 15609-1:2019 Clause 4.4.17 asks for the range of heat input or arc energy, if specified. The if specified is doing a lot of work in that sentence. Specify it. Say which one you mean, on the WPS, in words.
That is competence, not paperwork.
The thermal efficiency factors
From ISO/TR 17671-1:2002, Clause 19, Table 2, equivalent to EN 1011-1:2009.
| ISO 4063 | Process | k |
|---|---|---|
| 121 | Submerged arc, wire electrode | 1.0 |
| 111 | Manual metal arc (MMAW / SMAW) | 0.8 |
| 131 | MIG | 0.8 |
| 135 | MAG | 0.8 |
| 114 | Flux-cored, no gas shield | 0.8 |
| 136 / 137 | Flux-cored, active / inert gas shield | 0.8 |
| 138 / 139 | Metal-cored, active / inert gas shield | 0.8 |
| 141 | TIG / GTAW | 0.6 |
| 15 | Plasma arc | 0.6 |
Where these do not come from. ISO/TR 18491 contains no efficiency factors and never did. Its scope is measurement of arc energy, and its Clause 4 says as much: determination of heat input involves multiplication by a thermal efficiency factor, arc energy does not. If you have seen these values attributed to 18491, including on an earlier version of this page, that attribution was wrong.
Watch out for the ranges in the literature. You will find calorimetry-derived figures quoted as ranges: SAW 0.95 to 1.0, GMAW 0.8 to 0.9, MMAW 0.7 to 0.8, GTAW 0.6 to 0.7. Those are experimental measurements, not the standardised values. For WPS work use the single fixed values above and do not mix the two sets on one document. Pick a system, cite it, stay in it.
AWS does it differently, and the numbers are not the disagreement
In AWS practice, heat input equals arc energy. D1.1 does not apply a thermal efficiency factor, and the reasoning is practical rather than technical: arc efficiency is a calorimeter measurement, and asking a fabricator to apply a factor they cannot verify invites arithmetic nobody checks.
So AWS and ISO do not disagree about the physics. They disagree about what goes on the paperwork.
Practical consequence if you work to both codes: the same joint, same parameters, will carry two legitimate different energy figures depending on which code the document sits under. Neither is wrong. Both need labelling.
Units
With U in volts and I in amps:
| Travel speed in | Arc energy E (kJ/mm) | Heat input Q (kJ/mm) |
|---|---|---|
| mm/s | U x I / (1000 x v) | k x U x I / (1000 x v) |
| mm/min | 0.06 x U x I / v | k x 0.06 x U x I / v |
Cross-unit, for anyone working to AWS documents in imperial: 1 kJ/mm = 25.4 kJ/in, and 1 kJ/in = 0.03937 kJ/mm. The 60 in the mm/min form is the one people drop. If your answer comes out sixty times too small, that is where it went.
Worked example
MAG, solid wire, ISO 4063 process 135. 280 A, 29 V, 350 mm/min.
E = 0.06 x 29 x 280 / 350 = 1.39 kJ/mm arc energy
Q = 0.8 x 1.39 = 1.11 kJ/mm heat input
Same weld run in GTAW at 180 A, 12 V, 80 mm/min:
E = 0.06 x 12 x 180 / 80 = 1.62 kJ/mm arc energy, Q = 0.6 x 1.62 = 0.97 kJ/mm heat input
Note what happened there. The GTAW run has higher arc energy than the MAG run, 1.62 against 1.39, but lower heat input, 0.97 against 1.11. Compare the wrong pair of numbers across two processes and you will draw the wrong conclusion about which one is putting more heat into the joint.
Measuring it properly, and why your meter is probably lying to you
This is the part of ISO 18491:2026 that most of the industry has not caught up with yet, and it matters more than the arithmetic above.
There are three methods, not one.
| Method | How | Formula |
|---|---|---|
| A | Welding current, arc voltage and welding speed | E = (U x I) / (1000 x v) |
| B | Total instantaneous energy and length of the run | E = E_TIE / (1000 x L) |
| C | Average instantaneous power and welding speed | E = P_AIP / (1000 x v) |
With v in mm/s, L in mm, E in kJ/mm. Method A is the one everyone knows. It is also the one that breaks.
For waveform-controlled welding, ISO 18491 says method B or C should be used, because method A can introduce errors of up to 70 per cent.
Seventy. Not seven.
Now the bit that catches people. ISO 18491 states that all pulsed welding processes, pulsed GMAW included, are waveform-controlled. And that power sources sold as synergic, programmable or microprocessor-controlled are generally capable of waveform-controlled welding. Walk through a modern fabrication shop and count the machines that are not synergic. That is the size of this problem.
AS/NZS 1554.1 says the same thing. Clause 1.1 Note 2 states that GMAW includes waveform controlled welding such as synergic, programmable and microprocessor controlled processes, giving pulsed spray transfer and controlled short circuit transfer as examples. Clause 5.3.4(e) then tells you to use average voltage and average current for pulsed mode, and adds that where arc energy needs limiting you should seek advice from the welding machine supplier. That hedge and the ISO 70 per cent figure are the same problem, described from two directions.
Your clamp meter is part of it. ISO 18491’s introduction is blunt: the instruments inspectors generally use, TRMS clamp meters among them, can no longer be relied on to measure welding energy correctly, and differences exceeding 30 per cent against true energy can sometimes be found. Meters reporting true energy, true power or power factor are built for this. Meters labelled kVA, DC power or average power generally are not.
