Acoustic Monitoring in the Process of Pulsed Laser Paint Removal

Real-time monitoring of the laser paint removal process is critical to realize a good cleaning effect without causing damages to the base. Acoustic signals when the laser pulse continuously acts at one point were studied through an experiment. Results show that peak values and maximum frequency component of acoustic signals decreased to a stable value with paint removal, reflecting the cleaning process. According to an experimental study on cleaning acoustic signals during laser scanning, the cleaning process is divided into paint removal, clean finished, and excessive cleaning according to the damages threshold. The acoustic signals of the different processes were analyzed. Results show that variations of frequency components could reflect the paint removal process more accurately than time-domain signals. Extracting characteristic parameters of the maximum frequency component, spectrum, spectrum centroid amplitude, and spectrum centroid frequency from the frequency domain; introducing in the discriminant analysis; and establishing a Bayesian discriminant model according to cleaning characteristics can realize the quantitative discrimination of the paint removal process.