Hand-Eye System-Based Weld Overlay Repair Trajectory Generation for Edge-Cutting Dies
1. Definition and Fundamental Principles
The hand-eye system-based weld overlay repair trajectory generation method is an advanced robotic automation technique that integrates a vision-guided sensing subsystem (the "eye") with a multi-axis robotic welding manipulator (the "hand") to autonomously generate precise, adaptive weld overlay repair paths on edge-cutting dies. This methodology addresses the inherent challenges of repairing worn, chipped, or deformed cutting edges on industrial dies—components critical to metal forming, stamping, and sheet processing operations—by combining real-time visual inspection, geometric modeling, and intelligent path planning into a unified workflow.
The core principle relies on a hand-eye calibration framework, wherein the coordinate transformation between the camera's optical coordinate system and the robot's base coordinate system is mathematically established. This calibration enables the vision system to accurately map detected surface features—such as wear profiles, edge geometry deviations, and damage boundaries—into the robot's working frame, thereby generating executable welding trajectories that conform precisely to the required repair geometry.
1.1 Hand-Eye Calibration Fundamentals
Hand-eye calibration solves the equation AX = XB, where A represents the transformation from the robot base frame to the end-effector frame, X represents the unknown hand-eye transformation matrix, and B represents the transformation from the end-effector frame to the camera frame. Two primary configurations exist:
- Eye-in-Hand (EIH): The camera is mounted on the robot's end-effector. This configuration provides a dynamic field of view that moves with the welding torch, offering superior resolution for localized edge inspection but requiring continuous recalibration as the robot moves.
- Eye-to-Hand (ETH): The camera is mounted on a fixed external position. This configuration provides a stable global reference frame, simplifying calibration and enabling comprehensive pre-inspection of the entire die surface before the robot initiates welding.
1.2 Trajectory Generation Logic
The trajectory generation process follows a structured pipeline:
- Surface Acquisition: Structured light scanning or stereo vision captures the three-dimensional topography of the die cutting edge, producing a high-resolution point cloud dataset with sub-millimeter accuracy.
- Feature Extraction and Defect Identification: Point cloud processing algorithms segment the nominal geometry from actual geometry, identifying wear zones, chipping areas, and edge radius deviations that require overlay repair.
- Repair Volume Modeling: The geometric difference between the target nominal profile and the measured worn profile is computed to determine the volumetric material deposition required at each spatial location along the cutting edge.
- Weld Pass Planning: The repair volume is decomposed into individual weld passes, each optimized for the selected welding process (TIG or MIG) in terms of bead width, bead height, and interpass geometry to ensure adequate fusion and avoid excessive dilution or undercut.
- Robotic Path Synthesis: The weld passes are converted into continuous robotic motion trajectories with appropriate torch travel speed, torch angle, and oscillation parameters, incorporating collision avoidance and thermal management constraints.
2. Category and Business Positioning
This technology entry occupies a strategic position at the intersection of intelligent manufacturing automation and specialized cladding/overlay repair services. Within Cladding Technology Shanxi Co., Ltd.'s capability portfolio, it represents a significant advancement in the company's automation and digitalization maturity, transitioning from manual or semi-automated overlay repair to fully vision-guided robotic systems.
2.1 Positioning Within the Company's Technology Ecosystem
| Dimension | Positioning |
|---|---|
| Technology Maturity | Advanced / Intelligent Manufacturing Tier |
| Core Competency | Robotic Automation, Machine Vision, Process Intelligence |
| Value Chain Role | Upstream process design and trajectory optimization for overlay repair |
| Customer Segment | Heavy industry, automotive stamping, aerospace forming, mining equipment |
| Competitive Advantage | Repeatability, traceability, and precision unattainable through manual methods |
2.2 Strategic Business Value
Edge-cutting dies represent high-value assets in manufacturing operations, with single die replacement costs frequently exceeding USD 50,000–200,000 depending on complexity and material. The integration of hand-eye guided robotic overlay repair transforms the economics of die maintenance by:
- Reducing die replacement frequency through precise, repeatable repair that restores original cutting edge geometry.
- Minimizing material consumption through computed deposition volumes rather than operator-dependent estimation.
- Enabling rapid turnaround with automated pre-inspection, trajectory generation, and post-weld verification.
- Generating digital repair records that support predictive maintenance and asset lifecycle management.
3. Technical Purpose and Engineering Value
3.1 Primary Technical Objectives
The hand-eye system-based trajectory generation method serves several critical technical objectives in the context of edge-cutting die overlay repair:
- Geometric Fidelity Restoration: Achieve cutting edge profile restoration to within ±0.05 mm of the original design specification, ensuring consistent forming quality and tool life extension.
