Welded Joints: Geometry, Load Transfer and Structural Behaviour

Weld Geometry: Fillet, Butt, Groove, Lap, T and Corner

The geometry of a welded joint is determined by the relative position and the shape of the two components being joined. A fillet weld joins two components at an angle — typically a T-joint, a lap joint or a corner joint — and the weld is a triangular cross-section in the angle between the two components. A butt weld joins two components edge-to-edge — the edges are prepared (square, V, U, J, bevel) and the weld fills the groove between them. A groove weld is a general term for a full-penetration or partial-penetration weld in a prepared groove. A lap weld joins two overlapping components — typically with a fillet weld at the overlap. A T-joint joins two components at right angles — one component is the web, the other is the flange — and the weld is a fillet or a full-penetration weld on one or both sides. A corner joint joins two components at an angle to form a corner — the weld can be a fillet on the inside, a butt weld on the outside, or a combination. Each geometry has a different load path, a different stress flow and a different set of stress concentrations. The selection of the geometry depends on the load (tension, shear, bending, combined), the manufacturing constraints (access, position, fit-up), and the inspection requirements.

Weld Features: Throat, Leg, Toe, Root and Fusion Zone

A weld has several geometric features that are critical to its strength and its behaviour. The weld throat is the shortest distance from the root of the weld (the deepest point of penetration) to the weld face (the outer surface). For a fillet weld with equal leg lengths, the throat is the leg length × sin(45°) ≈ 0.707 × leg. The throat is the critical dimension for fillet weld strength — the throat area (throat × weld length) is the area that resists shear. The weld leg is the visible size of the fillet — the distance from the root to the toe along each component. The leg length is the specified weld size, but the throat is the dimension that determines the strength. The weld toe is the transition point from the weld face to the parent material — it is a stress concentration, and it is the primary fatigue crack initiation site. The weld root is the deepest point of the weld — for a partial-penetration weld, the root is an unfused edge and a stress concentration. The fusion zone is the region where the weld metal and the parent material have melted and mixed — the boundary between the fusion zone and the HAZ is the fusion line. The heat-affected zone (HAZ) is the region of parent material that has been thermally affected but not melted — the microstructure and the properties are changed by the thermal cycle, and the HAZ can be harder or softer than the parent material, with different toughness and fatigue properties.

The weld throat is the critical dimension for fillet weld strength — it is the shortest distance from the root to the face, and the throat area resists shear. For an equal-leg fillet, the throat ≈ 0.707 × leg. The weld toe is the primary fatigue crack initiation site — the transition from weld to parent material is a stress concentration.

Load Transfer in Fillet and Butt Welds

The load transfer mechanism is different for fillet and butt welds. In a butt weld (full penetration), the weld metal fills the entire groove and the load passes directly through the weld — the weld is in the same stress state as the parent material (tension, compression, or shear, depending on the load direction). A full-penetration butt weld, properly executed, is as strong as the parent material — the weld is not a discontinuity in the load path. In a fillet weld, the load transfers through the weld throat in shear. The load path is not direct — the load must change direction to pass through the throat, and the eccentricity of the load path (in a lap joint or a T-joint) produces local bending and additional stress. The fillet weld is inherently less efficient than the butt weld: the throat area is smaller than the parent material area, and the load path is indirect. The fillet weld is simpler and cheaper to make (no edge preparation, no backing, single pass for moderate sizes), but it is weaker and more fatigue-sensitive. For a T-joint with a fillet weld on both sides, the load transfers through both throats — the total throat area is the sum of the two throats. For a lap joint with a fillet weld, the load transfers through the throat in shear, and the eccentricity of the overlap produces bending in the plates and additional stress at the weld.

  Fillet weld throat (equal legs):

  throat = leg × sin(45°) ≈ 0.707 × leg

  Throat area (single fillet):

  A_throat = throat × L_weld = 0.707 × leg × L_weld

  →  The throat area resists shear in a fillet weld
  →  For double-sided fillet: A_total = 2 × A_throat
  →  A full-penetration butt weld has the full parent-material area

A full-penetration butt weld transfers load directly — the weld is as strong as the parent material. A fillet weld transfers load through the throat in shear, with an indirect load path and local bending from eccentricity. The fillet weld is simpler but weaker and more fatigue-sensitive.

