Residual Stresses in Welds: Causes, Effects and How to Eliminate Them
Residual stresses in welds are stresses that remain locked within the material after all external loads have been removed and the weld has fully cooled to ambient temperature. They are an almost inevitable consequence of the non-uniform thermal cycle that characterises every arc welding process, and their magnitude can reach the room-temperature yield strength of the base metal or weld metal. For welding engineers, fabricators, and inspectors, understanding how residual stresses form, what damage they cause, and how to reduce them to acceptable levels is not optional knowledge — it is a core professional competency.
The effects of unmanaged residual stress span a wide range of failure modes: accelerated fatigue crack initiation, stress corrosion cracking in corrosive service environments, hydrogen-assisted cold cracking in susceptible steels, dimensional instability during subsequent machining, and in extreme cases, brittle fracture without any applied external load. Every post-weld heat treatment requirement, preheat specification, and interpass temperature limit in a qualified welding procedure specification (WPS) exists — at least in part — to manage the residual stress state left behind by the welding arc.
This article covers the complete picture: the thermomechanical mechanism that produces residual stresses, the typical stress distributions in practical weld geometries, the quantitative effects on structural performance, and the three principal mitigation methods — thermal stress relief (PWHT), vibratory stress relief, and peening — with code references and inspector guidance. It also includes a worked PWHT parameter example for carbon steel under ASME Section VIII Division 1.
1. The Origin of Weld Residual Stresses
The thermomechanical sequence that produces residual stress in a weld can be described in three distinct phases. Understanding each phase is essential for selecting appropriate mitigation measures.
1.1 Phase 1 — Heating: Expansion Under Restraint
When the welding arc is applied, metal in the immediate weld zone is heated rapidly to temperatures at or above the solidus. The thermal energy causes intense atomic vibration, increasing the mean interatomic spacing — the metal expands. However, the adjacent cooler base metal is physically bonded to the hot zone and resists this expansion. The hot zone cannot expand freely; instead, it is placed under compressive stress by the cooler surrounding metal.
Because the hot metal has a substantially reduced yield strength at elevated temperature, it cannot sustain this compressive stress elastically. It deforms plastically in compression — it is “upset” or shortened in the longitudinal direction. This plastic compression of the hot zone is the critical event that sets up the residual stress state upon cooling.
1.2 Phase 2 — Cooling: Contraction Under Restraint
As the arc moves on and the weld zone cools, the contracted (plastically upset) metal wants to shrink further as temperature drops. It is again restrained by the surrounding base metal. Now the situation is reversed: the cooling weld zone is pulled in tension by the restraining base metal. Because it is already shorter than it would naturally be (due to the plastic compression in Phase 1), this restraint generates tensile stress.
As temperature drops to ambient, the cooling weld metal and HAZ are unable to contract freely. The resulting tensile residual stress is locked into the weld and HAZ. By equilibrium, the wider base metal region remote from the weld must carry complementary compressive stresses.
1.3 Phase 3 — Equilibrium: The Final Residual Stress State
The final residual stress state must satisfy internal equilibrium — the sum of forces and moments across any cross-section must be zero, because no external loads are applied. This means:
- Tensile residual stress — weld metal and HAZ (peak values near fusion boundary)
- Compressive residual stress — base metal remote from the weld (balancing the tension)
2. Residual Stress Distribution in Practical Weld Geometries
The distribution of residual stress is not uniform and depends on weld geometry, section thickness, degree of restraint, number of weld passes, and preheat conditions. Understanding typical distributions helps the engineer and inspector identify the highest-risk regions of a weldment.
