Microcomputer Fuzzy Control in TIG Inverter Power Source - Technical Study Note

Literature Overview

Published in Electronic Technology Application in 2000, this paper by Wang Jianping from Shenzhen University's School of Engineering Technology presents the application of microcomputer-based fuzzy control technology to TIG inverter power sources. The study addresses the challenge of achieving precise and stable arc control in TIG welding through advanced power electronics and control algorithms. This research was particularly significant during the early 2000s when inverter welding power sources were transitioning from analog to digital control, and fuzzy logic control represented an emerging approach to handling the nonlinear dynamics of the welding arc.

Core Technical Architecture

The TIG inverter power source with microcomputer fuzzy control comprises several integrated subsystems:

Power Conversion Stage

The inverter topology converts DC input voltage to high-frequency AC output suitable for welding arc generation:

Component	Function	Typical Specification
Rectifier	AC to DC conversion	3-phase bridge, 380V input
DC Link Capacitor	Energy storage and filtering	400-800V DC, 2000-4000 μF
IGBT Inverter	DC to high-frequency AC	20-50 kHz switching frequency
Output Transformer	Voltage transformation and isolation	Turns ratio 1:5 to 1:10
Output Rectifier	AC to DC welding current	Fast recovery diodes

Control System Architecture

The microcomputer fuzzy control system operates through the following hierarchy:

Sensing Layer: Current transducers, voltage sensors, and arc voltage monitors provide real-time feedback.
Processing Layer: Microcomputer (typically 8051 or similar MCU) executes fuzzy control algorithms.
Decision Layer: Fuzzy inference engine processes input variables and generates control outputs.
Actuation Layer: PWM (Pulse Width Modulation) signals drive the IGBT inverter switches.

Fuzzy Control Algorithm

The fuzzy control approach maps linguistic variables to control actions through a series of steps:

Step 1: Fuzzification

Input variables: Arc voltage error (e), error rate of change (de/dt)
Membership functions: Triangular or Gaussian distributions
Linguistic terms: Negative Large (NL), Negative Medium (NM), Zero (Z), Positive Medium (PM), Positive Large (PL)

Step 2: Rule Base

The fuzzy rule base contains IF-THEN rules that encode expert knowledge:

Rule No.	IF Arc Voltage Error	AND Error Rate	THEN Control Output
1	NL	NL	PL
2	NL	Z	PM
3	Z	NL	NM
4	Z	Z	Z
5	PL	PL	NL
6	PL	Z	NM

Step 3: Inference

Mamdani or Sugeno fuzzy inference methods
Defuzzification using centroid or mean-of-maxima methods

Step 4: Output Scaling

Convert fuzzy output to PWM duty cycle or frequency modulation signal

Performance Characteristics

The fuzzy control system demonstrates several advantages over conventional PID control:

Performance Metric	PID Control	Fuzzy Control	Improvement
Arc Voltage Stability (±V)	±2.5	±0.8	68% reduction
Current Regulation Time (ms)	15-20	5-8	60% faster
Arc Restart Reliability (%)	85-90	95-98	10-13% improvement
Current Overshoot (%)	15-20	5-8	50% reduction
Disturbance Rejection (dB)	20	35	15 dB improvement

Dynamic Response Analysis

The fuzzy control system exhibits superior dynamic response characteristics:

Step Response: For a 50% change in welding current demand, the fuzzy controller achieves settling within 10 ms compared to 25 ms for PID control.
Disturbance Rejection: Sudden changes in arc length (due to electrode wear or operator movement) are compensated within 3-5 ms, maintaining arc stability.
Soft Start/Stop: The fuzzy controller provides smooth current ramping at weld initiation and termination, reducing spatter and improving bead quality.

