Design of MIG Welding Arc Acoustic Signal Acquisition and Analysis System

Literature Overview

This 2010 paper published in the Journal of Harbin University of Science and Technology presents the design and development of an acoustic signal acquisition and analysis system for monitoring MIG welding arcs. The research was conducted at Northeast Forestry University in collaboration with Harbin University and Harbin University of Science and Technology, supported by the Heilongjiang Provincial Natural Science Foundation and other funding sources. The study addresses an important aspect of welding process monitoring: the use of acoustic signals as a non-intrusive means of assessing arc behavior and weld quality.

Core Technical Content

Acoustic monitoring of welding arcs is a non-intrusive technique that captures the sound emitted by the arc during welding. The acoustic signal contains information about arc stability, spatter generation, gas flow characteristics, and potential defects. By analyzing the frequency spectrum, amplitude, and temporal patterns of the acoustic signal, it is possible to detect process anomalies in real time and improve weld quality.

The acoustic signal from a MIG welding arc typically contains a broadband component (50 Hz-20 kHz) with dominant frequencies in the 1-10 kHz range. The signal amplitude and spectral content vary with welding parameters such as current, voltage, travel speed, and shielding gas flow rate. Process anomalies such as arc wandering, spatter bursts, and porosity formation produce characteristic changes in the acoustic signal that can be detected and classified.

Signal Acquisition System Design

The system design follows a modular architecture with separate stages for signal capture, conditioning, digitization, and analysis. The microphone is positioned at a fixed distance from the welding arc (typically 100-300 mm) to ensure consistent signal capture while avoiding damage from spatter and heat. The pre-amplifier provides gain adjustment to optimize the signal-to-noise ratio for the specific welding conditions.

The analog-to-digital converter samples the acoustic signal at a rate sufficient to capture the full frequency range of interest. A sampling rate of 44.1 kHz provides frequency resolution up to 22.05 kHz, which covers the entire audible range and the majority of the welding arc acoustic spectrum. The digitized signal is processed using fast Fourier transform (FFT) to extract the frequency spectrum, and wavelet transform to analyze the time-frequency characteristics of transient events.

The signal processing algorithms extract key features from the acoustic signal, including root mean square (RMS) amplitude, peak amplitude, spectral centroid, and spectral bandwidth. These features are correlated with welding parameters and weld quality indicators to develop a process monitoring model.

Signal Analysis and Process Monitoring

The acoustic signal analysis reveals distinct patterns associated with different welding conditions and process states. Stable arc conditions produce a consistent acoustic signal with narrow spectral bandwidth and stable RMS amplitude. Arc instability produces fluctuations in amplitude and spectral content, with increased high-frequency components. Spatter events produce transient bursts with broad spectral content and high peak amplitudes.