Fuzzy Comprehensive Evaluation of Belt Electrode Surfacing Forming Quality Based on Matlab-FIS

Literature Overview

This study by Guo Xiao, Xu Kai, and Zou Liwei from the Harbin Welding Research Institute of the Chinese Academy of Machinery Science and Technology develops a fuzzy comprehensive evaluation model for the forming quality of belt electrode surfacing using the Fuzzy Analytic Hierarchy Process (FAHP) methodology, implemented through the Matlab-FIS (Fuzzy Inference System) platform. Published in the Transactions of the China Welding Institution in 2013 (Vol. 34, No. 8, pp. 89–91), this research was funded by the National Science and Technology Major Project (2012ZX060004-21; 2011ZX04016-061).

Core Technical Findings

Evaluation Index System

The study establishes a four-criteria evaluation framework for belt electrode surfacing forming quality:

The selection of these four indices reflects the practical quality characteristics that determine the functional performance of surfacing layers: thickness affects wear life, width affects coverage area, edge contact angle affects adhesion and stress concentration, and straightness affects dimensional accuracy and downstream processing.

FAHP Methodology

The Fuzzy Analytic Hierarchy Process combines the hierarchical decomposition approach of the Analytic Hierarchy Process (AHP) with fuzzy logic to handle the inherent uncertainty and imprecision in quality evaluation:

1. Hierarchy construction: The evaluation problem is decomposed into the goal layer (forming quality), criteria layer (four indices), and sub-criteria layer (specific measurement values)
2. Fuzzy judgment matrix construction: Instead of crisp pairwise comparisons, fuzzy numbers are used to represent the relative importance of criteria, acknowledging the subjective nature of weight assignment
3. Weight determination: The weights of each evaluation index are calculated from the fuzzy consistent judgment matrix, ensuring mathematical consistency
4. Membership function definition: Based on practical welding experience, membership functions are defined for each index to convert measured values into fuzzy membership degrees (0 to 1)
5. Fuzzy inference: The Matlab-FIS platform is used to perform the fuzzy inference calculation, combining the weighted membership degrees into a comprehensive quality score

Matlab-FIS Implementation

The Matlab-FIS (Fuzzy Inference System) provides a computational framework for implementing the fuzzy evaluation model:

• Input variables: The four evaluation indices (thickness, width, edge contact angle, straightness)
• Membership functions: Triangular or trapezoidal membership functions defined for each input variable based on expert knowledge
• Rule base: Fuzzy rules that relate input conditions to output quality grades
• Inference engine: Mamdani or Sugeno-type fuzzy inference to compute the final quality score
• Defuzzification: Conversion of the fuzzy output into a crisp quality grade

The implementation in Matlab provides a reproducible and transparent calculation framework that can be easily modified as new quality data becomes available or as evaluation criteria are refined.

Validation Results

The evaluation model was validated through experimental tests, and the results showed good agreement between the fuzzy evaluation scores and actual observed quality assessments. This validation confirms that the FAHP methodology, combined with the Matlab-FIS computational framework, provides a reliable and objective approach to surfacing quality evaluation.

Engineering Practice Integration

Quality Control Applications

The fuzzy comprehensive evaluation model has direct applications in:

• In-process quality monitoring: Real-time assessment of surfacing quality during production to enable immediate corrective actions
• End-of-process inspection: Comprehensive quality grading of completed surfacing operations for acceptance/rejection decisions
• Process optimization: Identification of which quality indices most influence overall quality, guiding process parameter adjustments
• Supplier qualification: Objective evaluation of surfacing service providers based on measurable quality criteria

Quality Management System Integration

The evaluation model can be integrated into a broader quality management framework:

• PDCA cycle: The model provides the "Check" function, enabling comparison of actual quality against target specifications
• FMEA integration: The evaluation indices can be mapped to potential failure modes in surfacing operations, with the fuzzy scores serving as severity/occurrence/detection indicators
• Statistical process control: The comprehensive quality score can be used as a control chart variable for monitoring process stability

Belt Electrode Surfacing Process Characteristics

Belt electrode surfacing (also known as strip metal arc welding or SAWS with strip electrode) offers several advantages over conventional electrode surfacing:

• Higher deposition rate: The continuous belt electrode provides a larger cross-sectional area for metal transfer
• Better deposition efficiency: Reduced spatter and improved arc stability
• Lower heat input per unit length: Compared to conventional submerged arc welding with wire electrodes
• Consistent bead geometry: The continuous belt shape promotes uniform bead width and thickness

However, belt electrode surfacing also presents unique challenges:

• Edge quality: The contact angle at the bead edges is critical for adhesion and stress distribution
• Straightness control: The long belt electrode can be susceptible to lateral deflection during deposition
• Thickness uniformity: Variations in belt feed speed or arc length can cause thickness variations along the bead

Key Questions and Reflections

The study successfully demonstrates the applicability of fuzzy logic for surfacing quality evaluation, but several questions remain for practical implementation:

1. Subjectivity in membership function definition: The membership functions are based on "practical welding experience," which introduces subjectivity. Different experts may define different membership functions, leading to different evaluation results. Standardization of membership function definitions is necessary for inter-laboratory comparability.
2. Dynamic quality assessment: The model evaluates a single surfacing operation but does not address the cumulative quality effects of multi-pass surfacing or the interaction between adjacent beads. Extension to multi-pass evaluation would be valuable for practical applications.
3. Integration with process parameters: The model evaluates quality outcomes but does not directly relate them to process parameters (welding current, voltage, travel speed, belt feed rate). Establishing this relationship would enable predictive quality assessment and proactive process control.
4. Scalability: The model is developed for four evaluation indices. In practice, additional quality characteristics (surface roughness, porosity, dilution ratio, hardness) may need to be included. The model's scalability to handle additional indices without excessive complexity is an important consideration.

Study Insights and Implications

The fundamental contribution of this research is the development of a systematic, mathematically rigorous framework for evaluating surfacing quality that accounts for the inherent uncertainty and subjectivity in quality assessment. The FAHP methodology, implemented through the Matlab-FIS platform, provides a transparent and reproducible evaluation approach that can be adapted to different surfacing applications and quality requirements. For engineers involved in surface engineering quality control, this research demonstrates that fuzzy logic provides a powerful tool for bridging the gap between quantitative measurements and qualitative quality judgments, enabling more objective and consistent quality decisions in surfacing operations. The successful validation of the model against actual observations provides confidence in its practical applicability and supports its adoption in industrial quality management systems.