关键词: 激光切削, 水曲柳, 相关性, 回归性

Abstract:

Objective: By analysing the technical parameters and cutting effect of laser cutting Fraxinus mandshurica, the optimal regression equation model of cutting effect was established to provide a theoretical basis for estimating the cutting effect. Method: Taking the technical parameters of the laser machine lens height, cutting speed and light intensity as the influencing factors, and the seam depth and seam width as the cutting effect indicators, the two were analysed by Spearman correlation analysis using SPSS 27.0 software, and further multivariate linear and non-linear regression analyses were carried out by using MATLAB R2020a software programming based on the principle of the least squares method. Result: 1) The lens height was significantly correlated with the seam depth and width, with correlation coefficients of −0.677 and 0.962, indicating that the lens height was negatively correlated with the seam depth and positively correlated with the seam width. The cutting speed was significantly correlated with the seam depth and not with the seam width, with correlation coefficients of −0.619 and −0.090, indicating that the correlation was negatively correlated with the seam depth and seam width. The light intensity was not significantly correlated with the seam depth and seam width, with correlation coefficients of 0.116 and 0.057, indicating a positive correlation with the seam depth and seam width. 2) Linear regression analysis showed that the goodness-of-fit R² of the regression model for seam depth was 0.771 54 (P<0.01). The goodness-of-fit R² of the regression model for seam width was 0.904 58 (P<0.01). Non-linear regression analyses yielded a goodness-of-fit R² of 0.936 69 (P<0.01) for the regression model of seam depth. The a goodness-of-fit R² of 0.942 41 (P<0.01) for the regression model of seam width. Both multivariate linear and non-linear regression models were well fitted, but the accuracy of the multivariate nonlinear regression model was relatively higher than that of the multivariate linear regression model. 3) From the comparison of the coefficients of the multiple regression model equations and the magnitude of the image changes, it can be obtained that the influence of the lens height on the seam depth and width is greater than that of the cutting speed and light intensity. 4) In the actual laser cutting Fraxinus mandshurica production, if the values of seam depth and seam width are only roughly and quickly calculated, the multiple linear regression model can be used for estimation; on the contrary, if the values of seam depth and seam width are calculated more accurately, the estimation effect will be better if the multiple nonlinear regression model is used for estimation. Conclusion: 1) Seam depth and seam width generally show a periodic change law with the combination of laser technical parameters. 2) In the production of laser cutting Fraxinus mandshurica, it is necessary to adjust the lens height first. 3) The influence of the combination of laser parameters on the seam depth and seam width shows more significant non-linear characteristics.

Key words: laser cutting, Fraxinus mandshurica, correlation, regression