CLADDING TECHNOLOGY SHANXI CO., LTDMIG Welding Power Source Design Based on Fuzzy and PI Control
Literature Overview
Published in Control Engineering in 2012 by Pang Qingle and Wang Yongqiang from Shandong Institute of Business and Technology, this paper presents a hybrid control strategy combining fuzzy logic and proportional-integral (PI) control for MIG welding power sources. Supported by the National Natural Science Foundation (Grant No. 50777040), the China Postdoctoral Science Foundation (Grant No. 20090461204), and the Shandong Provincial Natural Science Foundation (Grant No. ZR2010EL030), this research addresses the challenge of achieving both robustness and precision in welding power source control.
Core Technical Content
Traditional PI controllers for welding power sources provide good steady-state performance but suffer from poor transient response when process disturbances occur. Pure fuzzy logic controllers offer excellent adaptability to nonlinearities but may exhibit oscillatory behavior around the setpoint. The hybrid approach proposed in this paper combines the strengths of both methods: fuzzy logic handles large deviations and process nonlinearities, while the PI controller provides precise regulation near the setpoint.
Control Architecture
The hybrid control system operates as follows:
- Error detection: The difference between the measured arc voltage and the setpoint is continuously monitored.
- Fuzzy reasoning: When the error exceeds a threshold (typically ±2V), the fuzzy controller determines the appropriate control action based on predefined membership functions and rule base.
- PI regulation: When the error is within the threshold, the PI controller takes over to provide precise regulation with minimal overshoot.
- Smooth transition: A switching mechanism ensures seamless handover between fuzzy and PI control modes to prevent control instability.
Fuzzy Controller Design
| Linguistic Variable | Membership Functions | Range |
|---|---|---|
| Error (e) | NB, NM, NS, ZO, PS, PM, PB | -5V to +5V |
| Error rate of change (de/dt) | NB, NM, NS, ZO, PS, PM, PB | -10V/s to +10V/s |
| Control output (u) | NB, NM, NS, ZO, PS, PM, PB | -100% to +100% |
The fuzzy rule base contains 49 rules (7×7) that map the input linguistic variables to the output control action. The most critical rules for welding applications include:
- IF error is PB AND error rate is PS, THEN output is PB (large positive error with increasing trend requires maximum correction)
- IF error is ZO AND error rate is ZO, THEN output is ZO (system at setpoint, maintain current state)
- IF error is NS AND error rate is NB, THEN output is NM (small negative error with rapidly decreasing trend requires moderate positive correction to prevent overshoot)
PI Controller Parameters
The PI controller parameters are tuned for each welding application using the Ziegler-Nichols method as a starting point, followed by manual optimization:
| Application | Kp | Ki (1/s) | Response Time | Overshoot |
|---|---|---|---|---|
| Root welding, carbon steel | 2.5 | 4.0 | 50 ms | <2% |
| Fill welding, stainless steel | 2.0 | 3.0 | 80 ms | <3% |
| Overlay welding, nickel alloy | 1.5 | 2.0 | 100 ms | <2% |
| Thin gauge aluminum | 3.0 | 5.0 | 30 ms | <1% |
Performance Evaluation
The paper presents comparative performance data between the hybrid controller and conventional PI-only control:
| Performance Metric | PI Only | Fuzzy Only | Hybrid (Fuzzy + PI) |
|---|---|---|---|
| Settling time (ms) | 120 | 80 | 45 |
| Maximum overshoot (%) | 5.2 | 3.1 | 1.8 |
| Steady-state error (V) | 0.1 | 0.3 | 0.05 |
| Disturbance rejection (s) | 0.8 | 0.4 | 0.2 |
| Arc stability index | 0.72 | 0.85 | 0.94 |
The hybrid controller demonstrates superior performance across all metrics, particularly in disturbance rejection and arc stability. The improved arc stability translates directly to better weld quality through more consistent heat input and droplet transfer.
Application to Cladding and Overlay Welding
For weld-overlay cladding operations, the control precision offered by the hybrid fuzzy-PI approach has direct implications for overlay quality:
Dilution Control
The dilution ratio in the first overlay layer is highly sensitive to arc voltage fluctuations. A ±0.5V variation in arc voltage can cause a ±5% change in dilution. The hybrid controller's ability to maintain arc voltage within ±0.1V provides the precision necessary to achieve consistent dilution across the entire overlay area.
Multi-Layer Overlay Sequence
In multi-layer overlay welding, the process conditions change significantly between layers:
- First layer: High dilution, deep penetration, requires aggressive control
- Intermediate layers: Moderate dilution, transitional conditions
- Final layer: Low dilution, shallow penetration, requires precise surface finish
The hybrid controller can be programmed with different control parameters for each layer, automatically transitioning as the welding sequence progresses.
Quality Assurance Integration
| Quality Parameter | Control Strategy | Acceptance Criteria |
|---|---|---|
| Arc voltage stability | Hybrid fuzzy-PI control | ±0.2V from setpoint |
| Wire feed speed consistency | Feedback-controlled motor | ±2% of setpoint |
| Contact tip-to-workpiece distance | Arc voltage feedback | 8–12 mm (solid wire) |
| Heat input per unit length | Current × voltage / speed | Within procedure limits |
Engineering Practice Considerations
Implementing hybrid fuzzy-PI control in production welding environments requires attention to several practical aspects:
- Power source hardware: The control algorithm requires a digital signal processor (DSP) or microcontroller operating at minimum 10 kHz sampling rate to capture the full arc voltage waveform.
- Sensor quality: Arc voltage and current sensors must have bandwidth exceeding 5 kHz to provide meaningful feedback for the control loop.
- Parameter calibration: The fuzzy rule base and PI parameters must be calibrated for each specific welding application, material combination, and joint configuration.
- Safety interlocks: The control system must include hardwired safety limits that cannot be overridden by software, ensuring that current and voltage never exceed equipment ratings.
Study Insights and Future Directions
This research represents a significant step toward intelligent welding power sources that can maintain optimal process conditions regardless of external disturbances. For the cladding and overlay welding community, the implications are profound: consistent dilution control, uniform overlay composition, and predictable mechanical properties can all be achieved through superior power source control.
The hybrid approach also opens the door to more advanced applications such as:
- Real-time dilution monitoring through arc voltage analysis combined with material-specific models
- Adaptive multi-layer control that automatically adjusts parameters between overlay layers
- Predictive control that anticipates process changes based on welding position and geometry
The fundamental insight from this work is that welding quality is not just a function of the welding parameters selected but of how precisely those parameters are maintained throughout the welding process. In applications where the cost of failure is high — such as pressure vessel cladding for hydrogen service or nuclear power plant components — the investment in advanced power source control technology is fully justified by the improved reliability and reduced rework rates that result.
This paper demonstrates that combining different control philosophies can overcome the inherent limitations of any single approach, providing a powerful framework for developing the next generation of welding equipment for critical fabrication applications.