- Material Optimization: Deposition of overlay alloy material limited to the minimum volume required for geometric restoration, reducing material cost and thermal distortion.
- Process Consistency: Elimination of operator variability through algorithmic trajectory generation, ensuring identical weld parameters and bead geometry regardless of shift or operator assignment.
- Complex Geometry Adaptation: Capability to handle multi-axis cutting edges with varying radii, compound angles, and non-planar surfaces without manual programming intervention.
3.2 Quantifiable Engineering Value
| Performance Metric | Manual/Conventional Method | Hand-Eye Robotic Method | Improvement |
|---|---|---|---|
| Edge profile accuracy | ±0.15–0.30 mm | ±0.03–0.05 mm | 60–80% improvement |
| Repair cycle time (per die) | 8–16 hours | 2–5 hours | 60–75% reduction |
| Overlay material utilization | 40–55% | 75–85% | 40–50% improvement |
| Post-weld machining requirement | Extensive grinding | Minimal finishing | 50–70% reduction |
| Batch consistency (Cpk) | 0.8–1.0 | 1.33–1.67 | Statistical process control compliant |
4. Key Process and Implementation Points
4.1 System Architecture and Hardware Configuration
A production-grade hand-eye welding system for edge-cutting die repair comprises the following integrated subsystems:
- Industrial Robot: Six-axis articulated robot with payload capacity of 200–600 kg, repeatability of ±0.03 mm, and minimum link length sufficient to access all die edge regions without collision.
- Structured Light 3D Scanner: Laser triangulation or fringe projection scanner with point cloud resolution of 0.02–0.05 mm and measurement accuracy of ±0.03 mm for pre-weld surface acquisition.
- Infrared Thermal Camera: Real-time monitoring of weld pool temperature distribution for process feedback and thermal distortion prediction.
- Welding Power Source: TIG or MIG power supply with digital control interface supporting trajectory-synchronized parameter adjustment.
- Industrial Computer and PLC: High-performance computing unit running vision processing software, trajectory generation algorithms, and robot control interface.
4.2 Hand-Eye Calibration Procedure
Accurate hand-eye calibration is the foundation of the entire system. The calibration procedure follows these critical steps:
- Calibration Target Deployment: A high-precision calibration board with known fiducial marker patterns (minimum 12 markers with sub-millimeter manufacturing tolerance) is positioned at multiple poses relative to the robot workspace.
- Data Acquisition: The robot moves to at least 15 different poses while the camera captures the calibration target. Each pose records both the robot's kinematic state (joint angles, end-effector pose) and the camera's extracted marker positions.
- Mathematical Solution: The AX = XB equation is solved using Tsai-Lenz or Park-Martin methods, yielding the hand-eye transformation matrix with angular error less than 0.5° and translational error less than 0.1 mm.
- Verification and Validation: The calibrated system is verified by commanding the robot to touch known reference points on the calibration target, confirming end-to-end accuracy within specification.
4.3 Trajectory Generation Algorithm Details
4.3.1 Point Cloud Processing Pipeline
| Processing Stage | Algorithm/Method | Output | Quality Criterion |
|---|---|---|---|
| Raw Point Cloud Denoising | Statistical outlier removal + Moving Least Squares (MLS) | Clean point cloud | Outlier ratio < 0.5% |
| Surface Reconstruction | Poisson Surface Reconstruction or Ball Pivoting Algorithm | Watertight mesh model | No holes, smooth normals |
| Nominal Geometry Registration | Iterative Closest Point (ICP) alignment to CAD model | Aligned coordinate system | Registration error < 0.1 mm |
| Deviation Mapping | Color-coded height deviation field computation | Repair volume map | Resolution ≤ 0.05 mm |
| Weld Pass Decomposition | Level-set segmentation + offset contour generation | Ordered pass sequence | Uniform pass spacing |
4.3.2 Weld Pass Planning Parameters
For TIG weld overlay repair of edge-cutting dies, the following parameter ranges are typically applied, adjusted based on the specific overlay alloy and base material:
| Parameter | Typical Range | Notes |
|---|---|---|
| Welding current | 80–200 A | Adjusted per pass based on deposition volume |
| Travel speed | 200–600 mm/min | Correlated with current and wire feed rate |
| Shielding gas flow | 8–15 L/min (Ar or Ar/He mix) | Higher for outdoor or draft-affected environments |
| Interpass temperature | ≤ 150°C (general); ≤ 80°C (heat-sensitive) | Monitored via IR camera feedback |
| Torch angle | 10–20° from vertical | Forward or backward depending on repair geometry |
| Weld bead width | 3–8 mm | Determined by current, speed, and oscillation amplitude |
| Weld bead height | 1.0–3.0 mm per pass | Limited by fluidity and undercut prevention |
4.4 Real-Time Process Monitoring and Adaptation
The hand-eye system incorporates closed-loop feedback mechanisms to maintain process stability during multi-pass overlay welding:
- Weld Pool Tracking: Infrared or visible-light weld pool monitoring provides real-time position feedback, enabling automatic trajectory correction for drift or thermal distortion.