2.1 Longitudinal vs Transverse Residual Stress
| Stress Component | Direction | Typical Peak Magnitude | Most Critical Location |
|---|---|---|---|
| Longitudinal | Parallel to weld run | Up to yield strength | Weld centreline and HAZ |
| Transverse | Perpendicular to weld run | 30 – 60% of yield strength | Weld ends; highly restrained joints |
| Through-thickness | Normal to weld surface | Significant in thick sections | Mid-thickness of heavy sections (>50 mm) |
2.2 Effect of Section Thickness
In thin sections (below approximately 12 mm), the through-thickness temperature gradient is small, and the residual stress state is essentially biaxial (longitudinal and transverse). In thick sections, the slower heat extraction rate creates significant through-thickness gradients, and a triaxial stress state develops at the weld root region. Triaxial tension is particularly damaging because it suppresses the plastic deformation that would otherwise redistribute stress, reducing effective fracture toughness and making the weld root region the most vulnerable location in heavy-section vessels.
2.3 Effect of Joint Restraint
A weld made in a freely supported, low-restraint configuration will distort significantly but retain lower residual stress, because the weld metal and HAZ can contract by moving the joint. A highly restrained weld — such as a closure weld in a heavy pressure vessel or a weld in a stiff structural frame — cannot distort and therefore develops much higher residual stresses. This is the engineering basis for the general principle that high restraint requires more conservative welding procedures, including lower heat input to reduce the volume of material subject to the thermal cycle, mandatory preheat, and often mandatory PWHT.
3. Effects of Residual Stress on Weldment Performance
3.1 Fatigue Life Reduction
Fatigue failure occurs by progressive crack initiation and propagation under cyclic loading, even at stress levels well below the static yield strength. The driving force for fatigue crack growth is the stress range — the algebraic difference between maximum and minimum stress in each cycle. Tensile mean stress (including residual stress) is additive to the applied cyclic stress, increasing the effective mean stress and therefore accelerating crack initiation.
This is the reason why welded joints are classified with significantly lower allowable fatigue stress ranges than equivalent plain (unwelded) sections in fatigue design standards such as BS 7608 and ASME VIII Appendix 2. The fatigue classification accounts for the combined effect of the weld geometry stress concentration and the tensile residual stress.
3.2 Stress Corrosion Cracking (SCC)
Stress corrosion cracking requires three simultaneous conditions: a susceptible material, a corrosive environment, and tensile stress. Residual stresses in welds routinely satisfy the tensile stress condition — and they do so at stress levels that can equal the yield strength — even before any service loads are applied. The two most commercially significant SCC failure modes driven by weld residual stress are:
- Chloride SCC of austenitic stainless steel — occurs in the presence of chlorides, temperatures above ~60 °C, and tensile stress. Weld HAZ residual stresses in 304 and 316 type stainless steel piping are a well-documented cause of this failure mode in chemical plant and oil and gas service. See the stainless steel weld decay guide for the related sensitisation mechanism.
- Caustic SCC of carbon steel — high-temperature caustic (NaOH) solutions can cause SCC in carbon steel welds with high residual stress, particularly in heat exchangers and boiler equipment. PWHT is routinely specified to mitigate this risk.
In sour service environments (H₂S-containing hydrocarbon systems), NACE MR0175 / ISO 15156 limits the hardness of weld metal and HAZ precisely because high hardness correlates with high residual stress and susceptibility to hydrogen-induced cracking (HIC) and sulphide stress cracking (SSC).
3.3 Hydrogen-Assisted Cold Cracking
Hydrogen-assisted cold cracking (HACC), also called hydrogen-induced cracking (HIC) or delayed cracking, requires four concurrent factors: hydrogen in the weld/HAZ, a susceptible (typically martensitic) microstructure, tensile stress (residual or applied), and temperature below approximately 200 °C. Tensile residual stresses — which can equal the yield strength in the HAZ nearest the fusion line — provide the stress component that drives crack opening once a hydrogen-embrittled region forms.
The P91 creep-resistant steel welding guide on WeldFabWorld details how HACC risk is managed in high-alloy power plant steels through mandatory preheat, immediate PWHT after welding, and strict hydrogen control procedures.