For instantaneous measurement the standard sets a sampling floor: typically a multiple of the waveform frequency and at least 5 kHz, with a narrow exception down to ten times the waveform frequency where the error against high-frequency sampling has been shown to be no greater than 5 per cent.
Three practical points that cost nothing and get missed.
Record where you measured the voltage. ISO 18491 requires the position of the connectors to be recorded in the WPQR so production measurement matches qualification. It sets preferred locations by process group: the wire feeder connection for gas-shielded metal arc, the welding head for SAW, and the power source connection device for MMA, GTAW and plasma. Measure as close to the arc as practical, because cable size, length and connection quality all move the reading.
Measure all the cables. Where multiple earth or current lead cables are used, the measurement must include all of them. Reading one cable and multiplying by the number of cables is not valid, and neither is mixing earth and current leads in one measurement. Below a clamp meter’s range, loop the cable through the clamp and divide by the number of loops.
Length of the run ends at the middle of the crater, not the end of it. Calibrate to ISO 17662 and IEC 60974-14.
Where the number actually gets used
Procedure qualification range. ISO 15614-1 and AS/NZS 3992 set qualification ranges off the recorded energy. Get the basis wrong and the range you think you have qualified is not the range you have qualified.
Preheat determination. AS/NZS 1554.1 Clause 5.3.4 works from Q as defined by its Equation 5.3.4(2), which carries no thermal efficiency factor. Note that the axes of Figures 5.3.4(B) and (C) are labelled heat input while the clause text calls the same quantity arc energy. Do not apply k before reading those charts. See the preheat calculator.
Impact properties. High heat input coarsens the HAZ grain structure and drops toughness. This is why AS/NZS 1554.1 caps arc energy at Clause 4.6.1.1 for SAW and FCAW on certain grades, and at 2.5 kJ/mm for seismic service under Clause 4.6.1.3.
Q&T and TMCP steels. Too much energy destroys the properties the mill put in. There is a ceiling as well as a floor, and on those steels the ceiling is usually the binding constraint.
Distortion and cost. More energy per millimetre, more distortion, more rework. The economics sit alongside the metallurgy, and mostly get ignored until the fabrication is out of tolerance.
Standards and editions
- ISO/TR 17671-1:2002, Recommendations for welding of metallic materials, Part 1: General guidance for arc welding. Clause 19, Table 2 carries the k values. Current, last reviewed and confirmed 2016.
- EN 1011-1:2009, the same values, European equivalent.
- ISO 18491:2026, Welding and allied processes, Measurement of arc energies. First edition, June 2026. Prepared by ISO/TC 44/SC 10 with CEN/TC 121 under the Vienna Agreement. Cancels and replaces ISO/TR 18491:2015, technically revised, and elevated from a Technical Report to a full International Standard.
- ISO 15609-1:2019, Clause 4.4.17, on specifying the range on a WPS.
- AS/NZS 1554.1:2014 incorporating Amendments 1 and 2, for the Australian structural treatment.
ISO 18491 does not contain thermal efficiency factors. Its Note 2 to entry 3.3 states that where arc energy is used to quantify effect on a material it is multiplied by a thermal efficiency factor k, and refers the reader to the ISO/TR 17671 series. That is the correct chain: 18491 for how to measure, 17671-1 for the factor.
What is verified, what is derived, what is a modelling choice
Verified against the primary standards. Technoweld holds licensed copies of ISO 18491:2026 and AS/NZS 1554.1:2014 with Amendments 1 and 2, both read 16 August 2026. Confirmed directly from ISO 18491: the first-edition status and June 2026 date; that it cancels and replaces ISO/TR 18491:2015; that it contains no thermal efficiency factors and refers k to the ISO/TR 17671 series; the three determination methods at Clause 6; the up-to-70-per-cent method A error on waveform-controlled welding; the scope of what counts as waveform-controlled; the over-30-per-cent TRMS clamp meter discrepancy; the 5 kHz sampling floor; the voltage measurement locations and WPQR recording requirement; the multiple-cable rule; length of run to the middle of the crater; and calibration to ISO 17662 and IEC 60974-14.
Verified elsewhere: the k values, read from ISO/TR 17671-1 Clause 19 Table 2 and independently corroborated by AWS’s own Welding Digest quoting EN 1011-1.
Derived: all worked examples and the cross-process comparison.
Modelling choice: the calculator applies the single standardised k values, not calorimetry ranges. It implements method A, which is the appropriate method for non-waveform-controlled welding. Rounding is to two decimal places.
Not verified: whether an AS/NZS adoption of ISO/TR 17671 exists. It appears to exist only as an ISO Technical Report, and ISO 18491 lists it as ISO/TR 17671 (all parts) with no AS/NZS designation.
Limitations
Indicative only. This calculates a number, it does not qualify a procedure.
This tool implements method A. That is correct for non-waveform-controlled welding. If your power source is synergic, programmable, microprocessor-controlled, or running any pulsed mode, method A can be out by up to 70 per cent and you need method B or C off a meter that reports true energy or true power. The tool cannot tell what machine you are standing in front of. You can.
It also assumes steady-state welding at the parameters you enter. Real runs vary, and what a data logger records across a run is not what a single spot reading suggests. Confirm against the applicable WPS, PQR and code acceptance limits.
Related tools
- Preheat Calculator – AS/NZS 1554.1 method, and where arc energy feeds in
- Pipe Schedule and Wall Thickness
- Shielding Gas Waste Calculator
- Volts, Amps and the GMAW Arc
- GTAW Electrode Angle and Penetration
If you are developing procedures and want the qualification ranges right the first time, that is what I do.