An arc strike is an unintended localized fusion event outside the intended weld area — a stray arc from an electrode touching base metal, a slip of the electrode holder, or contact from an improperly clipped ground cable. They look minor: a small discoloration, maybe a tiny depression or crater. They are not minor.
Under AWS D1.1:2025 an arc strike anywhere outside the weld groove is a visual inspection rejection criterion. The repair sequence is straightforward, but it must be executed correctly and documented. This article walks through why arc strikes matter metallurgically, what the repair requires, and what records a compliant shop keeps.
Rule library based on AWS D1.1:2025; verify against your governing edition (the AHJ or contract may specify 2020 or earlier).
The metallurgical problem: local martensite and hydrogen traps
When an arc strike occurs, a small volume of steel melts and then self-quenches into the surrounding base metal. The surrounding plate is cold relative to the molten pool, so cooling rates through the 800–500°C range (the martensite transformation range) are extremely fast — much faster than the controlled cooling that occurs during normal welding with proper heat input and preheat.
The result is a hardened, brittle zone at the arc strike location. On A36 plate with low carbon equivalent, this zone is small and the effect is modest. On A572 Grade 50, A913, or A514 high-yield steels — or on thick plate where the mass acts as an even larger heat sink — the hardened zone is larger and more brittle. Under AWS D1.1:2025, these higher-strength base metals require careful attention to arc strikes precisely because the through-thickness stress in tension connections creates the conditions for crack propagation from a brittle initiation point.
In dynamically loaded structures, this matters even more. Fatigue crack initiation at surface discontinuities is a well-documented failure mode; a hardened arc strike in the web of a crane runway girder or in the flange of a seismic moment connection is a structural liability, not just a cosmetic defect.
Visual identification during inspection
Arc strikes are a required check item during visual inspection of structural fabrication. The CWI or QC technician should look for:
- Discoloration or heat tinting — a blue/straw/purple area outside the weld zone indicates localized heating. Not all heat tints are arc strikes (thermal cutting produces similar tints), but they warrant closer examination.
- Surface depressions or craters — small pits, often surrounded by a narrow discolored ring, are characteristic of arc strikes.
- Hardness to scratch test — experienced inspectors can sometimes detect the hardened zone with a scratch tool, but this is a screening technique, not an acceptance method.
- Areas near electrode holder rest points — stray contact is most common where welders rest their holders on the work or clip them to the joint during breaks.
In-process inspection is more effective than final inspection for catching arc strikes, because they are easier to spot on unpainted clean steel than on a blast-cleaned surface with a primer coat applied.
Repair procedure: grinding and blending
The standard repair for an arc strike is removal by grinding. The sequence:
1. Mark the location. The CWI marks every rejected arc strike before repair begins so they can be individually re-inspected after. Marking with paint pen or chalk on the adjacent area (not over the strike itself) is standard practice.
2. Grind flush and below. The arc strike material — the remelt crater and the immediately surrounding HAZ — must be removed. Grinding should remove the visible discoloration and go slightly below the original surface. The goal is to remove the hardened zone, not just smooth the surface. This typically means removing 1/32 to 1/16 inch of material beyond the visible arc strike boundary.
3. Blend smoothly. The ground area must blend smoothly into the surrounding plate surface. Abrupt transitions, notches, or tool marks left by grinding create new stress concentration points. Blend grinding with progressively finer discs to leave a smooth finish.
4. Check remaining thickness. On plate with limited section, confirm that grinding has not reduced the thickness below the minimum specified in the design documents or the applicable tolerance in the fabrication specification. If it has, weld fill is required.
When weld fill is required
If grinding removes material to the point of a measurable depression — typically defined as greater than 1/32 inch below the plate surface — or if the remaining section thickness falls below the design minimum, weld fill is required before re-inspection.
