Study On The LIBS Monitoring Of Laser Layer Accurate Paint Removal From Carbon Fiber Composite Skins

When repairing the carbon fiber reinforced polymer（CFRP）skin of an airplane,selective removal of specific layers of external coatings is required in order to protect functional coatings or underlying composite material structures.This is known as controlled layer-by-layer paint removal.In order to address the uncontrollable paint removal results from traditional techniques,research has been conducted on laser-controlled layer-by-layer paint removal technology.The reliability and controllability of laser-controlled layer-by-layer paint removal depend on effective online monitoring technology.During the laser paint removal process,laser-induced breakdown spectroscopy（LIBS）technology is used for online monitoring of the paint removal area to ensure complete removal of the coating layer and to avoid unnecessary damage to the material surface.Based on the established high-repetition-rate laser-induced breakdown spectroscopy（LIBS）online monitoring platform for paint stripping,this study investigates the evolution and interpretation of LIBS signals during the laser ablation process on the surface paint layers of CFRP aircraft tail fins.By adjusting the laser power,the depth control of paint stripping is achieved,and LIBS spectra excited by laser ablation of both the topcoat and primer are collected online.The characteristic peaks in the LIBS spectra are interpreted,and the Pearson linear correlation between the online LIBS spectra and reference spectra is compared.Discrimination and prediction models based on principal component analysis（PCA）and partial least squares（PLS）are established using 60 sets of online spectra to study the classification and discrimination of LIBS spectra during the laser ablation process and achieve laser ablation discrimination.The specific research content of this study is as follows:1.The laser power is adjusted using a single-factor variable method,and controlled laser ablation of the topcoat layer is achieved with an accuracy of approximately 3μm.Inductively coupled plasma emission spectroscopy tests reveal that Ti accounts for 98.24%of the detected metallic element mass in the topcoat,while Ba,Sr,Ti,and Cr in the primer account for 36.62%,27.73%,19.57%,and 12.75%of the detected metallic element mass,respectively,totaling96.67%.The main characteristic elements excited in the topcoat layer by laser are Ti,while Ba,Sr,Ti,and Cr are the main characteristic elements excited in the primer layer.2.Through the analysis of characteristic peaks in the LIBS spectra,it is found that the characteristic peaks at 357.48 nm（Cr I）,359.00 nm（Ba I）,360.13 nm（Cr I）,425.43 nm（Cr I）,427.46 nm（Cr I）,460.66 nm（Sr I）,478.35 nm（Sr I）,481.16 nm（Sr I）,483.13 nm（Sr I）,520.42nm（Cr I）,520.56 nm（Cr I）,520.79 nm（Cr I）,687.68 nm（Sr I）,and 689.18 nm（Sr I）can effectively discriminate the laser ablation layers.By comparing the differences in LIBS spectra at cleaning depths of 44.79μm and 51.50μm,it is found that the Sr I signal intensity at 460.66nm is high,the signal-to-noise ratio is high,and the peak is singular.Therefore,the characteristic peak at 460.66 nm（Sr I）is selected as the monitoring feature peak for laser ablation layers,which is used to monitor the layer boundaries of laser paint stripping on composite material tail fins.3.The Pearson correlation coefficient between the reference spectrum and the online spectrum is used for analysis.The Pearson correlation coefficient r₁ between the online laser paint stripping spectrum and the reference spectrum of white topcoat S_t decreases from 0.98 to0.87 as the laser cleaning depth increases.The Pearson correlation coefficient r₂ between the online laser paint stripping spectrum and the reference spectrum of yellow primer S_p increases from 0.80 to 0.89 as the laser cleaning depth increases.When r₁ decreases to 0.87 and r₂increases to 0.89,it reaches the layer boundary of laser paint stripping without damaging the primer.The laser ablation depth is approximately 44.79-51.50μm,which matches the actual sprayed thickness of 45μm for the white topcoat.This indicates that the Pearson correlation coefficient r can be used to determine laser ablation layers.4.Based on principal component analysis（PCA）and partial least squares（PLS）linear discriminant methods,relevant information is extracted from the spectra to identify the spectral features of the topcoat and primer.The first two principal components of the PCA model account for a cumulative contribution rate of 79.2%,while the first two principal components of the PLS-DA model account for a cumulative contribution rate of 85.5%.The root mean square error of estimation（RMSEE）of the PLS regression model is 0.142923,the root mean square error of cross-validation（RMSEcv）is 0.152053,and the root mean square error of prediction（RMSEP）is 0.142421.When predicting a mixed dataset of 20 laser-stripped topcoat and primer samples,the prediction accuracy reaches 100%.The results show that the PLS discriminant model performs better in classification and discrimination of paint layers based on LIBS spectra compared to the PCA model.The PLS prediction model exhibits good predictive accuracy in evaluating and classifying LIBS spectra of paint layers.The paper validates the intrinsic correlation between the controllability of layer-by-layer paint stripping and the changes in LIBS spectra through the evolution of characteristic peaks in LIBS spectra,Pearson linear correlation,and cluster analysis based on PCA and PLS.It demonstrates the feasibility of using LIBS technology for online monitoring and feedback in laser paint stripping.This study provides technical support for the automated and intelligent laser paint stripping process through LIBS online monitoring.