Residual Stresses in Welds and Their Effect on Fatigue Life

Welding is one of the most common joining processes in engineering, but it inevitably introduces residual stresses into the component. These locked-in stresses — particularly the tensile residual stresses near the weld toe — can dramatically reduce fatigue life, sometimes by a factor of two or more. This article explains where weld residual stresses come from, why they are so damaging for fatigue, how they can be estimated or measured, and what strategies are available to mitigate their effects.

Sources of Residual Stresses

Residual stresses are stresses that remain locked in a material after the original cause has been removed. Unlike stresses from external loads, they result from manufacturing processes that involve plastic deformation, temperature gradients or phase transformations. Machining, casting, forging and welding all generate residual stresses. Of these, welding produces some of the highest magnitudes — tensile residual stresses in a welded joint can approach the yield strength of the base material.

How Residual Stresses Arise During Welding

During welding, the localised heat input creates steep temperature gradients. The heated zone expands, but the surrounding cool material restrains this expansion, causing plastic compression of the hot weld metal. As the weld cools and contracts, the plastically deformed zone tries to shrink but is restrained by the surrounding base material, leaving behind tensile residual stresses in the weld and heat-affected zone (HAZ), balanced by compressive stresses further away (see Figures 1 and 2).

Cooling down process of a weld resulting in contraction of the weld material and tensile residual stresses — **Figure 1.** Contraction of the weld material during cooling, resulting in tensile residual stresses (click for larger image).

Residual stresses along a weld - the compressive and tensile stresses are shown — **Figure 2.** Tensile (red) and compressive (green) residual stresses along the length of a weld (click for larger image).

Effects on Fatigue and Durability

Tensile residual stresses near the weld toe are particularly harmful for several reasons:

Fatigue life reduction: tensile residual stresses raise the mean stress of the fatigue cycle, accelerating both crack initiation and propagation. In unwelded components, only the tensile portion of a load cycle typically drives fatigue crack growth. In a welded joint with high tensile residual stresses, the entire cycle — including what would otherwise be a compressive phase — can become fully tensile and contribute to incremental damage (see Figure 3). This is why weld fatigue standards such as BS 7608 and EN 1993-1-9 require the full stress range to be used in the fatigue assessment, regardless of the R-ratio of the applied loading.
Stress corrosion cracking: in corrosive environments, tensile residual stresses can initiate and drive stress corrosion cracking (SCC), where cracks develop under the combined action of sustained tensile stress and a corrosive medium.
Distortion: the non-uniform distribution of residual stresses can warp the component, affecting dimensional accuracy and fit during assembly.

Effect of a residual stress on the mean stress of a stress cycle — **Figure 3.** **Left:** applied stress cycle with tensile and compressive parts — typically only the tensile part drives fatigue crack growth. **Right:** a tensile residual stress shifts the entire cycle into the tensile regime.

It is worth noting that residual stresses are not always detrimental. Compressive residual stresses at the surface retard crack initiation and can significantly improve fatigue life. This principle is exploited deliberately in processes like shot peening. However, since residual stresses must be in equilibrium across the cross-section, introducing compressive residual stresses at one location inevitably creates tensile stresses elsewhere — a trade-off that must be understood and managed.

Estimating and Measuring Residual Stresses

Finite Element Analysis

Finite Element Analysis can simulate the welding process — heat input, thermal gradients, phase transformations and elasto-plastic material response — to predict the magnitude and distribution of residual stresses. This capability allows engineers to evaluate different welding sequences, joint designs or clamping strategies before fabrication, reducing the need for costly trial-and-error on real components.

Experimental Measurement

Several experimental techniques exist, grouped by the degree of material removal required:

Destructive methods: the contour method produces a two-dimensional residual stress map by making a precise cut through the component and measuring the resulting surface deformation; the slitting method determines through-thickness stress by incrementally deepening a narrow slit and recording strain release at each step.
Semi-destructive methods: centre hole drilling and deep hole drilling measure residual stresses by drilling a small hole and recording the strain relief around it. Centre hole drilling is suited to near-surface stresses; deep hole drilling can probe stresses deep within thick-walled components.
Non-destructive methods: X-ray diffraction measures surface residual stresses by detecting changes in lattice spacing caused by stress; neutron diffraction uses the deep penetration of neutrons to map stresses within the bulk of the material; ultrasonic testing infers stresses from changes in sound wave velocity.

Mitigation Strategies

Several strategies can reduce or redistribute weld residual stresses to improve fatigue and durability performance:

Pre-welding treatments: preheating reduces the thermal gradient between the weld zone and the base material, which in turn reduces the magnitude of residual stresses that develop during cooling.
Post-weld heat treatment (PWHT): stress relief annealing at a controlled temperature allows the material to relax plastically, reducing residual stresses significantly. PWHT is required by many design codes for thick-walled pressure equipment.
Mechanical post-weld treatments: shot peening, laser peening, hammer peening (HFMI) and roller burnishing introduce beneficial compressive residual stresses at the weld toe, directly opposing the harmful tensile stresses and greatly improving fatigue life.
Optimised welding parameters: lower heat input, controlled interpass temperatures, balanced welding sequences and appropriate joint design all help to minimise the residual stress field without the need for post-processing.

Conclusion

Weld residual stresses are an unavoidable consequence of the welding process, but they are not an uncontrollable one. Understanding the mechanism — localised plastic deformation during cooling creates tensile stresses that shift the fatigue cycle into a more damaging regime — is the first step towards managing the problem. By combining FEA prediction, targeted measurement and appropriate mitigation (heat treatment, mechanical peening or optimised welding parameters), engineers can ensure that welded structures achieve their intended design life.

For more on fatigue assessment of welded and unwelded components, see our course Introduction to Fatigue Analysis with FEA. For projects requiring specialist weld fatigue engineering, our fatigue analysis team can help.

Frequently asked questions

Common questions about weld residual stresses and fatigue.

Weld fatigue design curves in standards such as BS 7608 and EN 1993-1-9 are derived from test data on as-welded specimens that already contain high tensile residual stresses. Because these residual stresses shift the local stress cycle into the fully tensile regime regardless of the applied R-ratio, the compressive part of an externally applied cycle still contributes to fatigue damage. The codes account for this by requiring the full stress range to be used, which implicitly assumes the worst-case residual stress state.

In heavily restrained joints, longitudinal tensile residual stresses can approach or even reach the yield strength of the weld metal or the heat-affected zone. In structural steel welds this can mean 300–500 MPa of tensile residual stress at the weld toe. Transverse residual stresses are generally lower but still significant. The exact magnitude depends on the joint geometry, plate thickness, degree of restraint, welding process and heat input.

Not completely, but a properly executed PWHT typically reduces residual stresses to 10–30 % of the as-welded level, which is usually sufficient to remove their detrimental effect on fatigue. The effectiveness depends on the hold temperature, hold time, heating and cooling rates, and the material. Some design codes, including EN 13445 for pressure vessels, give explicit PWHT requirements and allow a more favourable fatigue assessment when PWHT has been performed.