Machine Vision-Based Robotic Weld Overlay System for Roller Press Roll Sleeves
1. Definition and Technical Principles
A machine vision-based robotic weld overlay system for roller press roll sleeves is an advanced automated welding architecture that integrates real-time optical sensing, adaptive path planning, and multi-axis robotic motion control to deposit protective or functional cladding layers onto the cylindrical surfaces of roller press rolls. The system employs structured-light or laser triangulation sensors to generate a three-dimensional point cloud of the roll surface, enabling the controller to dynamically compensate for geometric deviations, surface contamination, and residual mill scale that are inherent in as-received roller press components.
The core principle relies on a closed-loop feedback architecture: a pre-weld vision scan maps the roll surface topography; the robot controller processes this data against the programmed weld track geometry; during welding, a real-time seam-tracking sensor continuously monitors the arc position relative to the target path and sends correction signals to the robot servo drives at frequencies exceeding 1 kHz. This eliminates the cumulative positioning errors that plague open-loop robotic welding on large-diameter cylindrical parts.
2. Category and Business Positioning
This technology falls squarely within the TIG/MIG weld overlay technology route of Cladding Technology Shanxi Co., Ltd., specifically addressing the automated and semi-automated overlay welding segment. Within the company's broader capability portfolio, it occupies the intersection of:
- Process engineering — developing qualified welding procedures (WPS/PQR) for specific overlay alloys on high-carbon roll steel substrates
- Equipment integration — designing and commissioning robotic welding cells with vision-guided systems
- Quality assurance — ensuring consistent overlay thickness, microstructure, and bond strength across production batches
The business positioning targets large-scale industrial customers in iron ore pelletizing, coal preparation, mineral processing, and cement grinding operations where roller press availability is directly tied to production throughput and revenue.
3. Technical Purpose and Value
3.1 Primary Engineering Objectives
- Wear resistance enhancement: Depositing carbide-containing alloys (e.g., Cr-C-Mo high-chromium white iron, Ni-Cr-Mo-B, or tungsten carbide composite overlays) to extend roll service life by 3–8× compared to bare steel
- Friction coefficient control: Achieving surface roughness Ra 25–125 μm to optimize ore/cake pickup and release characteristics in the nip zone
- Dimensional restoration: Rebuilding worn roll diameters to within manufacturing tolerances without full roll replacement
- Production efficiency: Reducing overlay cycle time by 40–60% compared to manual or semi-automatic torch methods
3.2 Customer Value Proposition
For end-users operating roller presses in continuous duty cycles, each hour of unplanned downtime costs $5,000–$20,000 depending on commodity value. A vision-guided robotic overlay system delivers:
- Reduced roll downtime from 72–120 hours (manual welding) to 24–48 hours per overlay cycle
- Elimination of operator fatigue-related defects in multi-shift welding operations
- Traceable process data enabling predictive maintenance scheduling
- Consistent overlay quality independent of operator skill level
4. Key Process and Implementation Points
4.1 System Architecture Components
| Component | Specification / Function | Performance Requirement |
|---|---|---|
| Industrial Robot | 6-axis articulated, payload ≥25 kg, reach ≥2.5 m | Positioning accuracy ±0.05 mm, repeatability ±0.03 mm |
| Pre-Weld Vision Sensor | Structured light 3D scanner (e.g., Keyence, Micro-Epsilon) | Point density ≥0.5 mm, accuracy ±0.02 mm |
| Seam Tracking Sensor | Laser line or arc sensing (e.g., IPG, Hitachi) | Tracking bandwidth ≥1 kHz, correction lag <5 ms |
| Weld Power Source | Pulsed TIG or GMAW, 500–800 A range | Current stability ±2%, pulse frequency 1–10 kHz |
| Roll Rotary Positioner | Variable speed 0.5–10 rpm, brake torque ≥5000 N·m | Speed stability ±0.5%, encoder resolution ≥1024 pulses/rev |
| Controller/Software | Real-time kinematics solver with vision fusion | Path planning cycle time <50 ms |
4.2 Weld Overlay Process Parameters
| Parameter | Typical Range (TIG) | Typical Range (MIG) | Notes |
|---|---|---|---|
| Base material | High-carbon steel (C ≥0.6%), quenched & tempered | Same | HRC 45–60 pre-weld hardness |
| Overlay alloy | Cr-C-Mo, Ni-Cr-B, Co-Cr-W (Stellite-type) | Same | Selected per wear mechanism |
| Wire/rod diameter | Φ2.4–3.2 mm | Φ1.2–1.6 mm | — |
| Travel speed | 30–80 mm/min | 150–400 mm/min | — |
| Current | 120–250 A (pulsed) | 180–350 A | — |
| Heat input | 1.5–4.5 kJ/mm | 0.8–2.5 kJ/mm | Critical for dilution control |
| Overlay thickness per pass | 1.5–3.0 mm | 1.0–2.5 mm | Multi-pass builds to 5–20 mm total |
| Preheat temperature | 150–350°C | 100–250°C | Prevents cracking in high-C base |
| Interpass temperature | ≤250°C (max) | ≤200°C (max) | Monitor with IR pyrometer |
| Shielding gas | Ar or Ar/2%O₂ | Ar/5–8%CO₂ or Ar/15%CO₂ | — |
4.3 Vision System Implementation Sequence
- Pre-weld scanning: 3D scanner captures full roll surface geometry including existing wear patterns, cracks, and surface defects. Data is imported into the robot programming software as a digital twin.
