Real-time LIBS monitoring of laser-based layered controlled paint removal from aircraft skin based on random forest

Laser-induced breakdown spectroscopy (LIBS) is expected to be used for real-time monitoring and closed-loop control of laser-based layered controlled paint removal (LLCPR) from aircraft skin. However, the LIBS spectrum must be rapidly and accurately analyzed, and the monitoring criteria should be established based on machine learn-ing algorithms. Hence, this study develops a self-built LIBS monitoring platform for the paint removal process utilizing a high-frequency (kilohertz-level) nanosecond infrared pulsed laser and collects the LIBS spectrum dur -ing the laser removal process of the top coating (TC), primer (PR), and aluminum substrate (AS). After subtracting the spectrum's continuous background and screening the key features, we construct a classification model of three types of spectra (TC, PR, and AS) based on a random forest algorithm, and the real-time monitoring criterion based on the classification model and multiple LIBS spectra was established and verified experimentally. The results show that the classification accuracy is 98.89%, the time-consuming classification is about 0.03 ms per spectrum, and the monitoring results of the paint removal process are consistent with the macroscopic observation and micro-scopic profile analysis results of the samples. Overall, this research provides core technical support for the real-time monitoring and closed-loop control of LLCPR from aircraft skin. (c) 2023 Optica Publishing Group