Vision-Based Weld Pool Width Detection in TIG Welding of Copper-Clad Aluminum Cable

1. Definition and Technical Principles

1.1 Core Concept

Vision-based detection of weld pool width refers to the deployment of machine vision systems—typically comprising a high-resolution camera, optical filter, and real-time image processing algorithm—to monitor, measure, and control the width of the molten weld pool during TIG (Tungsten Inert Gas) welding operations on copper-clad aluminum (CCA) cable assemblies. The technique captures the geometric profile of the weld pool in real time and feeds this data back into the welding control system to ensure process stability, dimensional accuracy, and metallurgical integrity of the joint.

1.2 Operating Principle

The fundamental principle relies on the optical contrast between the molten weld pool and the surrounding solid-state material. The weld pool, due to its high temperature, emits intense visible and near-infrared radiation. By equipping the camera with appropriate band-pass filters (commonly in the 500–700 nm visible range or 700–1000 nm near-IR range), the system isolates the weld pool region from background interference such as arc light, spatter, and oxide films. Image processing algorithms then perform edge detection (e.g., Sobel, Canny, or threshold-based segmentation) to identify the left and right boundaries of the pool, calculate the width, and compare it against a predefined setpoint or tolerance window.

In the specific context of copper-clad aluminum cable welding, the challenge is compounded by the dissimilar nature of the constituent metals—copper (outer layer) and aluminum (core)—which exhibit significantly different thermal conductivity, melting points, and emissivity. Copper has a thermal conductivity of approximately 401 W/(m·K) and a melting point of 1085 °C, while aluminum has a thermal conductivity of approximately 237 W/(m·K) and a melting point of 660 °C. These differences result in non-uniform heat distribution and asymmetric weld pool geometry, making vision-based monitoring particularly valuable for maintaining process control.

1.3 Signal Acquisition and Processing Chain

• Image Acquisition: A CCD or CMOS camera positioned at a fixed standoff distance (typically 30–80 mm) above or to the side of the weld zone captures frames at 30–120 fps.
• Optical Filtering: Band-pass or long-pass filters suppress arc glare and select wavelength bands where the weld pool has maximum contrast against the substrate.
• Image Pre-processing: Noise reduction (Gaussian or median filtering), contrast enhancement, and background subtraction prepare the image for edge extraction.
• Edge Detection and Width Calculation: Algorithmic segmentation identifies pool boundaries; the horizontal distance between left and right edges yields the instantaneous weld pool width.
• Feedback Control: The measured width is compared against a target value; deviations trigger adjustments to welding current, travel speed, torch angle, or shielding gas flow.

2. Category and Business Positioning

2.1 Technology Classification

This technology belongs to the category of intelligent welding process monitoring and control systems, specifically under the sub-domain of sensor-based feedback control for TIG welding operations. It is classified as an advanced manufacturing technology that bridges traditional welding practice with Industry 4.0 digital manufacturing capabilities. Within the broader context of Cladding Technology Shanxi Co., Ltd.'s capability portfolio, this technology falls under the TIG/MIG weld overlay route and represents the company's commitment to precision manufacturing and quality assurance.

2.2 Business Positioning

The deployment of vision-based weld pool monitoring serves multiple strategic purposes for the company:

• Quality Assurance Differentiator: Demonstrates the company's capability to deliver high-reliability dissimilar metal joints with statistically controlled dimensional accuracy, distinguishing the company from competitors relying on manual inspection or post-weld testing alone.
• Process Qualification Enhancement: Provides documented, traceable process data that supports WPS (Welding Procedure Specification) qualification and PQR (Procedure Qualification Record) generation under standards such as ASME Section IX and AWS D1.1.
• Customer Confidence: Enables real-time process transparency for end customers in demanding industries such as power transmission, rail transit, and marine engineering, where cable joint reliability is mission-critical.
• Intellectual Property Development: The study and adaptation of academic research into production-applicable systems builds proprietary expertise and supports patent filings in intelligent welding monitoring.