3.4 Dimensional Instability During Machining
When material is removed from a weldment by machining — for example, to achieve tight dimensional tolerances on a pressure vessel flange face or a pump casing — the residual stress equilibrium is disrupted. The remaining material redistributes the unbalanced internal forces by deforming. This is why rough-machined weldments commonly spring or warp when finish-machined, and why stress relief before precision machining is standard practice for complex fabrications.
3.5 Brittle Fracture Without Applied Load
In susceptible steels at low temperatures or in the presence of brittle microstructures, residual stresses alone can provide sufficient driving force for brittle fracture. This is particularly relevant in heavy-section reactor pressure vessels and storage tanks during hydraulic pressure testing at low ambient temperatures, where the combination of residual stress, stress concentration at weld defects, and reduced fracture toughness can create fracture conditions despite the absence of applied operating stress. The requirement for UG-84 Charpy impact testing in ASME Section VIII Div. 1 is partly a response to this risk.
4. Methods for Reducing or Eliminating Residual Stresses
Three principal methods are used in industry to reduce weld residual stresses. Each has specific applications, advantages, and limitations that determine its suitability for a given fabrication.
Heating the weldment or a defined band around the weld to a specified temperature, holding, then cooling uniformly. The most reliable and code-recognised method. Required by ASME, AWS, EN 13445, and most pressure equipment codes for susceptible materials and above threshold thicknesses.
Applying resonant vibration or mechanical compressive forces to redistribute peak residual stresses. No furnace required. Effective for certain applications, but not universally code-accepted as equivalent to thermal treatment.
Controlled impact on intermediate weld passes during multipass welding. Counteracts shrinkage stresses between passes. Applied during welding, not after. Subject to specific restrictions on which passes may be peened.
4.1 Thermal Stress Relief (PWHT) — Detailed Treatment
Thermal stress relief works because the yield strength of steel decreases significantly with temperature. When the weldment is uniformly heated to the PWHT soak temperature, the metal can no longer sustain the room-temperature residual stress level — it yields locally in a slow, controlled manner, reducing peak stress toward the yield strength at the soaking temperature, which may be only 30 to 50% of the room-temperature yield strength.
Key Parameters the Inspector Must Monitor
- Thermocouple placement and calibration — Thermocouples must be positioned to verify that the entire weld zone and a minimum band of base metal on each side reach and maintain the specified soak temperature. Calibration records must be current.
- Temperature uniformity — The temperature difference between any two thermocouples monitoring the same weld during soaking must be within the code-permitted range (typically ±15 °C for the soak band).
- Heating and cooling rate compliance — Rates that are too fast cause thermal shock and can introduce or worsen distortion. Charts or data logger output must confirm compliance.
- Soak time verification — The soak time starts when the last thermocouple reaches the minimum specified temperature and must run continuously until all thermocouples remain within the specified range for the full hold period.
- Protection from environment — Stainless steel and nickel alloy welds being PWHT’d must be protected from sulphur contamination (e.g., marker compounds, lubricants) which can cause liquid metal embrittlement or sulphidation at elevated temperature.
4.2 Local vs Furnace PWHT
For large vessels or site-welded piping, furnace PWHT may not be practical. Local PWHT using electrical resistance heating elements, induction heaters, or gas-fired equipment is code-permitted under specific conditions, provided the soak band width meets code requirements:
| Method | Typical Application | Soak Band Width (each side of weld) | Advantages | Limitations |
|---|---|---|---|---|
| Furnace PWHT | Pressure vessels, headers | Full component | Best temperature uniformity; most thorough stress relief | Size-limited; transport required; costly |
| Local PWHT (resistance) | Piping welds, nozzles | Min. 3× weld width or 76 mm, whichever is greater | Site-applicable; no transport; flexible | Temperature gradient at band edges; inspector vigilance required |
| Local PWHT (induction) | Piping, vessel nozzles | Same as resistance | Fast heating; good control; low oxidation | Equipment cost; limited to smaller diameters |
4.3 Vibratory and Mechanical Stress Relief
Vibratory stress relief (VSR) involves mounting the weldment on vibration isolators and exciting it at one or more of its natural resonant frequencies using a variable-speed eccentric mass exciter. The resulting cyclic stresses cause local plastic deformation at stress concentration points (notably the weld toes and weld root), redistributing and reducing residual stress peaks. Typical treatment times are 20 to 60 minutes per resonant frequency.
