Numerical Simulation of Martensitic Transformation Effects on Transverse Residual Stress in Weld Overlay

1. Definition and Fundamental Principles

1.1 Technical Definition

The numerical simulation of martensitic transformation effects on cooling transverse stress in weld overlay represents a computational metallurgical and mechanical analysis capability that models the coupled thermomechanical behavior of weld overlay deposits during solidification and cooling. Specifically, this technology addresses how the diffusionless, displacive phase transformation from austenite (γ-Fe) to martensite (α'-Fe) during post-weld cooling generates volumetric expansion, which interacts with thermal contraction to produce complex residual stress fields—particularly transverse (lateral) residual stresses that are critical to overlay bond integrity, crack susceptibility, and dimensional stability.

1.2 Governing Physical Mechanisms

The residual stress state in a weld overlay is governed by two primary mechanisms acting during the cooling cycle:

• Thermal contraction stress: As the weld metal and heat-affected zone (HAZ) cool from the solidus temperature to ambient, differential thermal contraction between the deposited material, base metal, and previously deposited layers generates compressive and tensile stress components. The transverse direction experiences significant restraint from adjacent deposited material and the substrate.
• Transformation stress: When the cooling path crosses the martensite start temperature (Ms), the austenite-to-martensite transformation produces a volumetric expansion of approximately 2–5% (depending on composition and alloying elements). This expansion acts as a strain source that partially offsets thermal contraction, potentially converting tensile transverse stresses to compressive values or reducing peak tensile magnitudes.

1.3 Mathematical Framework

The total strain tensor in the simulation is decomposed as:

ε_total = ε_elastic + ε_plastic + ε_thermal + ε_transformation

Where the transformation strain is expressed as:

ε_transformation = (ΔV/V_γ) × f(ξ) × (1/3) × I

Here, ΔV/V_γ is the volumetric expansion ratio of martensite relative to austenite, f(ξ) is the volume fraction of martensite as a function of cooling history, and I is the identity tensor. The volume fraction evolution is typically modeled using the Koistinen-Marburger equation:

f(ξ) = 1 − exp[−α(Ms − T)]

where α is a material-specific constant (typically 0.011–0.017 °C⁻¹ for Fe-C alloys) and T is the current temperature. The simulation solves coupled heat transfer and mechanical equilibrium equations iteratively at each time step, updating material properties (elastic modulus, yield strength, thermal conductivity, density) at every increment based on the current temperature and phase fraction.

2. Category and Business Positioning

2.1 Technical Classification

This capability falls under the category of Computational Welding Metallurgy and Process Simulation, which serves as a critical enabler across all three manufacturing technology routes of Cladding Technology Shanxi Co., Ltd. It bridges the gap between empirical process development and physics-based predictive modeling, allowing the company to:

• Optimize welding parameters for minimum residual stress in production WPS development
• Predict overlay bond stress states for design verification
• Establish quantitative acceptance criteria for residual stress levels
• Accelerate WPS qualification by reducing the number of physical trials
• Provide engineering justification for customer design reviews and code compliance

2.2 Strategic Business Value

Within the company's three core technology platforms—TIG/MIG weld overlay, hydraulic explosive bonding, and explosion welding—the numerical simulation capability provides differentiated value at distinct touchpoints:

• TIG/MIG Weld Overlay: Directly applicable to predicting and controlling residual stresses in multi-pass overlay builds, particularly for martensitic stainless steels (e.g., 410, 420, 17-4PH overlay), high-strength steels, and nickel-based alloys where transformation effects are pronounced.
• Hydraulic Explosive Bonding: Supports post-bond stress analysis and prediction of stress relief requirements following hydraulic explosive cladding, where differential thermal expansion and any post-process heat treatment must be evaluated against the bond interface integrity.
• Explosion Welding: Enables prediction of residual stress states in explosion-welded clad plates and pipes, informing post-weld heat treatment (PWHT) parameters and stress relief specifications to meet code requirements.

