Fuzzy Control Inverter TIG Welding Power Supply Based on Microcontroller

Literature Overview

This research by Ai Sheng, Zhang Jian, and Ma Caixia from Northwestern Polytechnical University was published in Mechanical Science and Technology in 1999 and was supported by the Aviation Science Foundation. The study investigates the development of a fuzzy control inverter TIG welding power supply based on a microcontroller (single-chip microcomputer). This represents an early but significant contribution to intelligent welding power supply technology, addressing the need for adaptive control of welding parameters to maintain consistent weld quality under varying conditions.

Core Technical Concepts

The research addresses a fundamental challenge in TIG welding: maintaining consistent arc characteristics and weld quality despite variations in workpiece thickness, joint fit-up, welding position, and environmental conditions. Traditional constant-current or constant-voltage power supplies require manual adjustment of parameters, which is time-consuming and prone to operator error. The fuzzy control approach implements an intelligent control algorithm that automatically adjusts welding parameters in real-time based on feedback from the welding process.

Fuzzy logic control is particularly well-suited for welding applications because:

• It does not require a precise mathematical model of the welding process, which is inherently nonlinear and difficult to model accurately.
• It can handle imprecise sensor data and noisy measurements through linguistic variables and fuzzy sets.
• It can encode expert knowledge about welding in the form of fuzzy rules, making it accessible to experienced welders who may not have formal control theory training.
• It provides smooth, continuous control actions rather than the abrupt switching of conventional on/off controllers.

System Architecture and Control Strategy

The system architecture typically consists of the following components:

The fuzzy control algorithm operates by:

1. Fuzzification: Converting crisp sensor inputs (current, voltage, speed) into fuzzy linguistic variables (e.g., current is "low," "medium," or "high").
2. Rule evaluation: Applying a set of fuzzy rules (e.g., IF current is "low" AND voltage is "high" THEN increase current) to determine control actions.
3. Defuzzification: Converting the fuzzy control output into crisp values for the inverter PWM duty cycle.

The inverter TIG power supply itself operates by converting AC mains power to DC through a rectifier, then using a high-frequency switching circuit (typically IGBT-based) to generate a variable DC output. The switching frequency of 20–50 kHz allows for precise control of the output current and voltage with fast dynamic response.

Engineering Relevance to Cladding and Overlay Welding

The development of intelligent welding power supplies has direct relevance to cladding and overlay welding applications, where consistent overlay quality is critical. In overlay welding, the following parameters must be tightly controlled:

• Welding current: Affects penetration depth and dilution rate. Too high a current increases dilution, reducing the effectiveness of the overlay layer. Too low a current results in incomplete fusion and poor bond strength.
• Arc voltage: Affects weld bead width and profile. Inconsistent voltage leads to variable bead geometry and potential defects.
• Travel speed: Affects heat input per unit length. Too fast results in incomplete fusion; too slow leads to excessive dilution and distortion.

The fuzzy control approach can maintain optimal parameters by continuously monitoring arc characteristics and adjusting the power supply output accordingly. For example, if the workpiece thickness varies along the weld path (as in a tapered joint or a plate with thickness variation), the fuzzy controller can detect the change in arc voltage and adjust the current to maintain consistent penetration.

For pressure vessel fabrication, where overlay welding is used to provide corrosion-resistant linings on carbon steel vessels, the intelligent power supply offers several advantages:

• Reduced operator dependence: The system can maintain optimal parameters even with operator fatigue or inattention.
• Improved consistency: Automated parameter adjustment ensures uniform overlay quality across long welds.
• Faster cycle times: Elimination of manual parameter adjustments reduces non-welding time.
• Better documentation: The system can record parameter settings and adjustments for quality traceability.

Key Technical Challenges and Study Insights

The implementation of fuzzy control in welding power supplies faces several technical challenges:

• Sensor reliability: Arc voltage and current sensors must operate reliably in the harsh welding environment, with electromagnetic interference, high temperatures, and mechanical vibration.
• Rule development: The fuzzy rule base must be developed through extensive experimentation and expert knowledge, which is time-consuming and requires experienced welders.
• Computational requirements: Real-time fuzzy inference requires sufficient processing power, which was a constraint in 1999 but is no longer a limitation with modern microcontrollers.
• System stability: The feedback loop must be carefully designed to avoid oscillations or instability, particularly when the process dynamics change rapidly.

The study's contribution to the field is significant as an early demonstration of intelligent control in welding. The principles established in this research have evolved into modern adaptive welding systems that incorporate data analysis, neural networks, and advanced sensor fusion. However, the fundamental concept of using fuzzy logic to encode expert knowledge and implement adaptive control remains highly relevant and widely used in industrial welding applications.

For cladding engineers, the key takeaway is that intelligent power supply technology can significantly improve overlay welding quality by maintaining optimal parameters under varying conditions. This is particularly important for large-scale overlay applications where manual parameter control is impractical, and for applications where overlay quality directly impacts the service life and safety of pressure vessels and heat exchangers.