VSR is particularly attractive for large weldments where furnace PWHT is impractical, for materials that are sensitive to the thermal cycling of PWHT (such as some age-hardened aluminium alloys), and for applications where the lead time of PWHT is a production bottleneck. However, VSR cannot address susceptibility to hydrogen cracking in the same way as thermal treatment because it does not facilitate hydrogen diffusion out of the HAZ.
Thermal Stress Relief (PWHT)
- Code-accepted under all major pressure equipment standards
- Reduces peak residual stress to ~30 – 50% of yield strength
- Tempers martensite; improves toughness
- Promotes hydrogen diffusion out of HAZ
- Requires furnace or controlled local heating equipment
- Risk of over-tempering for some alloy steels (e.g., P91)
Vibratory Stress Relief (VSR)
- Not universally code-accepted as equivalent to PWHT
- Typically reduces peak residual stress by 30 – 60%
- No microstructure change; no tempering effect
- Does not facilitate hydrogen diffusion
- No furnace required; applicable on-site
- Lower energy cost; shorter cycle time
4.4 Peening
Peening is a mechanical impact treatment applied to the surface of a deposited weld pass using a pneumatic hammer fitted with a rounded peening tip (not a sharp chipping tip). The impact causes the weld surface to spread laterally, introducing compressive plastic deformation that partially counteracts the shrinkage tensile stress that would otherwise accumulate between passes. Peening is most effective for heavy-section, highly-restrained multipass welds.
- Root pass — must NOT be peened. The root pass may be in a brittle condition (rapid cooling, possible martensite) and peening can initiate cracking that will propagate through subsequent passes.
- Final cap pass — must NOT be peened with heavy impact. Peening cold-works the surface, potentially closing over surface discontinuities and making them undetectable by visual inspection, PT, or MT.
- Intermediate passes only — peening is restricted to intermediate (fill) passes in a multipass weld.
- Equipment — use a pneumatic hammer with a round-nosed tip, NOT a chipping hammer. Chipping hammers can introduce notches.
5. Residual Stress Measurement Techniques
In most production fabrications, residual stress is managed by controlling the welding procedure and verifying PWHT rather than by direct measurement. However, direct measurement is performed in research, fitness-for-service assessments, and high-consequence applications such as nuclear pressure vessels and aircraft structures. The principal techniques are:
| Technique | Method | Destructive? | Depth Capability | Typical Accuracy |
|---|---|---|---|---|
| X-ray Diffraction (XRD) | Measures lattice strain from diffraction peak shift | Non-destructive | Surface only (~10 μm) | ±20 – 50 MPa |
| Neutron Diffraction | Same principle, but neutrons penetrate deeply | Non-destructive | Up to 50 mm in steel | ±20 MPa |
| Hole Drilling | Strain gauge rosette measures relaxation on hole drilling | Semi-destructive | Up to 2 mm depth | ±20 – 40 MPa |
| Contour Method | Cut + measure surface profile; compute from elastic relaxation | Destructive | Full cross-section | ±20 MPa |
| Ultrasonic (EMAT/LCR) | Measures stress-induced velocity change in ultrasonic waves | Non-destructive | Surface to mid-thickness | ±30 – 70 MPa |
6. Residual Stress Management in Specific Applications
6.1 Pressure Vessels and Piping (ASME Code)
ASME Section VIII Division 1 requires PWHT for carbon and low-alloy steel vessels above certain weld thickness thresholds (generally 38 mm for P-No. 1, or for any thickness where material composition exceeds defined limits). The specific requirements are tabulated in UCS-56 (carbon steels), UHT-56 (high-tensile steels), and related tables. PWHT is mandatory regardless of thickness for some P-Numbers and environments where SCC or HACC risk is explicitly identified in the design.