3. Technical Purpose and Value

3.1 Engineering Objectives

The primary engineering objectives of this simulation capability are:

1. Residual Stress Prediction: Quantify the magnitude, distribution, and sign (tensile vs. compressive) of transverse residual stresses in weld overlay deposits and HAZ regions with engineering accuracy (target deviation ≤15% from measured values).
2. Crack Susceptibility Assessment: Identify regions where peak tensile transverse stresses exceed material fracture toughness thresholds, predicting potential cracking locations before physical testing.
3. Process Parameter Optimization: Determine optimal heat input ranges, interpass temperature windows, and travel speed parameters that minimize detrimental residual stress states.
4. Dimensional Stability Prediction: Forecast post-weld distortion and warpage that affects subsequent machining tolerances and assembly fit-up.
5. WPS Qualification Support: Provide computational evidence to support welding procedure qualification submissions to NDT and certification bodies.

3.2 Value to Product Delivery

This capability directly contributes to product delivery quality and customer confidence through:

• Reduced risk of overlay spalling, cracking, or delamination in delivered clad products
• Shortened qualification timelines from 8–12 weeks to 4–6 weeks for new overlay specifications
• Quantitative residual stress data packages included with product delivery documentation
• Ability to demonstrate code compliance (ASME, NB/T, API) through predictive analysis supplemented by verification testing
• Customer-specific stress analysis reports for critical-service applications (nuclear, pressure vessels, offshore platforms)

4. Key Process and Implementation Points

4.1 Simulation Workflow

The implementation follows a structured computational workflow:

1. Geometry and Mesh Development: Create finite element models of the overlay configuration (single-bead, multi-pass, or full component) with appropriate mesh density (element size ≤2 mm in the weld zone, transitioning to 10–20 mm in the substrate). Adaptive mesh refinement is applied around the heat source.
2. Material Property Input: Define temperature-dependent properties for base metal and weld metal including elastic modulus, yield strength, thermal conductivity, specific heat, thermal expansion coefficient, and phase transformation parameters (Ms, Mf, volumetric expansion).
3. Heat Transfer Analysis: Perform sequential thermal analysis using a moving heat source model (Gaussian or double-ellipsoidal) to generate temperature histories at all integration points.
4. Phase Transformation Modeling: Calculate martensite volume fraction evolution at each node using the Koistinen-Marburger or Scheil equation, incorporating the effect of prior austenite grain size and cooling rate.
5. Mechanical Analysis: Perform elastic-plastic analysis using the elastic-plastic strain decomposition, with transformation strain superimposed as an eigenstrain source.
6. Result Post-Processing: Extract residual stress distributions (σ_xx, σ_yy, σ_zz), stress relaxation sequences, and transformation-induced stress components.

4.2 Key Parameters and Their Influence

Parameter	Typical Range	Influence on Transverse Stress	Optimization Direction
Heat Input (Q)	0.5–5.0 kJ/mm	Higher Q → lower cooling rate → reduced transformation fraction → more tensile stress	Minimize within weldability limits
Interpass Temperature	80–300°C	Higher interpass → reduced thermal gradient → lower peak stress but potential grain coarsening	Balance with grain control requirements
Travel Speed	30–200 mm/min	Higher speed → lower Q → faster cooling → more martensite → partially offset tensile stress	Coordinate with penetration requirements
Ms Temperature	200–600°C	Lower Ms → transformation at lower T → larger ΔT for thermal contraction before transformation → higher tensile component	Material selection consideration
ΔV/V (Volumetric Expansion)	2.0–5.5%	Higher expansion → greater compressive contribution → net stress reduction	Maximize through alloy design
Number of Layers	1–20 passes	More layers → stress relaxation in lower layers → surface layers carry higher residual stress	Design layer sequence for stress management
Preheat Temperature	0–400°C	Higher preheat → reduced thermal gradient → lower peak stress	Maximize within metallurgical limits
Backing Material	Steel/Insulation	Thermal backing → symmetric cooling → reduced transverse stress asymmetry	Use where feasible

4.3 Critical Implementation Considerations

The accuracy of the simulation depends on several critical factors that must be rigorously controlled:

