Zhuojin Pipe Fitting Co., LtdMicrocomputer Fuzzy Control Applied to TIG Inverter Power Supply
Literature Overview
This paper, published in Electronic Technology Application in 2000, presents a novel control strategy for TIG welding inverter power supplies using fuzzy logic control implemented on a microcontroller. The approach addresses the inherent challenges of maintaining stable arc characteristics in TIG welding by employing a dual-input single-output fuzzy controller that regulates arc current in real time. This represents an early application of intelligent control theory to welding power source design.
Core Technical Content
Control Architecture
The system architecture comprises:
- Real-time arc current detection and feedback
- Fuzzy controller with dual inputs: arc current deviation (e) and rate of change of deviation (ec)
- Single output: phase-shift PWM duty cycle adjustment voltage
- Microcontroller-based implementation with real-time computational capability
| Component | Function | Specification |
|---|---|---|
| Input 1 (e) | Arc current deviation from setpoint | Real-time measurement |
| Input 2 (ec) | Rate of change of current deviation | Differentiated signal |
| Output | Phase-shift PWM duty cycle voltage | Controls inverter power |
| Controller | Fuzzy logic inference | Rule-based decision making |
| Platform | Microcontroller (MCU) | Embedded real-time processing |
Fuzzy Controller Design
The fuzzy controller operates through the following steps:
- Fuzzification: Converting crisp input values (e, ec) into fuzzy linguistic variables using membership functions.
- Rule base evaluation: Applying IF-THEN rules to determine appropriate control action.
- Defuzzification: Converting fuzzy output into a crisp control voltage for PWM modulation.
The dual-input approach (current deviation plus its rate of change) provides the controller with information about both the current error magnitude and the trend, enabling proactive rather than purely reactive control. This is analogous to a PID controller's P and D terms but implemented through fuzzy logic rules that can handle nonlinearities and imprecise system models.
Application to TIG Welding
TIG welding presents unique control challenges:
- Arc stability is sensitive to electrode protrusion, gas flow rate, and workpiece geometry
- Arc length variations cause significant current fluctuations
- Different welding positions and joint configurations require different current levels
- Arc blow effects in certain configurations require adaptive compensation
The fuzzy controller's ability to handle these nonlinearities without requiring an exact mathematical model of the welding process makes it particularly suitable for TIG applications where process variability is inherent.
Engineering Practice Integration
Comparison with Conventional Control
| Control Method | Advantage | Limitation |
|---|---|---|
| PID control | Well-understood, widely available | Requires accurate tuning, limited to linear systems |
| Fuzzy control | Handles nonlinearities, robust to parameter variation | Rule design requires expertise, less transparent |
| Adaptive control | Self-adjusting to changing conditions | Complex implementation, computational demands |
| Fuzzy-PID hybrid | Combines both strengths | Increased complexity |
Practical Implementation Considerations
- Sampling rate: The microcontroller must sample arc current at sufficient rate (typically > 1 kHz) to capture arc dynamics.
- Rule base design: Rules must be derived from expert knowledge of TIG welding behavior across different conditions.
- Anti-windup protection: Integral-like effects in the fuzzy rules must be managed to prevent control saturation.
- Noise filtering: Arc current signals contain high-frequency noise from arc oscillation that must be filtered before fuzzy processing.
- Safety interlocks: The fuzzy controller must operate within hard limits to prevent dangerous overcurrent conditions.
Impact on Weld Quality
The implementation of fuzzy control in TIG power supplies contributes to:
- More consistent arc length maintenance
- Reduced spatter and arc instability
- Improved weld bead uniformity
- Better adaptability to varying welding conditions
- Potential for reduced operator skill requirements
Study Insights and Reflections
This paper represents an early and significant contribution to intelligent welding power source control. The application of fuzzy logic to TIG arc current regulation demonstrates the potential of rule-based intelligent control in manufacturing processes where mathematical models are incomplete or difficult to establish. While the paper is now over two decades old, the fundamental approach remains relevant—modern welding power sources increasingly incorporate adaptive and intelligent control strategies. The dual-input fuzzy controller design provides a template that can be extended to multi-parameter control (current, voltage, travel speed) for comprehensive welding process optimization. The key challenge remains translating expert welding knowledge into effective fuzzy rules—a task that requires both metallurgical understanding and control engineering expertise.