- Bead Width Monitoring: Post-pass visual inspection of completed weld beads validates actual bead geometry against planned dimensions, triggering parameter adjustment for subsequent passes.
- Thermal Management: Infrared thermography monitors base plate temperature distribution, activating cooling interventions (forced air, water spray) when interpass temperature limits are approached.
- Spark and Arc Monitoring: Acoustic and optical sensors detect arc instabilities, enabling immediate process interruption and fault diagnosis.
5. Applicable Standards and Acceptance Criteria
5.1 Welding Procedure Standards
| Standard | Scope of Application | Key Requirement |
|---|---|---|
| ASME Section IX | Welding procedure qualification for pressure-containing components | WPQ/WPQR documentation, essential variables control |
| ASME BPV Code Section II Part D | Welding procedure qualification for boiler and pressure vessel | Procedure qualification record (PQR) and welding procedure specification (WPS) |
| ASTM A240 / A240M | Stainless steel plate/sheet material specification | Material composition and mechanical properties for overlay alloys |
| GB/T 985.1 | Welding symbol marking on technical drawings | Standardized weld documentation for repair specifications |
| GB/T 3375 | Welding terminology | Unified terminology for overlay welding documentation |
| NB/T 47014 | Qualification test of welding procedure for pressure vessels | Chinese national standard for WPS qualification in pressure vessel industry |
| ISO 15614-1 | Qualification testing of welding procedures for metallic materials | International standard for welding procedure qualification |
| ISO 3834-2 | Quality requirements for fusion-welding of metallic materials | Comprehensive quality management requirements for welding operations |
| API 16F | Specification for weld overlay of hardfacing | Weld overlay qualification for hardfacing applications |
| NACE MR0175 / ISO 15156 | Materials for use in H2S-containing environments | Hardenability and hardness limits for overlay alloys in sour service |
5.2 Non-Destructive Testing (NDT) Acceptance Criteria
Post-overlay repair NDT inspection follows these acceptance criteria:
- Visual Testing (VT) per ASME BPV Code Section V Article 4: No cracks, undercut deeper than 0.1 mm or 10% of weld thickness (whichever is less), porosity exceeding 0.25 mm in diameter or 1% of weld area, or surface irregularities exceeding 0.2 mm.
- Magnetic Particle Testing (MT) per ASME BPV Code Section V Article 7: No indications of linear defects (cracks, laps, folds) of any length. Round indications limited to 3 mm length or diameter.
- Penetrant Testing (PT) per ASME BPV Code Section V Article 6: Applied to non-magnetic overlay alloys (stainless steel, nickel-based). Acceptance criteria equivalent to MT.
- Hardness Testing per ASTM E18/E92: Overlay hardness within specified range (typically HV 300–600 for wear-resistant overlays). Maximum hardness in heat-affected zone (HAZ) ≤ 400 HV per NACE MR0175 for sour service applications.
- Dimensional Verification: Cutting edge geometry verified via 3D scanning against original CAD model with tolerance of ±0.05 mm for critical forming surfaces and ±0.10 mm for non-critical surfaces.
5.3 Quality Management System Standards
- ISO 9001:2015: Quality management system certification ensuring systematic control of repair processes, documentation, and continuous improvement.
- ISO 3834-2: Comprehensive welding quality requirements including personnel qualification, equipment maintenance, and process monitoring.
- ASME "Q" / "U" Stamp: For pressure-containing component repair, demonstrating organizational capability and procedural compliance.
- ISO 14731: Welding personnel qualification and certification ensuring operator competency for robotic welding systems.
