Using TinyML and Sound Event Detection for Weld Anomaly Detection in Manufacturing

Learn about an innovative manufacturing project presented at the tinyML Summit 2023 that combines tinyML and sound event detection for identifying welding anomalies. Explore how Shiloh Industries partnered with Imagimob to develop acoustic detection methods for robotic gas metal arc welding defects, including porosity, burn through, and low deposition issues. Discover the complete project implementation process, from data collection and model building to hardware integration and PLC/SPS systems. Gain insights into how Large Language Models (LLM) accelerated the development process and understand the advantages of this approach over conventional electronic detection methods. Master the concept of continuous learning in manufacturing processes and learn about the specific workflows developed for improving quality control and efficiency in welding operations.