For P91 (P-No. 5B) chromium-molybdenum creep-resistant steel, PWHT is mandatory for all thicknesses and the soak temperature window is tight (730 – 775 °C depending on base material heat treatment condition). Over-tempering P91 below the minimum temperature produces incomplete tempering; exceeding the maximum temperature re-austenitises and effectively anneals the weld, eliminating the creep strength that is the whole purpose of the material selection.
6.2 Duplex Stainless Steel
Duplex stainless steel welds are generally not subject to thermal PWHT in the conventional sense, because the soak temperatures used for carbon steel (593 – 649 °C) would precipitate sigma phase and other intermetallics, severely embrittling the weld. Solution annealing at full austenitising temperature followed by water quenching can be used for extreme cases, but this is not practical for large fabrications. Residual stress management in duplex stainless steel instead relies on controlling heat input, interpass temperature, and performing compressive surface treatments (shot peening in compression mode) where SCC in chloride service is a concern.
6.3 Structural Steelwork (AWS D1.1)
In structural steelwork governed by AWS D1.1, PWHT is not mandatory for all joints. Weld procedures for prequalified joints in common structural grades do not routinely specify PWHT unless the material or environment specifically demands it. Residual stress effects on fatigue are addressed through joint category classification in fatigue design, which already incorporates the effect of typical weld toe residual stresses. Post-weld improvement treatments such as high-frequency mechanical impact (HFMI) are increasingly used on critical fatigue-loaded structures to introduce compressive surface residual stresses and improve weld toe geometry simultaneously.
7. The Inspector’s Role in Residual Stress Management
The welding inspector is the key line of defence ensuring that residual stress mitigation measures are correctly applied. Inspector responsibilities include the following:
- Pre-PWHT: Verify that the correct thermocouple type, quantity, and placement are specified and implemented. Confirm calibration records are current. Check that the heating equipment rating is adequate for the section thickness.
- During PWHT: Monitor the chart recorder or data logger output for heating rate, soak temperature band, uniformity between thermocouples, and hold time compliance. Flag any exceedance of maximum temperature or violation of heating rate limits immediately — these may require engineering disposition.
- Post-PWHT: Confirm the cooling rate through the controlled range. Verify that the PWHT record is signed, dated, and includes all required data (equipment calibration ref, thermocouple positions, temperature vs time chart).
- Peening: If peening is specified, verify that it is applied only to permitted intermediate passes, using correct equipment. Confirm the root and cap passes are not peened.
- Documentation: Ensure all PWHT and peening records are incorporated into the weld data package and are traceable to the specific weld joint(s) treated.
Recommended Technical Resources
Disclosure: WeldFabWorld participates in the Amazon Associates programme (StoreID: neha0fe8-21). If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.
Frequently Asked Questions
What causes residual stresses in welded joints?
Residual stresses in welds arise from the non-uniform thermal cycle of welding. The weld zone is heated rapidly to very high temperatures and expands, but is restrained by the cooler surrounding base metal. Because the hot metal is weaker, it deforms plastically in compression rather than forcing the surrounding metal to expand. On cooling, the weld zone contracts but is again restrained, locking in tensile residual stresses in the weld and HAZ, and compensating compressive stresses in the adjacent base metal. The magnitude of these stresses can approach the room-temperature yield strength of the material.
How large can residual stresses be in a weld?
Longitudinal residual stresses (parallel to the weld run) can reach the yield strength of the weld metal or HAZ material at the weld centreline — commonly 250 to 550 MPa in carbon and low-alloy steels. Transverse residual stresses (perpendicular to the weld) are typically lower but can still be significant, particularly in highly restrained joints. Through-thickness stresses in thick sections add a triaxial stress state that further reduces material toughness. This is why high-restraint joints in thick sections are the highest-risk combination for hydrogen-assisted cold cracking.