3. Technical Purpose and Value

3.1 Primary Technical Objectives

1. Weld Pool Dimensional Control: Maintain weld pool width within a target range (typically ±10–15% of nominal) to ensure uniform penetration, adequate fusion, and consistent bead geometry.
2. Real-Time Process Stability Monitoring: Detect anomalies such as porosity, undercut, incomplete fusion, or excessive dilution before they become permanent defects.
3. Welding Parameter Optimization: Provide empirical data to refine welding parameter windows for copper-clad aluminum cable joints, reducing trial-and-error during procedure development.
4. Process Documentation and Traceability: Generate timestamped records of weld pool dimensions throughout the welding sequence, supporting quality audits and root-cause analysis.

3.2 Value to Product Delivery

In copper-clad aluminum cable applications, the weld joint between the CCA cable and a copper terminal or busbar is a critical connection point. Failure of this joint can lead to overheating, increased electrical resistance, and catastrophic system failure. Vision-based monitoring directly contributes to:

• Reducing weld defect rates from typical industry averages of 3–5% to below 1% through closed-loop process control.
• Increasing production throughput by enabling faster welding speeds with confidence in joint quality, as the system compensates for parameter drift in real time.
• Minimizing rework and scrap costs associated with post-weld NDT failures (e.g., ultrasonic testing or X-ray radiography revealing internal defects).
• Enabling consistent production across multiple operators and shifts, reducing the dependence on individual operator skill.

3.3 Value to Qualification Building

The integration of vision-based monitoring into the company's TIG welding procedures strengthens the technical foundation for:

• WPS/PQR Qualification: Documented weld pool width data serves as supplementary evidence of process control during qualification welding, demonstrating that the procedure produces consistent results across the full range of qualified variables.
• ISO 3834 / ISO 3836 Compliance: The system supports the documented quality management requirements of these international welding standards, particularly regarding process monitoring and control.
• ASME Section IX Compliance: Provides objective data for procedure qualification records, supporting the evaluation of essential variables and their effects on weld quality.
• NB/T 47014 and GB/T 19866 Compliance: Meets Chinese national and industry standards for welding procedure qualification of pressure vessels and equipment, providing traceable process parameter records.

4. Key Process and Implementation Points

4.1 System Hardware Configuration

Component	Specification	Function
Camera Sensor	Industrial CCD/CMOS, resolution ≥ 1280×960, frame rate ≥ 60 fps	High-speed image capture of weld pool region
Optical Filter	Band-pass filter, center wavelength 550–650 nm (green channel) or 700–900 nm (near-IR)	Suppress arc glare; maximize pool-to-substrate contrast
Lens	Fixed focal length, 8–16 mm, depth of field optimized for standoff distance	Focus image on weld pool plane with minimal distortion
Mounting	Fixed-position bracket on welding torch or gantry, standoff 40–80 mm	Maintain consistent viewing geometry throughout weld
Processing Unit	Industrial PC or embedded controller, real-time OS, ≥ 2 GHz CPU	Execute image processing and feedback algorithms
Communication Interface	RS-232/RS-485 or EtherCAT link to welding power source controller	Transmit width measurements and control signals to welder

4.2 Key Welding Parameters for CCA Cable TIG Welding

Parameter	Typical Range	Effect on Weld Pool Width	Control Strategy
Welding Current	60–150 A (DC)	Directly proportional; higher current increases pool width	Primary control variable for width adjustment
Travel Speed	50–200 mm/min	Inversely proportional; higher speed narrows pool width	Secondary control variable; adjusted with current
Torch Angle	0°–20° from vertical	Asymmetric angle creates asymmetric pool; increases effective width on leading side	Maintained constant; monitored for drift
Torch Standoff	3–6 mm	Larger standoff increases arc spread, widening pool	Mechanically controlled; verified periodically
Shielding Gas Flow	10–20 L/min (Ar or Ar/He mix)	Indirect effect; inadequate flow causes oxidation, altering pool surface appearance	Monitored via flowmeter; not a primary width control
Weld Pool Width Target	3–8 mm (depending on cable diameter)	—	Setpoint for vision-based feedback control