Full Penetration versus Partial Penetration

The depth of penetration — how far the weld metal fuses into the joint — determines whether the weld is full penetration or partial penetration. A full-penetration weld fills the entire thickness of the joint — the weld metal extends from one surface to the other, and the load-carrying area is the full thickness. A full-penetration weld is the strongest — it is equivalent to the parent material in static strength, and it has the best fatigue performance (no root defect). A partial-penetration weld fills only part of the thickness — the root of the weld is an unfused region, and the load-carrying area is the fused area only. The partial-penetration weld is weaker — the reduced area increases the stress, and the unfused root is a stress concentration and a potential crack initiation site. Partial-penetration welds are used where full penetration is not required (the load is low, or the weld is not primary), or where full penetration is not practical (very thick sections where multiple passes would be required, or where access for welding from the back side is not possible). The engineer must specify the required penetration and verify that the welding procedure achieves it. The penetration is verified by ultrasonic testing or radiography — the inspection must confirm that the fused area meets the specification. A partial-penetration weld designed as full penetration (because the inspection was inadequate) is a critical defect — the root defect can initiate fatigue cracks that propagate undetected.

A full-penetration weld fills the entire thickness and is as strong as the parent material. A partial-penetration weld fills only part of the thickness — the unfused root is a stress concentration and a crack initiation site. Partial penetration is used where full penetration is not required or not practical. Specify and verify the penetration by inspection.

Intermittent and Circumferential Welds

An intermittent weld is a fillet weld with gaps — the weld is deposited in segments along the joint, with unwelded gaps between segments. Intermittent welds are used where a continuous weld is not required — for attaching stiffeners, brackets and secondary components where the load is low and the continuous weld would be over-designed. The intermittent weld reduces the welding time, the heat input and the distortion. The design of an intermittent weld checks the stress in each weld segment and the stress in the plate between segments. The segment length and the gap must be specified — the segment is typically 25–75 mm, and the gap is similar. The intermittent weld is not suitable for fatigue-critical joints or for joints where the load varies along the length (the gaps create stress concentrations at the segment ends). A circumferential weld is a weld around a circular joint — a pipe butt weld, a nozzle-to-shell weld, a shaft splice. The load transfer in a circumferential weld is axisymmetric — the load is distributed around the circumference, and the stress state is uniform (for axial load) or varies sinusoidally (for bending). The circumferential weld is a butt weld (for pipe splices) or a fillet/flare weld (for nozzle attachments). The residual stress and the distortion in a circumferential weld are axisymmetric — the weld shrinkage produces a circumferential contraction that can ovalise the pipe or the shell.

Intermittent welds (segmented fillets) reduce welding time, heat input and distortion for low-load attachments. They are not suitable for fatigue-critical joints — the segment ends are stress concentrations. Circumferential welds (pipe splices, nozzle attachments) have axisymmetric load transfer and axisymmetric residual stress and distortion.

Residual Stress and Distortion

Welding inherently produces residual stress and distortion because of the localised heating and cooling. The weld metal and the HAZ are heated to melting temperature and then cool to ambient. As the hot metal contracts, it is restrained by the cooler surrounding parent material, and the contraction produces tensile residual stress in the weld and the HAZ — typically up to the yield strength of the material. The surrounding parent material is put into compression to balance the tension. The residual stress is superimposed on the applied stress — at the weld toe, the residual tension adds to the applied tension, and the total stress can reach yield even under moderate applied load. This affects the fatigue life (the tensile residual stress reduces the fatigue life) and the static capacity (the residual stress can reduce the plastic capacity if the section is already at yield before the load is applied). The distortion is the visible effect of the residual stress — the structure bends, twists or buckles as the welds contract. The distortion affects the fit-up of the assembly, the alignment of components and the appearance. The distortion is controlled by weld sequencing (welding in a balanced pattern to cancel the shrinkage), by pre-setting (pre-bending the components so that the distortion brings them back to the desired position), by clamping (restraining the components during welding), and by post-weld heat treatment (stress relief to reduce the residual stress). The residual stress and distortion are covered in detail in Article 26.