• Material Property Accuracy: Temperature-dependent properties must be experimentally determined for the specific alloy system, not assumed from generic databases. The company's metallurgical laboratory supports this through dilatometry, thermomechanical testing (Gleeble), and X-ray diffraction phase analysis.
• Phase Transformation Kinetics: The Koistinen-Marburger equation assumes isothermal transformation; for continuous cooling, the Scheil equation or more advanced models (e.g., JMA-type) must be employed. The constant α must be calibrated against experimental dilatometry data for each alloy system.
• Yield Surface Evolution: The yield strength of the material changes dramatically during transformation (martensite is significantly harder than austenite). The simulation must update the yield surface at each time step based on the current phase fraction, using a rule of mixtures or more sophisticated approaches.
• Mesh Convergence: Results must be verified for mesh independence, particularly in the weld zone where stress gradients are steepest. A mesh convergence study with at least three refinement levels is mandatory.
• Boundary Condition Fidelity: The actual clamping conditions, backing support, and thermal boundary conditions during production welding must be accurately represented. Simplified boundary conditions can introduce errors of 20–40% in predicted stress magnitudes.

4.4 Validation Protocol

All simulation results must be validated against experimental measurements before being used for engineering decisions:

• X-ray Diffraction (XRD): Surface residual stress measurement using sin²ψ or cosα methods with spot spacing ≤5 mm across the overlay width and length.
• Hole-Drilling Method: Sub-surface stress measurement at critical depths (overlay/substrate interface, mid-overlay) per ASTM E837.
• Neutron Diffraction: Bulk stress measurement for thick components where surface methods are insufficient.
• Strain Gauge Measurement: Real-time stress monitoring during welding for thermal cycle and stress evolution verification.
• Target Accuracy: Predicted vs. measured stress deviation should be within ±15% for peak stresses and ±20% for stress gradients.

5. Applicable Standards and Acceptance Criteria

5.1 Governing Standards

Standard	Applicability	Key Requirements
ASME Section VIII Div. 2	Residual stress in pressure vessel welds	Allowable residual stress limits, PWHT requirements, stress relief criteria
ASME Section IX	WPS/PQR qualification for overlay welding	Essential variables, performance qualification, weld metal properties
NB/T 20912	Nuclear-grade clad plate weld overlay	Residual stress limits at bond interface, NDT requirements, PWHT specifications
NB/T 20913	Weld overlay for nuclear pressure vessels	Crack resistance testing, residual stress acceptance criteria
ASTM A388	Stainless steel clad plate specification	Clad material properties, bonding requirements, testing methods
ASTM E837	Hole-drilling strain gauge method	Residual stress measurement procedure, calibration requirements
ASTM E975	X-ray diffraction residual stress measurement	sin²ψ method, data reduction, accuracy requirements
GB/T 19420	Weld residual stress measurement (Chinese standard)	Measurement methods, acceptance criteria for welded structures
API 579	Fitness-for-service assessment	Residual stress factors in fracture assessment, stress intensity correction
NACE MR0175/ISO 15156	Sulfide stress cracking resistance	Hardness and residual stress limits for sour service
ISO 15156	Materials for H₂S-containing environments	Residual stress and hardness interaction for SSC resistance
GB/T 4675	Weld residual stress measurement by XRD	Chinese standard for XRD residual stress measurement procedures

5.2 Residual Stress Acceptance Criteria

Typical acceptance criteria for weld overlay residual stresses, informed by code requirements and the company's internal quality standards:

• Nuclear applications (NB/T 20912): Longitudinal and transverse residual stresses at the overlay/bond interface shall not exceed 50% of the yield strength of the clad material at service temperature. For 304/304L overlay on low-alloy steel, this typically translates to a maximum tensile residual stress of approximately 120–150 MPa.
• Pressure vessels (ASME VIII Div. 2): Residual stresses shall be relieved to within 50% of the material yield strength through PWHT, or shall be accounted for in the fracture assessment per API 579.
• Sour service (NACE MR0175): Combined effect of hardness and tensile residual stress shall be evaluated; if hardness exceeds 22 HRC, residual stress relief is mandatory regardless of absolute stress level.
• General industrial overlay: Peak transverse tensile stress should be limited to 200 MPa or less for overlay materials with fracture toughness below 50 MPa·m^½, and to 350 MPa for tougher materials.

