Fuzzy Control System for MIG Weld Bead Width

Literature Overview

This paper by Zhang Jiaying, Jiang Lipeng, and Zhang Xianghong from Beijing Institute of Armored Force Engineering and Beijing Institute of Petrochemical Technology presents a fuzzy control system for real-time regulation of weld bead width in MIG welding. Published in the Journal of Welding in 1999, Volume 20, Issue 1, pages 61-67, the work builds upon a dual-wavelength filter front-view real-time weld bead width detection technology.

System Architecture and Control Strategy

The fuzzy control system employs a dual-input, single-output fuzzy controller model. The two input variables are the weld bead width change and its rate of change, while the output variable is the welding current. The system uses dual-wavelength filter front-view detection to measure weld bead width in real time during the welding process.

Fuzzy Control Design Process

The design process involves fuzzification of the input variables, establishment of fuzzy control rules, and defuzzification of the output. The authors provide the fuzzy control state table and fuzzy control action table, which map the linguistic input variables to specific current adjustments. The measured data on the influence of welding current on weld bead width serves as the basis for the fuzzification process and quantification scheme.

The use of both the current deviation and its rate of change as input variables is analogous to the proportional and derivative terms in a PID controller. However, the fuzzy logic approach offers greater flexibility in handling nonlinear relationships and uncertain process conditions. The dual-wavelength filter technique for front-view detection is particularly effective because it can isolate the molten pool region from the surrounding solidified weld metal, providing accurate real-time measurements even in the harsh optical environment of an active welding arc.

Engineering Significance

This work represents an early application of fuzzy logic control in welding process automation. The approach is particularly valuable because weld bead width is a critical quality indicator that affects joint strength, penetration, and appearance. Real-time control of bead width can compensate for variations in joint fit-up, material properties, and environmental conditions that would otherwise lead to inconsistent weld quality.

For engineers in production welding environments, the key insight is that fuzzy control provides a robust solution for processes with inherent nonlinearities and uncertainties. Unlike traditional PID controllers, fuzzy logic can handle the complex, nonlinear relationship between welding current and bead width without requiring an exact mathematical model. This makes the approach particularly suitable for field conditions where process parameters may drift due to consumable wear, environmental changes, or material lot variations.

The study demonstrates that combining real-time optical sensing with intelligent control algorithms can significantly improve welding consistency, which is essential for applications demanding high reliability such as armor manufacturing and military equipment fabrication.