4.3 Implementation Steps

1. Baseline Parameter Establishment: Conduct preliminary welding trials on CCA cable specimens without vision monitoring to establish baseline weld pool width characteristics as a function of current and travel speed. Record pool width measurements for each parameter combination using post-weld optical microscopy.
2. Camera System Calibration: Mount the camera at the designated standoff distance and viewing angle. Calibrate the pixel-to-mm conversion factor using a reference target (e.g., a calibrated graticule or known-dimension test coupon). Verify focus and field of view cover the expected weld pool region.
3. Filter Selection and Optimization: Test multiple filter wavelengths to identify the band that provides maximum contrast between the weld pool and surrounding material for the specific CCA cable geometry and welding parameters. Document the optimal filter selection.
4. Image Processing Algorithm Development: Implement edge detection and width calculation algorithms. Validate algorithm accuracy against manual measurements of known weld pools. Tune threshold values and edge detection parameters for robust performance across the expected range of pool geometries.
5. Closed-Loop Control Integration: Connect the processing unit to the welding power source controller. Program the feedback control logic: if measured width deviates from target by more than the tolerance threshold, adjust welding current (primary) or travel speed (secondary) to correct the deviation. Implement anti-oscillation damping to prevent over-correction.
6. Validation Welding Trials: Conduct a series of validation welds comparing vision-controlled welding against conventional parameter-fixed welding. Evaluate weld pool width consistency, bead geometry, microstructure, and mechanical properties. Document results for WPS qualification support.
7. Production Deployment and Operator Training: Integrate the system into production welding stations. Train operators on system startup, parameter setpoint configuration, alarm response, and basic troubleshooting. Establish daily calibration check procedures.

4.4 Image Processing Algorithm Flow

1. Capture raw image frame from camera sensor.
2. Apply noise reduction filter (e.g., 3×3 median filter) to suppress random noise.
3. Apply band-pass filter to isolate selected wavelength band.
4. Perform contrast enhancement (histogram equalization or CLAHE) to maximize pool-to-background contrast.
5. Apply adaptive thresholding to segment the weld pool region from the background.
6. Identify connected components; select the largest component as the weld pool candidate.
7. Extract left and right boundary coordinates using edge detection (Sobel or Canny operator).
8. Calculate weld pool width as the horizontal distance between left and right boundaries.
9. Compare measured width against target setpoint and tolerance window.
10. If deviation exceeds tolerance, generate control signal to adjust welding parameters.
11. Log measurement data with timestamp for process documentation.

5. Applicable Standards and Acceptance Criteria

5.1 Welding Procedure Standards

• ASME Section IX: Governs qualification of welding procedures and welders for pressure vessels and piping. Weld pool width data supports evaluation of essential variables (current, travel speed, electrode type) and their effects on weld quality.
• AWS D1.1/D1.1M: Structural welding code for steel. While primarily for structural steel, the quality management principles and procedure qualification requirements apply by analogy to dissimilar metal cable welding.
• GB/T 19866: Chinese national standard for qualification of welding procedures for steel, nickel, and their alloys. Provides the framework for WPS/PQR qualification in Chinese manufacturing contexts.
• NB/T 47014: Chinese industry standard for welding procedure qualification of pressure vessels. Requires documented process parameter control and weld quality verification.
• ISO 15614-1: International standard for qualification testing of welding procedures for metallic materials. Specifies test methods and acceptance criteria for weld qualification.

