A weld failure is either a fabrication defect the inspection missed or an in-service mechanism nobody anticipated — and the joint itself usually shows which one it was.
Start a conversation with our AI Research Concierge, already scoped to weld failures. Pick a starting point, or describe your situation directly.
A welded joint is metallurgically the most complicated few centimetres in most structures — a fusion zone that solidified from a molten pool, a heat-affected zone whose microstructure was rewritten by a rapid thermal cycle, and a base metal that may carry locked-in residual stress from restraint during welding. That complexity is precisely where the failure evidence lives. A fabrication defect — lack of fusion, porosity, a hydrogen crack in the heat-affected zone — has a geometry and a location dictated by the welding process itself, and it is distinguishable from an in-service crack that initiated at a weld toe or defect after years of loading. Establishing which one occurred, and whether the procedure, the welder, the consumables, or the design was responsible, is what this specialization does.
Some weld failures originate at fabrication; others originate in a sound weld once it is placed into service. The distinction is usually visible in the defect geometry.
Inadequate heat input, joint fit-up, or travel speed leaving a planar, unfused discontinuity that behaves like a built-in crack.
Trapped shielding gas, contamination, or moisture forming voids in the weld metal that reduce the effective load-bearing cross-section.
Diffusible hydrogen from moist electrodes or flux combining with a hard heat-affected-zone microstructure and residual stress, cracking hours or days after welding is complete.
Low-melting-point segregates concentrating along the weld centerline as the pool solidifies, cracking under the contraction strain of a highly restrained joint.
Through-thickness, step-like cracking in the base metal beneath a restrained weld, driven by banded non-metallic inclusions and low through-thickness ductility.
Undercut, HAZ hardening or softening, and locked-in residual stress from welding sequence and restraint, setting up conditions for later cracking or distortion.
Weld investigations combine non-destructive mapping of the defect with destructive sectioning to explain why it formed.
A weld defect rarely stays confined to the joint itself:
Grinding out or re-welding a joint before it is documented destroys the defect geometry and heat-affected-zone microstructure the investigation depends on. Preserve the joint as-found, with the welding procedure and consumable records.
By the defect’s geometry and location. Fabrication defects — lack of fusion, porosity, slag inclusions — have shapes and positions dictated by the welding process itself: planar and along the fusion line, or rounded and distributed through the weld metal. In-service cracks typically initiate at a stress concentration such as a weld toe or an existing defect and show progressive growth features, such as beach marks, corrosion products, or branching, that a fabrication defect does not have.
Hydrogen introduced from moisture in the electrode coating, flux, or base-metal surface diffuses into the hot, hardened heat-affected zone during and after welding. Because diffusion is slow at room temperature, the hydrogen keeps concentrating at points of high residual stress until it reaches a critical level and a crack initiates — which is why hydrogen cracking is a delayed phenomenon rather than an immediate one, and why codes require a waiting period before final inspection on susceptible materials.
Frequently, yes. Defects consistent with a sound, properly qualified procedure but present in only one welder’s work — inconsistent penetration, poor bead placement, contamination from technique — point toward workmanship. Defects present across multiple welders using the same procedure, consumables, or preheat practice point toward the procedure or the materials themselves. Comparing the defective weld against other welds from the same job, procedure, and welder is usually how this is resolved.
Lamellar tearing occurs in the base metal, not the weld metal — it is a step-like, through-thickness separation that follows bands of non-metallic inclusions beneath a heavily restrained weld, driven by the through-thickness contraction strain the joint imposes on the plate. It is identifiable by its characteristic terraced fracture appearance and its location in the base metal rather than the fusion zone or HAZ, and it is controlled through joint design and restraint management rather than through welding parameters alone.
It depends on the application. AWS D1.1 governs structural steel welding, AWS D1.5 governs highway bridges, API 1104 governs pipeline welding, and ASME Section IX governs welding procedure and welder qualification for pressure equipment along with the acceptance criteria in the relevant ASME Boiler and Pressure Vessel Code section. Each sets specific, measurable limits for porosity, undercut, lack of fusion, and other discontinuities that a defect is evaluated against.
Technical briefings from our work in this area.
Welding procedure, qualification record and welder credential are examined twice — once for internal coherence, and once against the metallurgy of the joint, which keeps its own record of what happened.
readPorosity, slag, undercut and lack of fusion are discontinuities. Only the acceptance criteria of the governing code make one a defect — and only the inspection actually performed could have found it.
readThe weld deposit is often the strongest part of a welded joint. Failures frequently run through the narrow band of base metal beside it — never molten, but metallurgically rewritten by the arc.
readTell us what happened. We will triage it and connect you with the right expert — usually within one business day.