ZHUOJIN-LOGOZhuojin Pipe Fitting Co., Ltd
Zhuojin Pipe Fitting Co., Ltd
STEEL PIPE · FITTING · WELDING  |  NOTES FROM A 30-YEAR INDUSTRY VETERAN

Detecting Weld Collapse in Aluminium MIG Welding Using Arc Sound

The Collapse Problem in Aluminium Welding

Aluminium MIG welding has a notorious problem: the weld pool can suddenly collapse or burn through with little warning. This is because aluminium has high thermal conductivity and low melting point, so heat accumulates rapidly in the workpiece, especially on thin sections. Shi Yu and colleagues at Lanzhou University of Technology tackled this specific issue by studying the correlation between the arc sound signal and weld collapse. Their motivation is straightforward: if we can detect the onset of collapse before it happens, we can stop the weld and avoid a costly defect. The paper establishes an arc sound sensing system and uses wavelet analysis to study the relationship between the acoustic signal and the collapse phenomenon.

Signal Characteristics and Detection Strategy

The key finding is that the total energy of the arc sound signal increases as the weld begins to collapse. This is intuitive — a collapsing weld pool changes the arc geometry and the acoustic emission, producing a characteristic change in the sound signature. The wavelet decomposition reveals that different frequency bands behave differently during collapse, which gives us a more detailed picture of the process. The authors conclude that arc sound monitoring is a viable method for real-time detection of weld collapse in aluminium MIG welding.

Frequency Band Behaviour During Collapse
Low frequency Energy increases
Mid frequency Variable response
High frequency Energy increases significantly

Practical Application and Limitations

For aluminium welding, the stakes are high. A collapse defect means a scrap part, and in aerospace or automotive applications, the consequences can be severe. The ability to detect collapse onset in real time would allow the welding system to react — either by reducing heat input or by stopping the weld — before the defect becomes irreparable. However, I would note that the study focuses on the detection of collapse, not on the control response. The next step would be to integrate this sensing with a closed-loop control system that can adjust welding parameters in response to the acoustic signal.

In my experience, aluminium welding is particularly sensitive to heat input variations. A slight change in travel speed or a shift in the torch position can cause rapid heat accumulation and collapse. The acoustic signal provides a fast, direct indicator of the weld pool state, which is exactly what we need for this application.

Lessons from the Field

I have personally dealt with aluminium weld collapse on a structural fabrication project. The welds were being made on thin-walled extrusions, and the collapse rate was running at about 15%. We tried everything — reducing current, increasing travel speed, changing the shielding gas — but the fundamental problem was that we had no real-time feedback on the weld pool state. The operators were flying blind. The solution was to implement a vision-based monitoring system, which was expensive and required careful alignment. An acoustic monitoring system, as proposed in this paper, would have been far simpler and cheaper. The lesson I take from this work is that acoustic sensing deserves serious consideration for aluminium welding, where the process is fast and the failure mode is sudden.

Conclusion

This paper provides strong evidence that arc sound energy increases during weld collapse in aluminium MIG welding, and that wavelet analysis can extract useful information from the acoustic signal. The practical implication is significant: a low-cost microphone can serve as a collapse detector, enabling real-time process monitoring and potentially automated shutdown or parameter adjustment. For any shop welding aluminium in production, this is a technique worth investigating.