5.2 Weld Quality Acceptance Criteria

Acceptance Parameter	Criterion	Measurement Method	Governing Standard
Weld Pool Width Consistency	Deviation from target ≤ ±15% throughout weld length	Vision-based real-time measurement; post-weld optical microscopy verification	Internal specification; supports ISO 15614-1
Weld Penetration	Full fusion through CCA cable thickness; no incomplete fusion	Macrograph examination of cross-section	ASME Section IX; GB/T 19866
Weld Defects (Internal)	No porosity > 0.5 mm, no cracks, no slag inclusion > 0.5 mm	Ultrasonic testing (UT) or X-ray radiography (RT)	GB/T 3323; NB/T 47013
Weld Defects (Surface)	No undercut > 0.5 mm depth, no cracks, no excessive convexity/concavity	Visual testing (VT) and dye penetrant testing (PT)	ASME Section V; GB/T 11345
Microstructure	No brittle intermetallic phases (e.g., Al₂Cu, CuAl₂) exceeding acceptable limits; grain size within specification	Optical microscopy of etched cross-section	Internal specification; AWS D1.1 principles
Mechanical Properties	Tensile strength ≥ 80% of base metal strength; hardness within acceptable range	Tensile testing and microhardness testing	ASME Section IX; GB/T 228
Electrical Resistance	Joint resistance ≤ 1.2× resistance of equivalent length of unwelded cable	4-wire (Kelvin) resistance measurement	Internal specification; IEEE standards for cable joints

5.3 Quality Management Standards

• ISO 3834-1 through ISO 3834-3: Fundamental requirements for welding quality in steel, nickel, and their alloys. The vision-based monitoring system supports compliance with Clause 5 (Welding Procedure Specifications), Clause 6 (Welding Equipment), and Clause 7 (Welding Personnel).
• ISO 9001:2015: Quality management system requirements. The system provides objective evidence of process control, supporting the "Plan-Do-Check-Act" cycle required by the standard.
• ISO/IEC 17025: If the company performs welding procedure qualification testing for external customers, this standard governs the competence of testing laboratories. Vision-based monitoring data contributes to the documented technical competence of the testing capability.

6. Common Risks and Controls

6.1 Technical Risks

Risk	Description	Impact	Control Measure
Optical Interference from Arc	Intense arc radiation overwhelms camera sensor, causing saturation and loss of weld pool visibility	Loss of monitoring capability; potential undetected defects	Use high-density optical filters (ND 4–6); position camera to minimize direct arc viewing; implement automatic exposure control
Weld Pool Geometry Variability	Copper-aluminum dissimilar metals create asymmetric pool geometry; pool width may vary with cable position and geometry	Inaccurate width measurement; false alarm or missed defect	Calibrate algorithm for expected asymmetric pool shape; use multi-point width measurement (e.g., leading edge, trailing edge, center); implement adaptive thresholding
Spatter and Oxide Interference	Weld spatter and aluminum oxide particles on pool surface disrupt edge detection	Erroneous width measurements; control instability	Apply morphological operations (opening/closing) to remove small artifacts; use temporal filtering (moving average) to smooth measurements; optimize shielding gas coverage
Camera Drift or Misalignment	Thermal expansion or mechanical vibration displaces camera from calibrated position	Systematic measurement bias; loss of calibration	Mount camera on rigid bracket; implement periodic calibration check (e.g., daily reference target measurement); use thermal compensation if necessary
Control Loop Oscillation	Over-aggressive feedback control causes welding parameters to oscillate, destabilizing the weld pool	Weld quality degradation; potential defect formation	Implement PID control with tuned gain and damping; set maximum adjustment rate limits; use dead-band tolerance to prevent minor corrections
Algorithm Failure on Unusual Pool Geometry	Edge detection algorithm fails when pool geometry deviates significantly from training conditions (e.g., during start/stop transitions)	Loss of monitoring during critical weld transitions	Implement algorithm confidence scoring; flag low-confidence measurements; use rule-based fallback for start/stop transitions; train algorithm on diverse pool geometries