Welding produces tensile residual stress (up to yield) in the weld and HAZ, and corresponding distortion. The residual stress adds to the applied stress, affecting fatigue and static capacity. Control by weld sequencing, pre-setting, clamping and post-weld heat treatment. See Article 26 for the detailed treatment.

Stiffness, Eccentricity and the Load Path

The stiffness of the welded joint affects how the load flows through the structure and where the stress concentrations arise. A full-penetration butt weld has the same stiffness as the parent material — the load flows through without a discontinuity, and there is no additional stress concentration (beyond the weld toe geometry). A fillet weld introduces a local stiffness change — the weld metal is in a different geometry (triangular) and the load must change direction to pass through the throat. The stiffness change produces a local stress concentration at the weld toe, and the load redirection produces local bending in the plates. The eccentricity of the load path is a key factor: in a lap joint with a single fillet weld, the load path shifts from the centreline of one plate to the centreline of the other through the weld, and the offset produces a bending moment on the weld and the plates. In a T-joint with a single-sided fillet, the load on the web produces a bending moment on the flange through the weld. The eccentricity increases the stress in the weld (the bending adds to the shear) and the stress in the plates (the bending adds to the membrane stress). A double-sided weld (fillet on both sides of a T-joint, or two fillets on a lap joint) reduces the eccentricity by making the load path symmetric. The engineer should analyse the load path and the eccentricity of every welded joint — the local bending from eccentricity is often the governing stress.

The weld stiffness and the load path eccentricity affect the stress concentration and the local bending. A full-penetration butt weld has no discontinuity. A fillet weld introduces a stiffness change and a load redirection. The eccentricity of a single-sided weld produces bending that adds to the shear — double-sided welds reduce the eccentricity.

Local Stress Concentration at the Weld Toe

The weld toe is the primary stress concentration in a welded joint, and it is the location where most fatigue cracks initiate. The toe is the transition from the weld face to the parent material — the geometry of this transition determines the stress concentration factor. A sharp transition (high weld angle, small toe radius) gives a high stress concentration; a smooth transition (low weld angle, large toe radius) gives a lower concentration. The weld toe geometry depends on the welding process, the welding parameters and the welder technique. As-deposited welds have a variable toe geometry — the toe radius can be as small as 0.1–1 mm, and the weld angle can be 30–45°. The stress concentration factor at the toe can be 2–4 for a typical fillet weld, and higher for a weld with a sharp toe. The stress concentration can be reduced by weld toe improvement techniques: grinding the toe to a smooth radius (toe grinding), TIG dressing (remelting the toe to smooth the transition), hammer peening (introducing compressive residual stress at the toe), or shot peening. These techniques can improve the fatigue life by a factor of 2–4. The stress concentration is also affected by the weld misalignment and the axial misalignment of the plates — a misalignment introduces an additional bending stress at the weld. The local stress concentration at the weld toe is the basis for the fatigue assessment methods — the nominal stress, the hot-spot stress and the notch stress methods — which are covered in Article 25.

The weld toe is the primary fatigue-critical location. The toe geometry (radius, angle) determines the stress concentration — a sharp toe gives Kt = 2–4 or higher. Toe improvement (grinding, TIG dressing, peening) can improve the fatigue life by a factor of 2–4. The local stress concentration is the basis for the fatigue assessment methods in Article 25.

Key takeaways

• A welded joint forms a metallurgical bond by melting the parent material. The weld geometry — fillet, butt, groove, lap, T-joint, corner — determines the load path, the stress flow and the stress concentrations. The geometry must be selected for the load and the manufacturing constraints.
• The weld throat is the critical dimension for fillet weld strength — it is the shortest distance from the root to the weld face, and the throat area is the load-carrying area in shear. The leg length is the visible size of the weld, but the throat is what matters for strength.
• The weld toe is the primary fatigue-critical location — the transition from the weld face to the parent material is a stress concentration, and most fatigue cracks initiate at the toe. The weld root can also be critical, particularly for partial-penetration and fillet welds.
• Residual stress and distortion are inherent to welding. The thermal contraction of the weld metal and the HAZ produces tensile residual stress at the weld and distortion of the structure. These affect the fatigue life, the static capacity and the fit-up of the welded assembly.