- Path generation: The software generates helical weld tracks optimized for uniform bead overlap (typically 30–50% overlap between adjacent passes) and minimum total welding time.
- Defect avoidance programming: Detected cracks or voids are flagged; the path is modified to include tack-weld repairs or the area is routed around pending manual intervention.
- First-pass welding: Robot executes the first overlay pass with real-time seam tracking engaged. The tracking sensor monitors the torch position relative to the previous pass edge or a scribe line.
- Inter-pass vision check (optional): After every 2–3 passes, the vision system re-scans to verify accumulated overlay thickness and detect any porosity or lack-of-fusion that developed in earlier passes.
- Final pass and surface finishing: The last pass is programmed with reduced heat input and optimized travel speed to achieve the target surface roughness without excessive reinforcement.
4.4 Critical Process Controls
- Dilution management: For hardfacing overlays, dilution must be controlled to ≤25–35% to maintain the required microstructure (e.g., retained austenite + carbide network). The vision system monitors bead width-to-height ratio as a proxy for dilution in real time.
- Thermal management: Rolling the part continuously and/or using intermittent welding sequences prevents localized overheating that could soften the base metal beyond HRC 40 in the heat-affected zone.
- Spatter and contamination control: MIG processes generate significant spatter on the roll surface, which the vision system must distinguish from surface topography. Gas flow optimization (laminar flow shields, wire cup design) reduces spatter by 60–80%.
- Hydrogen control: Wire and surface moisture management prevents cold cracking. Pre-weld bake at 150–200°C for 2 hours removes adsorbed moisture from the base metal surface.
5. Applicable Standards and Acceptance Criteria
5.1 Governing Standards
| Standard | Scope | Relevance |
|---|---|---|
| GB/T 985.1 | Welding terminology and symbols | Documentation and drawing interpretation |
| GB/T 12467 | Welding procedure specification (WPS) and qualification record (PQR) | Procedure development and qualification |
| GB/T 19418 | Welding procedure specification and qualification — General requirements | WPS/PQR framework |
| ASME Section IX | Qualification of welding procedures, welders, and welding operators | Welder/operator qualification (if ASME-stamped equipment) |
| AWS D10.9 | Standard for qualification and certification of welding personnel | Robotic weld system operator qualification |
| AWS D10.9M | Metric version of D10.9 | Metric qualification requirements |
| ASTM A396 | Standard specification for carbon and alloy steel rolls for mining | Base material specification for roller press rolls |
| ASTM A397 | Standard specification for carbon and alloy steel rolls for hot rolling | Alternative base material reference |
| ISO 3834-2 | Quality requirements for fusion-welded products — Comprehensive requirements | Quality management system for welding operations |
| NACE SP0169 | Corrosion prevention in soil and freshwater environments — Cathodic protection | Applicable where overlay also serves corrosion protection |
| GB/T 3323 | Non-destructive testing — Radiographic testing of welds | NDT acceptance criteria |
| GB/T 11345 | Non-destructive testing — Ultrasonic testing of welds | NDT acceptance criteria |
| GB/T 11346 | Non-destructive testing — Magnetic particle testing | Surface defect detection |
| GB/T 6060 | Non-destructive testing — Visual examination | Visual acceptance criteria |
5.2 Acceptance Criteria for Overlay Welds
- Visual inspection: No cracks, undercut >0.5 mm, excessive reinforcement (>3 mm), or incomplete fusion visible to the naked eye. Conform to GB/T 6060 Level B.