6. Common Risks and Controls
6.1 Technical Risks
| Risk Category | Description | Control Measures |
|---|---|---|
| Calibration Drift | Hand-eye transformation matrix degrades over time due to mechanical wear, thermal expansion, or camera mount loosening | Implement daily calibration verification using reference targets; automate drift detection through periodic touch-test sequences; establish calibration interval based on usage frequency (typically every 200–500 operating hours) |
| Point Cloud Noise | Surface reflectivity variations, oxide scales, or coolant residue produce erroneous scan data leading to incorrect repair volume computation | Apply multi-scan fusion from multiple viewpoints; implement statistical filtering and outlier rejection; require manual surface preparation (grinding, cleaning) before scanning; use structured light scanners with adaptive exposure control |
| Thermal Distortion | Cumulative heat input during multi-pass overlay welding causes die body distortion, particularly in thin-walled or high-carbon steel dies | Implement interpass temperature monitoring with automatic cooling; use low-heat-input welding parameters (shorter arc, lower current); employ sequential welding strategy (welding in sequence from least constrained to most constrained areas); apply back-of-plate cooling |
| Weld Defects | Porosity, lack of fusion, cracking, or excessive dilution in overlay welds | Strict WPS adherence; pre-qualification testing per ASME Section IX; real-time arc monitoring; post-weld NDT per applicable code; overlay alloy selection based on dilution calculation |
| Collision Risk | Robot arm or torch collides with die body, fixture, or surrounding equipment during trajectory execution | Implement 3D collision detection software with virtual environment modeling; maintain minimum safety clearance of 100 mm; perform dry-run trajectory validation before actual welding; install physical safety barriers and light curtains |
| Alloy Segregation | Uneven distribution of overlay alloy elements due to rapid solidification or improper mixing | Select alloy compositions with appropriate solidification range; control cooling rate through interpass temperature management; verify microstructure via metallographic examination for critical applications |
6.2 Operational Risks
- Operator Over-Reliance on Automation: Risk of accepting erroneous trajectory plans without human review. Control: Implement mandatory human-in-the-loop verification of generated trajectories before execution, particularly for first-time die geometries or critical repair zones.
- Software Version Inconsistency: Different software versions may produce divergent trajectory outputs. Control: Establish version control protocols; maintain regression test suites for trajectory generation algorithms; document software configuration for each repair job.
- Data Integrity and Traceability: Loss or corruption of scan data, trajectory files, or process records. Control: Implement redundant data storage; establish digital signature protocols for process records; maintain complete audit trails per ISO 9001 documentation requirements.
7. Application Scenarios Across the Company's Three Technology Routes
7.1 TIG/MIG Weld Overlay Integration
The hand-eye system-based trajectory generation method serves as the intelligent programming backbone for the company's TIG and MIG weld overlay operations, particularly in the following scenarios:
- Automotive Stamping Die Edge Repair: High-volume production stamping dies with complex multi-radius cutting edges requiring precise overlay restoration. The hand-eye system generates multi-pass TIG overlay trajectories using 309L/316L stainless steel or Ni-based overlay alloys, restoring cutting edge geometry to ±0.03 mm tolerance for production resumption.
- Aerospace Forging Die Maintenance: Critical forging dies where overlay repair must comply with stringent aerospace specifications (AMS, NAS). The system generates trajectories incorporating preheat, interpass temperature control, and post-weld heat treatment sequences as part of the trajectory program.
- Small-Batch Custom Die Repair: For unique or one-off die geometries where traditional manual programming is impractical, the hand-eye system enables rapid trajectory generation from scanned data, reducing programming time from days to hours.
- Multi-Layer Overlay Build-Up: For dies requiring substantial material addition (exceeding 5 mm), the system generates sequential layer-by-layer trajectories with automatic parameter adjustment for each layer, ensuring consistent weld quality throughout the build-up.
7.2 Hydraulic Explosive Bonding Integration
While hydraulic explosive bonding is a solid-state joining process that does not involve welding, the hand-eye system contributes to this technology route in several critical ways:
- Pre-Bonding Surface Preparation Verification: The vision system inspects and maps the surface condition of both base and cladding layers prior to hydraulic explosive bonding, ensuring surface cleanliness, flatness, and dimensional accuracy meet bonding requirements. Deviations are documented and addressed before bonding proceeds.
- Post-Bonding Interface Quality Assessment: After hydraulic explosive bonding, the hand-eye system performs ultrasonic scan integration and visual inspection of bonded joints, generating trajectory plans for any localized repair welding required at debonded or defective regions.
- Clad Edge Machining and Finishing: Following hydraulic explosive bonding, the hand-eye system generates robotic trajectories for precision machining of clad plate edges, ensuring dimensional accuracy and surface finish specification compliance for subsequent assembly.
- Transition Layer Weld Planning: When hydraulic explosively bonded clad plates require additional transition layer welding (e.g., 309L transition layer between dissimilar metals), the hand-eye system generates optimized TIG trajectories for the transition layer, ensuring proper fusion and mechanical compatibility.
7.3 Explosion Welding Integration
The hand-eye system enhances the explosion welding process route through the following application scenarios:
- Explosion Welding Panel Edge Inspection: Post-explosion welding inspection of panel edges and corners, identifying any incomplete bonding zones that require supplementary weld overlay repair. The system generates precise repair trajectories for these localized areas.