This is where the WPS enters the picture. The weld fill must be performed under a WPS that is qualified for:
- The base metal group and specification (e.g., ASTM A572 Grade 50, Group II)
- The welding process being used for the repair
- The applicable position
The fill weld is typically a small single-pass SMAW deposit with an E7018 electrode on structural A36 or A572 base metal. The WPS for this repair weld should already be in the WPS library as a standard repair procedure. Shops that don't have a dedicated repair WPS end up scrambling to justify one at audit time — it's worth qualifying one as part of initial procedure qualification.
After weld fill, the repair is re-inspected the same as the original nonconformance: visual acceptance first, then NDE if required.
NDE requirements after arc strike repair
For mild carbon steel (A36, A572 Grade 42/50) on standard structural applications, visual re-inspection after grinding is typically sufficient if no depression is present and the surface is smooth.
For higher-strength steels or where the contract documents specify magnetic particle testing (MT) for surface condition acceptance, MT of the repaired area is required after grinding. MT is the most practical method for structural fabrication: it detects surface and near-surface cracks with high sensitivity on ferromagnetic materials.
Penetrant testing (PT) can substitute when the material is non-ferromagnetic (unlikely for AWS D1.1 structural steel, which covers ferrous materials). On austenitic stainless, which would fall under AWS D1.6, PT is the surface method.
Key point: do not interpret a visually clean ground surface as NDE-clean. Grinding can smear metal over a crack, making it appear healed on the surface while the crack continues subsurface. MT after grinding is the only reliable confirmation on susceptible materials. For an overview of NDE method selection for structural welds, see the related article.
Documentation: the nonconformance record
Each arc strike found during inspection should generate a nonconformance record (NCR) in the fabricator's quality system. The NCR documents:
- Location (joint ID, member ID, distance from reference point).
- Description of the nonconformance (arc strike, approximate size, presence or absence of cracking).
- Repair method (grinding only, or grinding + weld fill).
- WPS reference if weld fill was performed.
- Re-inspection results (visual acceptance, MT results if applicable, CWI initials and date).
- Disposition — accepted as repaired, or re-submitted for further repair.
This record connects to the weld map and traceability system. For projects with third-party inspection or owner inspector witness requirements, the NCR is the document the witness reviews to confirm the nonconformance was properly closed.
Shops that catch and close arc strikes without paperwork create an audit exposure: if NDE later reveals a crack at that location, there's no record showing the repair was ever performed. The fabricator's QC plan should require NCR documentation for all arc strike nonconformances, regardless of how minor they appear.
Prevention: electrode discipline and ground placement
The most effective arc strike control is prevention. Common sources and their controls:
Electrode holder contact during rest breaks. Require welders to use a dedicated rest block or insulated holder holder, never setting the live electrode holder directly on the work. During electrode changes, the holder must be insulated or set to the side — not clipped to the joint or laid on the base metal.
Stray arcs from ground cable placement. The work connection (return lead) should attach to the structural member as close to the weld as practical, on the same piece being welded. Long ground runs on connected members cause stray current paths that can arc at bolted connections or plate edges.
Tack welding during fit-up. All tack welds must be made by qualified welders under the same WPS as the production weld, positioned within the joint groove or at the weld termination. Stray tack welds outside the groove are arc strikes by another name.
Arc-air gouging startup. The carbon electrode contacts the base metal before the arc establishes; a stray arc at startup is easy to land outside the intended gouge line. Require that arc-air starts be made in the groove or over the area being removed, never on adjacent plate.
The cost of undocumented arc strikes
A single undiscovered arc strike on a crane runway girder flange can propagate into a fatigue crack over 18–24 months of operation, leading to a weld repair or member replacement in service. The cost of that repair — including the crane downtime, engineering evaluation, and re-inspection — is measured in thousands to tens of thousands of dollars, depending on the girder.
The cost of a proper arc strike inspection and NCR program is a few minutes per shift. Software that tracks nonconformances against weld IDs makes it even faster, tying the repair record to the weld map automatically so there are no gaps at final inspection or at the owner's close-out audit.
Arc strikes are a common fabrication nonconformance. The distinguishing mark of a competent fab shop is not that they never happen — it's that every one gets caught, repaired correctly, and documented.