- Ultrasonic testing: No internal defects (porosity, slag inclusion, lack of fusion) exceeding 2 mm equivalent diameter per GB/T 11345 Level 2.
- Magnetic particle testing: No linear indications (cracks, hot tears) exceeding 3 mm in length. Round indications ≤1.5 mm are acceptable if ≤3 per 100 mm of weld length.
- Hardness verification: Overlay surface hardness must meet specification (typically HRC 55–65 for Cr-C-Mo, HV 1200–1800 for tungsten carbide composites). Base metal HAZ hardness must not exceed HRC 45 to prevent cracking susceptibility.
- Dilution measurement: Metallographic cross-section analysis at weld toe showing ≤30% base metal in the first 0.5 mm of overlay depth.
- Overlay thickness: Final thickness within ±0.5 mm of specified nominal value, uniformity across the roll face within ±10%.
- Bond strength: Peel test or bend test demonstrating no separation at the overlay-base interface. For critical applications, bond strength ≥250 MPa per ASTM G102.
6. Common Risks and Controls
| Risk | Mechanism | Control Measure |
|---|---|---|
| Cracking in base metal (HAZ) | High carbon content + rapid cooling → martensitic transformation with high residual stress | Preheat to 200–350°C; interpass temp ≤250°C; post-weld stress relief at 550–650°C for 2 h; use low-heat-input parameters |
| Overlay cracking | High dilution → soft, ductile microstructure with thermal cracking susceptibility; or low dilution → brittle carbide network with cold cracking | Optimize dilution to 20–30%; ensure proper wire preheating; maintain gas flow ≥15 L/min; avoid excessive restraint |
| Porosity | Hydrogen absorption from wire moisture, surface contamination, or atmospheric ingress | Wire storage in heated cabinet (150°C); pre-weld surface cleaning (grinding + solvent wipe); verify gas flow and nozzle condition each shift |
| Inconsistent overlay thickness | Robot positioning drift; vision sensor calibration drift; thermal distortion of roll during welding | Implement thermal compensation algorithm in controller; perform vision calibration verification every 8 hours; use closed-loop thickness monitoring |
| Spatter-induced surface roughness | MIG spatter solidifies on overlay surface creating stress concentrators | Use optimized wire cup geometry; increase stand-off distance to 15–20 mm; consider pulsed MIG with reduced peak current; post-weld dressing pass |
| Roll distortion | Asymmetric heat input causes barrel distortion or cone distortion | Program symmetric heat input (alternating sides); implement rolling during welding; use back-of-weld cooling (indirect water spray) |
| Robot programming errors | Incorrect kinematic compensation for cylindrical coordinate transformation | Implement 5-axis coordinated motion (robot + rotary axis); verify with dry-run simulation before production; include collision detection |
7. Application Across the Company's Three Technology Routes
7.1 TIG/MIG Weld Overlay Route
The machine vision robotic system is the primary enabler for high-volume, repeatable weld overlay production. Specific applications include:
- Iron ore pelletizing roller press rolls: Multi-pass TIG overlay of Cr-C-Mo (e.g., D2 equivalent or proprietary alloy) achieving HRC 58–62 surface hardness. Typical overlay thickness 8–15 mm on rolls of Φ1200–2000 mm diameter.
- Coal preparation roller press rolls: MIG overlay of Ni-Cr-B alloy for moderate abrasion resistance with controlled friction coefficient for coal cake release.
- Cement grinding roller press rolls: Tungsten carbide composite overlay (80% WC + 20% Co binder) applied via TIG process for extreme abrasion resistance in wet grinding conditions.
- Transition layer welding: The vision system guides the deposition of a 309L/310L stainless steel transition layer between the high-carbon base metal and the hardfacing overlay, preventing cracking at the interface.
7.2 Hydraulic Explosive Bonding Route
While the robotic vision system is primarily a welding technology tool, it contributes to the hydraulic explosive bonding route in the following ways:
- Post-bond surface preparation: After hydraulic explosive bonding of a cladding plate to the roll sleeve, the interface may require weld repair at edges or defects. The vision-guided robot performs these localized TIG repairs with sub-millimeter precision.
