Thin Plate Weld Overlay Deformation Prediction Based on Initial Deformation

Literature Overview

This 2020 study by Guo Nan, Ma Xiqiang, Yin Xianqing, and Yu Yongjian—published in China Mechanical Engineering—addresses a fundamental challenge in weld overlay fabrication: the prediction and control of welding distortion in thin plate cladding applications. The research was supported by multiple Henan Provincial research grants, reflecting its significance to the region's mechanical manufacturing industry.

The authors propose a deformation prediction methodology that incorporates initial (pre-existing) plate deformation as a boundary condition, rather than assuming a perfectly flat starting geometry. This approach represents a significant advancement over conventional welding distortion models that typically assume ideal initial conditions and therefore fail to predict actual production outcomes accurately.

Technical Motivation and Problem Statement

In practice, clad plates and weld overlay substrates are rarely perfectly flat before welding begins. Sources of initial deformation include:

• Rolling mill shape errors (crown, camber, edge wave)
• Residual stress relaxation during storage and handling
• Thermal exposure during prior processing operations
• Mechanical damage during transportation and fabrication
• Residual warpage from prior welding operations on the same plate

Conventional welding distortion prediction methods—whether analytical (such as the modified beam theory approach) or numerical (finite element simulation)—typically assume a flat, stress-free initial plate. When the actual plate contains initial curvature or out-of-flatness, the predicted welding distortion can deviate significantly from measured values, sometimes by 30-50%.

The research team at Henan University of Science and Technology, working with collaborators at Xi'an Jiaotong University, developed a methodology that explicitly accounts for initial plate deformation as an input parameter to the welding distortion prediction model.

Methodology and Technical Approach

The study employs a combined analytical-numerical approach that integrates:

1. Initial deformation measurement: Using coordinate measuring machines (CMM) or laser scanning to characterize the pre-weld plate geometry with spatial resolution of 0.1-0.5 mm.
2. Finite element modeling: Creating a three-dimensional shell element model that incorporates the measured initial deformation as geometric initial conditions.
3. Thermo-mechanical coupling: Implementing a sequential or fully coupled thermal-structural analysis that accounts for:

• Heat input distribution from the welding process
• Temperature-dependent material properties
• Phase transformation effects (for martensitic substrates)
• Plastic strain accumulation and stress relaxation

1. Welding sequence optimization: Using the prediction model to evaluate different welding sequences (continuous, skip, step-back, zigzag) and identify the sequence that minimizes final distortion.

Key Technical Parameters

Analysis of Results and Key Findings

The study demonstrates that incorporating initial deformation into the prediction model significantly improves accuracy:

• Without initial deformation input: Average prediction error of 25-40% compared to measured values
• With initial deformation input: Average prediction error reduced to 5-12% compared to measured values

The research identifies several important relationships:

1. Phase relationship matters: When the initial deformation is in phase with the welding-induced deformation (both causing upward curling of the plate edge), the combined effect is additive, resulting in total distortion that exceeds predictions from either source alone.
2. Anti-phase cancellation: When initial deformation is anti-phase to welding-induced deformation (initial convexity at the weld line that will be reduced by welding), partial cancellation occurs, and the final distortion may be less than predicted from welding alone.
3. Non-linear amplification: The interaction between initial deformation and welding distortion is non-linear. Small initial deformations (below 0.5 mm/m) have negligible effect on final distortion, while larger initial deformations (above 2 mm/m) can amplify final distortion by factors of 1.5-3.0.
4. Sequence sensitivity: The effect of initial deformation on final distortion depends on the welding sequence. Continuous welding amplifies the interaction effect, while skip or step-back sequences partially decouple the initial deformation from the welding-induced deformation field.

Engineering Practice Applications

The practical applications of this research extend to several manufacturing scenarios:

Clad Plate Fabrication

For stainless steel/carbon steel clad plates produced by submerged arc welding overlay, the prediction model enables:

• Selection of substrate plates with acceptable initial flatness for specific welding sequences
• Optimization of welding parameters to compensate for measured initial deformation
• Reduction of post-weld leveling operations (flat rolling or shot peening)
• Improved yield rates by predicting which plates will meet flatness specifications after welding

Pressure Vessel Fabrication

For clad pressure vessel shells and heads, the methodology supports:

• Assessment of whether initial shell curvature (from forming operations) will be acceptable after overlay welding
• Prediction of distortion effects on joint fit-up and alignment
• Optimization of weld sequence for large vessel segments to minimize geometric deviations from design specifications

Thin-Wall Heat Exchanger Fabrication

For overlay-clad heat exchanger tubesheets and channel covers:

• Prediction of distortion effects on hole alignment and tube insertion
• Assessment of overlay layer thickness uniformity after distortion
• Selection of appropriate back-up fixtures and clamping strategies

Quality Control Integration

The study's methodology integrates naturally with modern quality control systems:

• Pre-weld inspection: CMM scanning of substrate plates to establish initial deformation baseline
• In-process monitoring: Comparison of measured thermal cycles with predicted values to validate model assumptions
• Post-weld verification: Comparison of predicted and measured final distortion to calibrate and improve future predictions
• Process documentation: Recording of initial deformation data as part of the welding procedure qualification record

The approach aligns with the PDCA (Plan-Do-Check-Act) quality management cycle:

• Plan: Measure initial deformation and run prediction model
• Do: Execute welding according to optimized sequence and parameters
• Check: Measure final distortion and compare with prediction
• Act: Update model parameters and refine predictions for subsequent production

Key Technical Challenges and Limitations

Several limitations of the proposed methodology deserve recognition:

1. Measurement accuracy: The quality of prediction depends critically on the accuracy of initial deformation measurement. Surface roughness, coating thickness, and magnetic properties can affect measurement precision.
2. Model complexity: Fully coupled thermo-mechanical finite element models require significant computational resources and expert knowledge to set up correctly. Simplified models may not capture all physical phenomena.
3. Material model accuracy: The constitutive models used to describe temperature-dependent material behavior must be validated against experimental data for the specific materials used.
4. Boundary condition representation: The actual clamping and fixture conditions during welding are difficult to represent accurately in numerical models, and simplifications may introduce prediction errors.
5. Multi-pass interactions: For multi-pass overlay welding, the interaction between successive passes and the evolving stress field adds complexity to the prediction model.

Study Insights and Forward Looking Perspectives

The fundamental insight of this research is that welding distortion prediction must account for the real-world initial conditions of the workpiece, not just the idealized conditions assumed in conventional models. This seemingly simple observation has profound implications for manufacturing quality and process optimization.

For engineers involved in clad plate and weld overlay fabrication, the practical recommendations are:

• Always measure and document initial plate deformation before welding
• Use prediction models that incorporate initial conditions for critical applications
• Develop tolerance criteria that account for the interaction between initial and welding-induced deformation
• Invest in measurement capabilities (CMM, laser scanning) that provide adequate spatial resolution for deformation characterization
• Build prediction models into the production workflow as a standard quality assurance tool

The research represents a mature approach to welding distortion management that bridges the gap between theoretical prediction and manufacturing reality. As manufacturing industries move toward digital transformation and process optimization, such integrated prediction methods will become increasingly essential for achieving consistent quality in weld overlay fabrication.