Engineering Implementation Considerations

Hardware Design

The microcomputer-based fuzzy control system requires careful hardware design:

Component	Specification	Design Consideration
Microcontroller	8051/80C196KC	40 MHz clock, 32KB Flash
ADC	12-bit, 100 kS/s	Fast sampling for arc voltage
PWM Generator	20-50 kHz	High frequency for switching
Current Sensor	Hall effect, 0-500A	Linear response, low drift
Power Supply	5V/12V isolated	Digital and analog separation

Software Implementation

The fuzzy control algorithm is implemented in assembly or C language with the following memory requirements:

Memory Section	Size	Content
Program Memory	8-16 KB	Control algorithm, rule base
Data Memory	2-4 KB	Variables, look-up tables
Timer Interrupt	1-2 ms period	Control loop execution

Protection Features

The control system incorporates multiple protection mechanisms:

Overcurrent Protection: Instantaneous cutoff at 120% of rated current
Overvoltage Protection: DC link voltage limit at 900V
Overtemperature Protection: IGBT junction temperature limit at 125°C
Arc Loss Detection: Automatic arc restart within 50 ms
Soft Start: Current ramping over 200-500 ms at weld initiation

Applications in Cladding and Bimetal Fabrication

The fuzzy-controlled TIG inverter power source offers several advantages for cladding and bimetal pressure vessel fabrication:

Precision Current Control

For weld overlay cladding, precise current control is essential for:

Maintaining consistent dilution rates across multiple cladding layers
Achieving uniform bead geometry for subsequent grinding
Minimizing thermal distortion in thin-section cladding

The fuzzy controller's fast response (5-8 ms) enables real-time adjustment of current based on arc voltage feedback, compensating for variations in workpiece geometry, material properties, and shielding gas flow.

Multi-Process Capability

Modern inverter power sources with fuzzy control can support multiple welding modes:

Mode	Current Range	Application
DC TIG	20-300 A	Cladding, thin-section welding
AC TIG	20-200 A	Titanium welding, oxide removal
Pulse TIG	50-200 A (pulse)	Reduced heat input, narrow beads
Manual TIG	20-300 A	Operator-controlled welding

The pulse TIG mode is particularly valuable for cladding applications where minimal heat input is required to reduce dilution and distortion.

Integration with Automated Systems

The digital control architecture facilitates integration with automated welding systems:

Seam Tracking: Arc voltage feedback can be used for basic seam tracking in robotic cladding
Parameter Programming: Pre-set welding parameters for different cladding layers
Monitoring and Data Logging: Real-time recording of welding parameters for quality traceability
Remote Control: Ethernet or fieldbus communication for integration with manufacturing execution systems

Key Reflections and Study Insights

This research represents a significant milestone in the evolution of welding power source technology, transitioning from analog control to digital fuzzy control. The application of fuzzy logic to welding arc control addresses the inherent nonlinearity and time-varying nature of the welding process, which challenges conventional linear control approaches.

However, several limitations and challenges remain:

Rule Base Development: The effectiveness of fuzzy control depends heavily on the quality of the rule base, which requires extensive expert knowledge and experimentation to develop.
Computational Resources: Early microcomputers had limited processing power, constraining the complexity of fuzzy algorithms that could be implemented in real time.
Scalability: As welding power sources increased in capacity (from 200A to 1000A+), the control challenges became more complex, requiring more sophisticated algorithms.

From a pressure vessel fabrication perspective, the precision and reliability offered by fuzzy-controlled TIG power sources are essential for meeting the stringent quality requirements of codes such as ASME VIII and NB/T 47002. The ability to maintain consistent welding parameters throughout multi-pass cladding operations directly impacts the metallurgical quality of the overlay layer and the overall integrity of the bimetal pressure vessel.

Summary

The study by Wang Jianping demonstrates the successful application of microcomputer-based fuzzy control to TIG inverter power sources, achieving significant improvements in arc stability, dynamic response, and process reliability. This technology represents a fundamental advancement in welding power source design, enabling the precision control required for high-quality cladding and bimetal pressure vessel fabrication. The principles established in this research continue to influence modern welding power source development, with contemporary systems employing more sophisticated digital signal processing and adaptive control algorithms built upon the foundational concepts of fuzzy logic control.