- Quality verification: The 3D vision system provides as-built geometry data of the bonded interface, enabling comparison against design specifications and identification of areas requiring additional processing.
- Edge grinding and finishing: Post-bond grinding operations can be guided by the same vision infrastructure, ensuring uniform cladding thickness across the roll face.
7.3 Explosion Welding Route
For explosion-welded roll sleeves (where a copper or stainless steel cladding plate is explosion-bonded to a steel backing plate, then fabricated into a roll sleeve), the vision system contributes to:
- Pre-weld inspection: 3D scanning of explosion-welded plates to verify bond quality indicators (wavy interface pattern, thickness uniformity) before fabrication into roll sleeves.
- Weld joint positioning: During circumferential and longitudinal welding of the cladded plate into a roll sleeve, the vision system ensures accurate weld joint alignment and tracks the weld seam through the multi-layer welding sequence.
- Post-fabrication dimensional verification: Final roll sleeve geometry is scanned and compared against CAD models to verify runout, out-of-roundness, and surface flatness specifications.
8. Contribution to Qualification Building, Product Delivery, and Customer Value
8.1 Qualification Building
- WPS/PQR development: The robotic system's repeatability enables efficient qualification testing. Multiple test coupons can be welded under identical conditions, producing statistically robust mechanical property data for procedure qualification per GB/T 19418 or ASME Section IX.
- Operator qualification: The vision-guided system reduces the skill barrier for robotic weld operators, simplifying the qualification process per AWS D10.9. Operators demonstrate system setup and monitoring competency rather than manual welding skill.
- Equipment qualification: Documentation of the vision system's calibration procedures, sensor accuracy verification, and controller software version control establishes traceability for customer audit requirements.
- ISO 3834 compliance: The system's data logging capability (welding parameters, vision data, robot paths) provides the objective evidence required for ISO 3834-2 Level 2 certification.
8.2 Product Delivery Enhancement
- Cycle time reduction: Automated overlay reduces roll rebuild time from 5–10 days (manual) to 2–4 days (robotic), enabling faster turnaround and reduced customer inventory requirements.
- Quality consistency: Coefficient of variation in overlay hardness reduced from ±5 HRC (manual) to ±2 HRC (robotic), ensuring uniform wear performance across the roll face.
- Scalability: The system can be configured for different roll diameters (Φ800–Φ3000 mm) and overlay alloys through software parameter changes, enabling the company to serve diverse customer bases without dedicated hardware for each application.
- Documentation package: Each completed overlay job generates a digital quality file including: pre-weld scan data, welding parameter log, in-process vision verification images, and post-weld NDT results — a powerful differentiator in competitive bidding.
8.3 Customer Value Creation
- Total cost of ownership reduction: Extended roll life (3–8×) combined with reduced downtime translates to 40–70% reduction in roll-related cost per ton of processed material.
- Process optimization: The vision system's data on roll wear patterns enables predictive maintenance recommendations, allowing customers to schedule roll rebuilds during planned maintenance windows rather than emergency shutdowns.
- Environmental benefit: Roll rebuild via overlay consumes 60–80% less material and energy compared to roll replacement, supporting customers' sustainability and ESG targets.
- Technical partnership: The company positions itself not merely as a welding contractor but as a process technology partner, providing continuous improvement data and optimization recommendations based on accumulated vision system analytics across the customer's fleet.
9. Summary and Forward Outlook
The machine vision-based robotic weld overlay system represents a critical capability enabler for Cladding Technology Shanxi Co., Ltd. It transforms the company from a traditional welding service provider into a technology-driven solutions partner capable of delivering quantifiable performance improvements to industrial customers. The integration of real-time sensing, adaptive control, and data analytics creates a foundation for Industry 4.0 compliance and positions the company to capture emerging markets in intelligent manufacturing and predictive maintenance.
Future development directions include:
- Integration of AI-based defect detection algorithms for real-time in-process quality assessment
- Development of adaptive parameter control that automatically adjusts welding parameters based on real-time thermal imaging feedback
- Extension of the vision system to multi-torch simultaneous welding for further cycle time reduction
- Development of digital twin capabilities enabling virtual commissioning and remote operation