5.3 Simulation Verification Acceptance

Before a simulation model is used for production engineering decisions, it must pass verification:

• Temperature field prediction accuracy: within ±15% of thermocouple measurements at corresponding locations
• Residual stress peak values: within ±15% of XRD or hole-drilling measurements
• Stress distribution shape (tensile/compressive regions): qualitatively matching experimental patterns
• Stress relaxation sequence between passes: correctly predicting layer-by-layer stress evolution

6. Common Risks and Controls

6.1 Technical Risks

Risk	Consequence	Control Measure
Inaccurate material property input (generic database data used instead of alloy-specific data)	Predicted stresses deviate >25% from actual; incorrect process recommendations	Mandatory experimental determination of temperature-dependent properties for each new alloy system; Gleeble thermomechanical testing
Oversimplified phase transformation model (ignoring bainite or pearlite formation)	Overprediction of martensite fraction; incorrect transformation stress	Dilatometry to determine actual transformation products; multi-phase transformation model implementation
Inadequate mesh resolution in weld zone	Artificial stress concentrations; non-convergent results	Mesh convergence study; minimum element size of 1 mm in weld nugget region
Incorrect boundary conditions (rigid clamping assumed when actual setup is flexible)	Overprediction of peak stresses by 30–50%	Accurate representation of production fixture and clamping; flexible boundary conditions with measured stiffness
Failure to account for stress relaxation during multi-pass welding	Incorrect prediction of final stress state; overconservative or under-conservative design	Sequential layer-by-layer analysis with proper constraint release between passes
Using room-temperature mechanical properties throughout the analysis	Grossly incorrect stress predictions (errors up to 100%)	Full temperature-dependent property implementation; property updates at every time increment

6.2 Quality Risks in Product Delivery

• Residual stress-induced cracking: If simulation predictions are not validated and process parameters are not optimized accordingly, transverse cracking may occur in martensitic overlay systems, leading to product rejection and customer downtime.
• Post-weld distortion: Unpredicted residual stress states lead to dimensional non-conformance after stress relief or during subsequent machining operations.
• Code compliance failure: Residual stresses exceeding code limits result in failed acceptance inspections, requiring costly rework (grinding, re-welding, or PWHT).
• Hydrogen-assisted cracking: High tensile residual stresses combined with hydrogen from welding significantly increase susceptibility to hydrogen-induced cracking (HIC) in susceptible microstructures.

6.3 Risk Mitigation Strategy

1. Mandatory simulation-to-experiment correlation: No simulation result shall be used for production decisions without validation against at least one physical weld coupon with measured residual stresses.
2. Conservative design margins: When simulation accuracy is uncertain, apply a safety factor of 1.3 on predicted peak tensile stresses for acceptance decisions.
3. Regular model recalibration: Update simulation models quarterly with new experimental data from production welds to maintain accuracy.
4. Independent review: All simulation studies supporting customer deliverables must be reviewed by a qualified metallurgical engineer independent of the original analysis.
5. Documentation and traceability: Complete simulation input data, software version, mesh files, and output results shall be archived for a minimum of 15 years (or per contract requirement).

7. Application Across Company Technology Routes

7.1 TIG/MIG Weld Overlay Applications

The numerical simulation capability is most directly applicable to TIG/MIG weld overlay, where the thermal cycle is well-defined and phase transformation effects are controllable through process parameters:

7.1.1 Multi-Pass Overlay Stress Management

For multi-pass overlay builds (typically 5–20 passes for thick overlay layers), the simulation enables:

• Prediction of stress build-up and relaxation sequence as each pass is deposited
• Identification of the critical pass number beyond which stress accumulation becomes detrimental
• Optimization of pass sequence (e.g., building from center outward vs. single-direction) to minimize peak stresses
• Determination of optimal interpass temperature to balance stress relief against grain coarsening risk

7.1.2 Martensitic Overlay Systems

For overlay systems involving martensitic materials (e.g., 17-4PH, 431, 420, 15-5PH, or martensitic stainless steels), the simulation is critical because:

• Martensitic transformation occurs at temperatures where the material is still significantly constrained by the substrate
• The volumetric expansion during transformation can either beneficially offset tensile stress or, if poorly managed, create complex stress states that promote cracking
• Post-weld heat treatment (PWHT) for aging or tempering will further alter the stress state, requiring prediction of the final stress state after all thermal cycles

7.1.3 Transition Layer Optimization

When overlaying dissimilar materials (e.g., austenitic stainless on low-alloy steel), the simulation helps optimize the transition layer composition and thickness to:

• Minimize the coefficient of thermal expansion mismatch between layers
• Control the transformation stress at the bond interface
• Ensure the transition layer itself does not develop unacceptable residual stresses that compromise bonding

7.1.4 Specific Process Applications

• GTAW overlay of 309L/310L transition layers: Simulation confirms that austenitic overlay metals with Ms below room temperature develop minimal transformation stress, but high thermal contraction stress due to low modulus at elevated temperatures.
• GMAW overlay of 625/626 nickel alloys: Modeling the transformation behavior of precipitation-strengthened nickel alloys, where the γ' precipitation during PWHT introduces additional volumetric changes.
• Multi-wire GMAW for thick overlay builds: Simulation of high-deposition-rate processes where thermal mass effects and rapid cooling create complex transformation scenarios.

7.2 Hydraulic Explosive Bonding Applications

While hydraulic explosive bonding is a solid-state process that does not involve melting, the numerical simulation capability contributes to:

7.2.1 Post-Bond Stress Analysis

• Prediction of residual stresses introduced during the hydraulic explosive bonding process itself (from plastic deformation of the bonding wave)
• Assessment of stress state after differential thermal expansion during subsequent cooling from process temperature (if any)
• Analysis of stress state after post-bond heat treatment for stress relief

7.2.2 Bond Interface Integrity

The simulation helps establish the maximum allowable residual stress at the bond interface that will not compromise metallurgical bonding, supporting:

• Design of bond interface geometry to minimize stress concentration
• Prediction of interface stress during service thermal cycling
• Verification that the bond interface stress state remains within acceptable limits for the specific application

7.2.3 Clad Plate Stress State for Subsequent Welding

When hydraulic explosively bonded clad plates are subsequently welded (e.g., for pipe fabrication or vessel construction), the pre-existing stress state from the bonding process must be accounted for in the welding simulation. The company uses the residual stress prediction from bonding as boundary conditions for the subsequent weld simulation.

7.3 Explosion Welding Applications

Explosion welding involves extreme thermomechanical loading, and the simulation capability supports:

7.3.1 Residual Stress Prediction in Explosion-Welded Clad Plates

• Prediction of the complex residual stress field in explosion-welded clad plates resulting from the combined effects of explosive impact loading, plastic deformation, and post-impact cooling
• Determination of stress relief requirements to meet code acceptance criteria (typically 50% of yield strength or less)
• Optimization of post-weld heat treatment parameters (temperature, duration, cooling rate) to achieve target stress relief without compromising bond integrity

7.3.2 Explosion-Welded Pipe Stress Analysis

For explosion-welded clad pipes (a critical product for oil and gas service), the simulation addresses:

• Axial and hoop stress distribution around the pipe circumference
• Stress concentration at the explosion seam and weld repair locations
• Prediction of stress state after subsequent welding operations (end welds, repair welds)
• Verification of compliance with API 5L, API 5CT, or customer specifications for residual stress limits

7.3.3 Multi-Layer Explosion Cladding

For multi-layer explosion cladding (e.g., nickel layer + stainless layer on carbon steel), the simulation predicts:

• Cumulative residual stress from sequential explosion events
• Interfacial stress between layers that could compromise bonding
• Required stress relief between explosion events (if applicable)

8. Contribution to Qualification Building

8.1 WPS Qualification Acceleration

The numerical simulation capability significantly accelerates welding procedure qualification by:

1. Reducing the number of physical trial welds: Instead of testing 5–10 parameter combinations experimentally, simulation narrows the search space to 2–3 candidates, reducing qualification time by 40–60%.
2.Providing engineering justification for parameter selections: When ASME Section IX or NB/T 20913 requires justification for non-standard parameters, simulation provides the quantitative basis. 3.Predicting weld metal properties: By modeling the thermal cycle accurately, the simulation predicts cooling rates and resulting microstructure, supporting mechanical property qualification without extensive coupon testing. 4.Supporting performance qualification: For fracture toughness or fatigue performance qualification, simulation identifies the most critical stress states and weld locations for testing.

