Fully Automatic Surfacing Welding Repair of Worn Mechanical Parts

Literature Overview

This 2002 paper by Kolasa A. (Warsaw University of Technology) and Zhao Jiangtao, Zhu Sheng (Academy of Armored Force Engineering), published in China Surface Engineering, presents a fully automated robotic surfacing welding system for the repair of worn mechanical components. The system represents an early example of integrating robotic manipulation, automated measurement, and adaptive welding control into a unified repair process. The paper is particularly notable for its focus on repairing parts with similar geometry but varying degrees of wear, a scenario that is extremely common in industrial maintenance operations.

Technical Context and Motivation

In heavy industry and military applications, mechanical components such as gears, bearings, shafts, and structural elements undergo progressive wear that eventually necessitates repair or replacement. The traditional approach involves:

• Manual measurement of wear depth and pattern
• Manual grinding of the worn surface to a uniform baseline
• Manual surfacing welding to rebuild the worn material
• Manual machining to restore final dimensions

This manual approach suffers from several deficiencies:

• Inconsistent quality due to operator skill variation
• Low productivity, particularly for high-volume repair operations
• Difficulty in maintaining consistent weld quality over extended production runs
• High labor costs in regions with skilled welder shortages

The robotic system described in this paper addresses these limitations through full automation of the measurement, planning, and execution phases of the repair process.

System Architecture

Robotic Platform

The system employs an industrial robot with the following configuration:

Measurement Subsystem

The measurement subsystem is critical for enabling the system to handle parts with varying wear patterns:

• Contact measurement: A probe attached to the robot wrist traces the surface geometry
• Non-contact measurement: Laser or optical displacement sensor for rapid scanning
• Data acquisition: Controller processes measurement data to generate a wear map
• Reference comparison: Measured geometry is compared against nominal CAD model or stored reference

Welding Subsystem

The welding subsystem includes:

• Torch system: GMAW (gas metal arc welding) or FCAW (flux-cored arc welding) torch with automatic wire feed
• Shielding gas supply: Argon/CO2 mixture or self-shielded flux-cored wire
• Power source: Inverter-type welding machine with programmable output
• Cooling system: Water-cooled torch and workpiece cooling fixtures

Control and Programming

The control system orchestrates the complete repair cycle:

1. Part identification and fixture alignment
2. Surface scanning and wear mapping
3. Weld path generation based on wear map
4. Weld parameter selection based on material and wear depth
5. Automated welding execution
6. Post-weld inspection (optional)

Process Flow

The automated repair process follows a structured sequence:

Phase 1: Part Identification and Fixturing

The part is placed in a dedicated fixture that positions it within the robot workspace. The identification process may involve:

• Manual selection from a library of known part types
• Automated recognition through barcode or RFID tagging
• Geometric matching against stored reference models

Phase 2: Surface Measurement

The robot moves the measurement probe across the surface of interest, collecting spatial data at intervals of 1-5 mm. The resulting point cloud is processed to generate:

• A three-dimensional surface model of the worn geometry
• A wear depth map indicating the material loss at each location
• A classification of wear severity (light, moderate, severe)

Phase 3: Weld Path Planning

The control system generates the welding path using the following algorithm:

• The wear map is segmented into zones of similar depth
• For each zone, a welding strategy is selected (single pass, multi-pass, spiral, raster)
• The weld path is optimized to minimize total travel distance while ensuring adequate coverage
• Overlap between adjacent passes is calculated to ensure complete fusion

Phase 4: Weld Execution

The robot executes the planned weld path with the following control features:

• Adaptive travel speed: Adjusted based on measured deposition rate
• Arc length control: Maintained within ±0.5 mm for consistent bead geometry
• Heat input management: Current and voltage adjusted to maintain target dilution
• Interpass temperature monitoring: Thermocouple feedback prevents overheating

Key Technical Parameters

The paper presents the following welding parameters for the automated system:

Application Examples

The paper describes several industrial applications of the system:

Gear Repair

Worn gear teeth are rebuilt through automated surfacing welding followed by machining. The system measures the tooth profile deviation and applies weld material only where needed, minimizing excess material and subsequent machining.

Bearing Race Repair

Worn bearing races are measured for dimensional deviation, and the system applies a uniform overlay layer that is subsequently ground to the required tolerance.

Structural Component Repair

Large structural elements with localized wear or damage are repaired by targeting the specific affected areas, reducing total weld volume and thermal input.

Quality Assurance

The automated system incorporates several quality assurance features:

• In-process monitoring: Arc voltage and current are continuously monitored for anomalies
• Post-weld inspection: Visual inspection and dimensional verification
• Statistical process control: Weld parameters are tracked to detect drift
• Traceability: Each repair is documented with complete parameter records

Comparison with Manual Methods

Critical Analysis

While the paper presents a compelling case for automated surfacing welding, several limitations should be acknowledged:

1. Programming complexity: The initial setup and programming of the system requires significant expertise in both welding and robotics.
2. Flexibility constraints: The system is optimized for parts with known geometries; highly irregular or novel parts may require extensive reprogramming.
3. Capital investment: The equipment cost is substantial and may not be justified for low-volume repair operations.
4. Maintenance requirements: The robot and associated systems require regular maintenance and calibration.

The paper's focus on parts with "similar shape but different wear degree" is astute, as this represents the most common scenario in industrial maintenance. The system's ability to adapt to varying wear patterns through measurement and path planning is its primary advantage over simpler programmed welding systems.

Modern Context and Evolution

Since the publication of this paper in 2002, significant advances have been made in automated welding systems:

• Real-time process monitoring: Modern systems use optical sensors, acoustic emission, and data analysis algorithms to detect and correct welding defects in real time.
• Collaborative robots: Newer collaborative robot platforms are safer and more flexible for integration into existing maintenance workflows.
• Digital twin technology: Virtual simulation of the repair process allows optimization before physical execution.
• Advanced sensors: Structured light scanners and 3D cameras provide faster and more accurate surface measurement.

However, the fundamental principles described in this paper—automated measurement, adaptive path planning, and closed-loop welding control—remain the foundation of modern automated repair systems.

Summary

The fully automatic surfacing welding system described in this paper represents a significant advancement in industrial repair technology. By integrating robotic manipulation, automated measurement, and adaptive welding control, the system achieves consistent, high-quality repairs at substantially higher productivity than manual methods. The approach is particularly well-suited to high-volume maintenance operations where similar components with varying wear patterns require frequent repair. For engineers evaluating automated repair solutions, this paper provides a foundational understanding of the system architecture, process flow, and performance characteristics that continue to define modern robotic welding systems.