Implementation of a Camera System for the Application of Artificial Intelligence in Monitoring the Cold Atmospheric Plasma Spray Process

Monitoring of the complex cold atmospheric plasma spraying process currently relies on technologically advanced systems, that often measure only a single parameter. This research introduces a more integrated approach, that uses a comprehensive set of parameters, consolidated in the characteristics of the plasma flame, is used to accurately classify particle temperature and coating materials. A simple optical camera system and image analysis based on a transfer learning model predicts with an 85% success rate. This approach simplifies and complements the monitoring process. Accuracy could be increased by using additional machine data already recorded, which could allow a complete replacement of additional the plasma flame monitoring.