Monitoring melted state of reinforced particle in metal matrix composite fabricated by laser melt injection using optical camera

Laser melt injection (LMI) is a promising technique for the fabrication of particle reinforced metal matrix composites (MMCs), in which process monitoring is highly demanded to ensure reliability. The objective of this work is to study the feasibility of using observed spatter and molten pool features for predicting the "invisible" reinforced particle melted state to explore the potential for in situ optical monitoring during LMI. To accomplish this, an in situ optical monitoring system was established and data-driven models were developed based on the analysis of the physical process and image signal formation mechanism in LMI. The image features had distinct behavioral characteristics at different reinforced particle melted states. Meanwhile, the different particle melted states determine the forming quality of the MMCs. The extracted particle spatter and droplet spatter features were proved to be significantly correlated with the particle melted state based on the correlation assessment; thus, the highly correlated spatter feature vectors were used as the input for the classification model. The test results showed that the overall classification accuracy of the prediction model has a high level from 85 to 95%, which illustrated the good generalization ability and robustness of the prediction model. The potential of inferring forming quality of MMCs based on image features is validated through the optical in situ monitoring system. This work contributed to the in-depth understanding of the LMI process and the further applications in process monitoring.