The “Smart and Lean Production Hairpin 4.0” project was conducted by Tecnomatic and the Department of Computer Science of UNIVPM – Università Politecnica delle Marche with the aim of studying the application of Artificial Intelligence (AI) in vision systems for post-weld inspection, with particular reference to the evaluation of hairpin welding quality in the stator.
Tecnomatic specializes in the design, development and implementation of automation and assembly systems. Building on active research and extensive experience, Tecnomatic has embraced electrification as an opportunity for growth and innovation. Since 1998, various studies on copper conductors with rectangular cross-section have provided considerable experience in hairpin technology.
Continuing its commitment to the advancement of the industry, Tecnomatic actively collaborates with research institutes and universities to promote the development of innovative projects including in the field of artificial intelligence, with a focus on improving the quality of products and production processes.
The Project
In 2020, Tecnomatic launched in collaboration with the Department of Computer Science of UNIVPM – Polytechnic University of Marche the Smart Hairpin project to explore the application of Artificial Intelligence (AI) in vision systems for post-weld inspection, with particular reference to the evaluation of the quality of hairpin welding in the stator.
Speaking of AI and solder flow, in a fork stator assembly line, the soldering process follows the twisting phase. During twisting, the pins are aligned and twisted into place. Parameter setting is critical to ensure weld quality and reliability, and here, the role of the operator setting the weld “recipe” is crucial.
The role of artificial intelligence
In the initial vision phase and during final inspection, so both before and after welding, the role of AI is key. By analyzing the welding data, AI creates a feedback loop that identifies the optimal parameters for future prototypes. The data collected during the initial vision process is processed, and the information obtained defines the settings for the final inspection.
In addition, artificial intelligence guides the laser beam and records the data, determining where the welder should perform the welding. Using tomography, the AI correlates the visual data with predetermined quality standards. During this phase, the AI focuses primarily on downstream control, ensuring a data-driven approach that increases the accuracy and continuous improvement of the welding process.
For high reliability of the fork stator assembly
Before welding, AI systems analyze historical data to recommend the best initial parameters. This ensures that the process starts with the settings that previously resulted in high-quality welds. After welding, however, artificial intelligence evaluates the results to refine and adjust the parameters. This continuous optimization improves the quality and efficiency of the welding process, contributing to the overall reliability of the fork stator assembly.
By Andrea Ciani and Martina Vallese