8.2 Certification System Integration

The simulation capability supports the company's quality management system (ISO 9001, ISO 3834, ASME QME-1) by:

• Providing documented engineering analysis as part of the WPS justification file
• Supporting NDT procedure development by predicting defect susceptibility zones
• Enabling systematic process control through defined parameter windows with predicted stress outcomes
• Fulfilling traceability requirements for all engineering decisions affecting product quality

8.3 Customer Qualification Support

For customer-specific qualification programs, the simulation provides:

• Residual stress prediction reports as part of qualification documentation packages
• Engineering analysis supporting deviation requests from standard specifications
• Life assessment data (combined with residual stress predictions) for fitness-for-service evaluations
• Design optimization recommendations that reduce qualification risk

9. Implementation Roadmap and Continuous Improvement

9.1 Current Capability Level

The company's current simulation capability includes:

• 2D and 3D sequential thermomechanical analysis using established FEA software
• Temperature-dependent material property databases for common overlay alloys (304, 309L, 310L, 625, 626, 410, 420, 17-4PH)
• Koistinen-Marburger transformation model with Scheil extension for continuous cooling
• Validation against XRD and hole-drilling measurements for at least 3 alloy systems
• Integration with WPS development workflow for new overlay specifications

9.2 Planned Enhancements

1. Coupled phase transformation kinetics: Implementation of advanced models (e.g., JMA, Avrami) for bainite and pearlite transformation alongside martensite, enabling simulation of hypoeutectoid and eutectoid systems.
2.Microstructure-sensitive modeling: Integration of prior austenite grain size effects on transformation temperature and transformation stress magnitude. 3.Multi-scale modeling: Coupling of continuum FEA with crystal plasticity models for localized stress prediction at grain boundaries and phase interfaces. 4.Machine learning acceleration: Development of surrogate models trained on FEA results for real-time stress prediction during process optimization. 5.Full-process simulation: Extension to include welding-induced distortion prediction, cutting and machining residual stresses, and PWHT stress relief in a single integrated workflow.

9.3 Knowledge Transfer and Standardization

The learning and documentation of this simulation capability contributes to organizational knowledge through:

• Development of internal simulation procedures and checklists for consistent application
• Training programs for welding engineers and metallurgists on simulation interpretation
• Creation of alloy-specific simulation templates that reduce setup time for new projects
• Contribution to industry standards development through participation in relevant committees (e.g., ASME BPV, ISO/TC 44)
• Publishing of technical papers and case studies that establish the company's technical authority in the field

10. Conclusion

The numerical simulation of martensitic transformation effects on cooling transverse stress in weld overlay represents a sophisticated computational capability that directly supports the company's product quality, qualification efficiency, and customer value proposition. By enabling accurate prediction and control of residual stress states in weld overlay deposits, this capability reduces the risk of cracking, spalling, and dimensional non-conformance while accelerating the development and qualification of new overlay specifications.

Across the company's three technology routes, the simulation capability provides differentiated value: direct process optimization for TIG/MIG weld overlay, bond interface stress verification for hydraulic explosive bonding, and post-process stress relief design for explosion welding. The integration of this analytical capability with the company's experimental metallurgy, NDT, and quality management systems creates a comprehensive technical framework that supports code compliance, customer confidence, and competitive differentiation in the cladding and overlay manufacturing market.

The ongoing investment in simulation capability enhancement—through advanced transformation models, multi-scale approaches, and machine learning acceleration—ensures that the company maintains technical leadership in computational welding metallurgy, positioning it as a preferred partner for critical-service cladding applications where residual stress control is paramount.