6.2 Operational Risks

• Operator Override: Operators may bypass the vision-based control system during production pressure, reverting to manual parameter control. Control: Integrate the system as an interlock in the welding power source; require authorized override with documented justification.
• Maintenance Neglect: Camera lens contamination, filter degradation, or sensor aging may go undetected, leading to gradual measurement drift. Control: Establish preventive maintenance schedule; implement automated calibration verification at each shift start.
• Environmental Conditions: Ambient lighting changes, temperature fluctuations, and vibration in the production environment may affect system performance. Control: Enclose the camera in a protective housing with controlled illumination; mount on vibration-isolated platform; implement environmental compensation algorithms.

7. Application Across the Company's Three Technology Routes

7.1 TIG/MIG Weld Overlay Route

Within the TIG/MIG weld overlay route, vision-based weld pool monitoring has direct and immediate applicability. For TIG welding of copper-clad aluminum cable joints, the technology provides real-time feedback control of weld pool dimensions, ensuring consistent penetration and fusion across the dissimilar metal interface. The monitoring system can be extended to other TIG weld overlay applications within the company's portfolio, including:

• Copper-clad steel pipe weld overlay: Monitoring weld pool width during overlay of copper or copper alloy cladding on steel pipe surfaces, ensuring uniform overlay thickness and adequate fusion.
• Stainless steel weld overlay on carbon steel: Monitoring pool width during application of 309L or 310 transition layers and 316L/321 overlay layers on carbon steel substrates for corrosion resistance.
• Aluminum weld overlay on copper substrates: Monitoring the delicate pool geometry when welding aluminum alloys onto copper surfaces, where thermal conductivity mismatch creates challenging pool control conditions.

For MIG (GMAW) weld overlay operations, the vision-based system can be adapted to monitor the larger, more dynamic weld pool characteristic of MIG welding. The higher deposition rates and greater heat input of MIG require more robust optical filtering and potentially higher frame rates, but the fundamental monitoring principle remains the same. The system can also monitor wire feed consistency and arc length indirectly through pool width variations.

7.2 Hydraulic Explosive Bonding Route

While hydraulic explosive bonding (also known as hydraulic explosion bonding or hydraulic explosive cladding) relies on high-pressure water jet impact rather than welding, the vision-based monitoring technology contributes to this route in complementary ways:

• Post-Bonding Weld Repair Monitoring: Hydraulic explosive bonding can produce localized defects or weak bond regions. Repair welding of these regions using TIG welding benefits from vision-based pool monitoring to ensure precise, controlled repair welds that do not compromise the surrounding bonded interface.
• Bond Interface Characterization: After hydraulic explosive bonding, the bond quality is verified through macrograph examination of the characteristic wave pattern at the interface. Vision-based systems can be adapted to automate the analysis of wave amplitude, wavelength, and coverage, providing quantitative bond quality assessment.
• Hybrid Process Development: For hybrid processes combining hydraulic explosive bonding with TIG welding (e.g., explosive bonding followed by weld overlay for additional thickness or corrosion resistance), the vision-based system provides unified monitoring across both process stages.

7.3 Explosion Welding Route

In explosion welding, the vision-based monitoring technology serves several important functions:

• Pre-Weld Preparation Monitoring: Before explosion welding, base plates and cladding plates may require edge preparation, including welding of backing plates or temporary fixtures. Vision-based monitoring ensures quality of these preparatory welds.
• Post-Weld Inspection Support: After explosion welding, the bond interface is examined through macrograph analysis. Automated vision systems can analyze the wave pattern morphology to assess bond quality, supplementing traditional NDT methods such as magnetic particle testing and ultrasonic testing.
• Repair and Retouch Welding: Expulsion welding (a variant of explosion welding) and other explosive cladding methods may leave regions requiring post-weld repair. TIG repair welding of these regions, monitored by the vision-based system, ensures that repair welds achieve proper fusion without excessive dilution of the explosion-welded interface.
• Process Development and Research: For R&D activities involving explosion welding of new material combinations, the vision-based system provides quantitative data on weld pool behavior during supplementary welding operations, supporting procedure development and qualification.