What is the difference between residual stress and distortion in a weldment?
Both residual stress and distortion arise from the same non-uniform thermal cycle, but they are complementary manifestations rather than independent problems. Distortion occurs when a weldment is free to move in response to the internal contraction forces — the structure deforms to partially relieve internal stress. Residual stress occurs when the structure cannot move (due to its own stiffness or external restraint) — the internal forces are locked in as stress rather than released as movement. Highly restrained joints exhibit high residual stress and little distortion; flexible, unrestrained joints show more distortion but lower locked-in stress.
What temperature is used for PWHT of carbon steel?
For carbon steels under ASME Section VIII Division 1, the minimum PWHT temperature is 593 °C (1100 °F) for P-No. 1 Group 1 and 2 materials, with a maximum of 649 °C (1200 °F) in most cases to avoid over-tempering. The hold time is typically 1 hour per 25 mm (1 inch) of weld thickness, with a minimum of 15 minutes. Heating and cooling rates are limited as specified in UCS-56. Specific requirements must always be taken from the applicable code and relevant WPS. For P91 steel (P-No. 5B), a different and tighter soak range applies — refer to the P91 welding guide.
Can residual stresses cause cracking without any applied load?
Yes. Residual stresses are a form of internal loading, and in susceptible situations they can drive crack initiation and propagation without any external service load. The three most significant examples are: hydrogen-assisted cold cracking, where tensile residual stress provides the stress component needed for crack propagation in the presence of hydrogen and a susceptible microstructure; stress corrosion cracking, where tensile residual stress combined with a corrosive environment causes environmentally-assisted cracking well below nominal yield strength; and delayed cracking in high-strength steels, where high locked-in residual stresses can cause cracking hours or days after welding is complete and while the component is in storage.
What does PWHT actually do to residual stresses at the atomic level?
At elevated PWHT temperatures, the yield strength of the metal decreases significantly — typically to 30 to 50% of the room-temperature value. Atoms in the highly stressed regions have sufficient thermal energy to creep — to slowly migrate toward lower-energy equilibrium positions. Dislocations become mobile and rearrange to reduce internal stress concentrations. The locked-in tensile stress cannot be sustained and the material yields locally in a slow, controlled manner, reducing peak residual stress to approximately the yield strength at the soaking temperature. Upon slow, uniform cooling, the material retains this reduced stress state. Hardness and brittleness of martensitic regions are also reduced by the tempering effect, which is a secondary benefit beyond stress relief.
Why is the root pass not peened during multipass welding?
The root pass is typically the thinnest, most restricted deposit and may be in a brittle condition due to rapid cooling and possible martensitic transformation in the surrounding HAZ. Peening involves striking the weld surface with significant impact force. Applied to a brittle root pass, this impact can initiate or propagate cracks that may not be detectable but will compromise the integrity of the final joint. Codes and good practice therefore generally prohibit peening of the root pass. Similarly, the final cap pass is not peened with heavy impact because peening cold-works the surface and can close over surface discontinuities, making them undetectable by visual inspection and surface NDE methods such as PT or MT.
How does residual stress affect stress corrosion cracking in stainless steel?
Austenitic stainless steels are susceptible to chloride stress corrosion cracking (Cl-SCC) when tensile stress, a chloride-containing environment, and temperatures above approximately 60 °C are simultaneously present. Tensile residual stresses in the weld and HAZ can reach the material yield strength and therefore easily satisfy the stress condition for SCC initiation, even when applied service stresses are low. Mitigation approaches include PWHT by solution annealing where practical, selecting SCC-resistant grades such as duplex stainless steels, applying compressive surface treatments, and controlling the chloride content and temperature of the service environment.