- Explosion Welded Pipe Component Repair: For explosion-welded clad pipes with localized damage or wear on the cladding surface, the hand-eye system enables robotic overlay repair of the cladding layer, restoring corrosion resistance and wear protection.
- Large-Format Clad Plate Processing: The hand-eye system guides robotic cutting, beveling, and machining operations on large explosion-welded clad plates, ensuring precise dimensional control and preservation of the cladding layer integrity during fabrication.
- Hybrid Clad Structure Assembly: For complex structures combining explosion-welded clad plates with welded joints, the hand-eye system generates welding trajectories that account for the thermal and mechanical behavior of the clad interface, preventing damage to the explosion-welded bond during subsequent welding operations.
8. Qualification Building and Customer Value
8.1 Qualification and Certification Advantages
The hand-eye system-based trajectory generation capability significantly strengthens the company's qualification position in the following areas:
- Welding Procedure Qualification (WPS/PQR): The system enables systematic qualification of welding procedures across a broader range of essential variables (travel speed, torch angle, current, voltage) with documented repeatability data, satisfying ASME Section IX and NB/T 47014 requirements for advanced qualification.
- ISO 3834-2 Certification: The automation and traceability provided by the hand-eye system demonstrate compliance with comprehensive quality requirements, supporting certification at the highest level (Level 2).
- Digital Manufacturing Certification: The capability supports qualification under emerging digital manufacturing standards and customer-specific requirements for Industry 4.0 readiness, including data traceability, process digitalization, and predictive quality assurance.
- Customer-Specific Qualification: Many OEM customers (automotive, aerospace, energy) require demonstration of automated welding capability with documented accuracy and repeatability. The hand-eye system provides the data foundation for such qualification programs.
8.2 Customer Value Proposition
The integration of hand-eye guided robotic overlay repair delivers measurable customer value across multiple dimensions:
| Value Dimension | Customer Benefit | Quantifiable Impact |
|---|---|---|
| Cost Reduction | Extended die service life through precise repair vs. frequent replacement | 30–60% reduction in die lifecycle cost |
| Downtime Minimization | Rapid repair turnaround reducing production line idle time | 50–75% reduction in repair cycle time |
| Quality Consistency | Uniform repair quality across all die units, reducing production variability | Cpk improvement from 0.8 to >1.33 |
| Sustainability | Material and energy savings through optimized deposition and reduced waste | 40% reduction in overlay material consumption; 30% reduction in post-weld machining energy |
| Digital Traceability | Complete digital repair records for asset management and predictive maintenance | 100% traceability from inspection to final verification |
8.3 Strategic Roadmap and Future Development
The hand-eye system-based trajectory generation method represents a foundation for continued technological advancement in the company's intelligent manufacturing capabilities. Future development priorities include:
- Artificial Intelligence Integration: Incorporation of machine learning algorithms for adaptive trajectory optimization based on historical repair data, enabling predictive parameter selection and automated quality prediction.
- Multi-Sensor Fusion: Integration of additional sensing modalities (laser displacement, ultrasonic thickness, electromagnetic sensing) for comprehensive real-time process monitoring and adaptive control.
- Digital Twin Development: Creation of virtual replicas of die components that simulate welding thermal and mechanical effects, enabling pre-weld distortion prediction and trajectory pre-compensation.
- Cloud-Based Process Intelligence: Development of cloud-connected platforms for fleet-wide process data aggregation, enabling cross-customer benchmarking, continuous improvement, and remote expert support.
- Cross-Route Integration: Extension of the hand-eye system's trajectory generation capabilities to support all three technology routes (TIG/MIG, hydraulic explosive bonding, explosion welding) within a unified digital manufacturing platform.
9. Conclusion
The hand-eye system-based weld overlay repair trajectory generation method for edge-cutting dies represents a transformative advancement in the company's intelligent manufacturing capabilities. By integrating machine vision, robotic automation, and intelligent trajectory planning into a unified system, this technology addresses the critical industry challenges of die maintenance cost, repair quality variability, and production downtime. The methodology is directly applicable across all three of the company's core technology routes—TIG/MIG weld overlay, hydraulic explosive bonding, and explosion welding—serving as a unifying intelligent automation layer that enhances process precision, quality assurance, and customer value delivery. Through rigorous adherence to international standards (ASME, ASTM, ISO, NB, GB, API, NACE), systematic qualification building, and continuous technological evolution, this capability positions the company at the forefront of advanced cladding and overlay repair manufacturing.