8. Contribution to Qualification Building, Product Delivery, and Customer Value

8.1 Qualification Building

The vision-based weld pool monitoring technology strengthens the company's qualification portfolio in three dimensions:

1. WPS/PQR Documentation: Real-time weld pool width data provides objective, quantifiable evidence of process control during qualification welding. This data can be incorporated into PQR records as supplementary information demonstrating that the welding procedure produces consistent, repeatable results across the full range of qualified variables.
2. Welder Qualification Support: Welder performance qualification under ASME Section IX and GB/T 19866 requires demonstration of consistent weld quality. Vision-based monitoring data provides additional evidence of welder performance, showing that weld pool dimensions remain within acceptable limits throughout the qualification weld.
3. Technology Capability Certification: The successful implementation of vision-based monitoring demonstrates the company's capability in intelligent manufacturing and advanced process control, supporting applications for certifications such as ISO 3834-2 (comprehensive requirements) or ISO 3834-3 (basic requirements), as well as industry-specific certifications for nuclear, aerospace, or automotive welding.

8.2 Product Delivery Enhancement

For product delivery, the technology delivers measurable improvements:

• Defect Rate Reduction: By maintaining weld pool width within tight tolerances throughout the welding process, the technology reduces the incidence of geometric defects (undercut, excessive convexity/concavity) and metallurgical defects (incomplete fusion, excessive dilution). Target reduction: from 3–5% to below 1%.
• Production Speed Improvement: With confidence in real-time process control, welding parameters can be optimized toward the upper end of the acceptable range, increasing deposition rates and reducing cycle times. Expected improvement: 10–20% throughput increase.
• Rework Elimination: By detecting and correcting process deviations in real time, the technology prevents the formation of defects that would require post-weld rework. Expected reduction in rework: 50–70%.
• Consistency Across Production: The automated control system ensures consistent weld quality regardless of operator, shift, or production volume, enabling reliable delivery of large production runs with uniform quality.

8.3 Customer Value

The technology creates direct value for the company's customers:

• Enhanced Reliability: Customers in power transmission, rail transit, and marine engineering receive cable joints and clad components with statistically controlled quality, reducing the risk of in-service failures and associated safety and economic consequences.
• Documentation and Traceability: Each welded component can be accompanied by a digital record of weld pool dimensions throughout the welding process, providing customers with full traceability for quality audits, regulatory compliance, and performance verification.
• Cost Reduction: Reduced defect rates, lower rework costs, and higher production efficiency translate to competitive pricing for customers without compromising quality.
• Technical Partnership: The availability of advanced process monitoring capabilities positions the company as a technology partner rather than a commodity supplier, enabling collaborative development of custom solutions for demanding applications.
• Regulatory Compliance Support: Customers subject to regulatory requirements (e.g., nuclear, aerospace, or transportation safety regulations) benefit from the comprehensive documentation and traceability provided by the vision-based monitoring system.

9. Conclusion

The vision-based detection of weld pool width in TIG welding of copper-clad aluminum cable represents a significant advancement in the company's intelligent manufacturing capabilities. By integrating machine vision technology with TIG welding process control, the company achieves real-time, quantitative monitoring of weld pool geometry, enabling closed-loop feedback control that ensures consistent weld quality across dissimilar metal joints. The technology contributes to qualification building by providing objective process documentation, enhances product delivery through defect reduction and throughput improvement, and creates substantial customer value through enhanced reliability, traceability, and cost efficiency. Its applicability extends beyond CCA cable welding to the company's full portfolio of TIG/MIG weld overlay, hydraulic explosive bonding, and explosion welding operations, establishing a unified framework for intelligent process monitoring across all technology routes.