Study On Three-dimensional Characterization And Prediction Model Of Surface Roughness Of Wood Sawing

Surface roughness is an important indicator for evaluating the surface quality of wood,which affects the surface state,appearance quality,decorative effect,friction and wear,and coating performance of wood products.Wood is fiber anisotropic material,and its machining surface quality is greatly affected by its own structure and cutting system parameters.In view of the problems that using surface roughness profile parameters to characterize the machining surface quality of wood can only express the profile height characteristics,and the measurement results depend on the fiber structure and measurement direction,taking the diamond woodworking circular saw blade as an example,fractal dimension is introduced to characterize the three-dimensional shape of the wood sawing surface.On this basis,a prediction model of wood sawing surface roughness considering the structural characteristics and cutting system parameters is established,which provides new ideas and tools for the three-dimensional characterization and prediction of wood processing surface quality.Based on the measurement standard of wood surface roughness,the influence of wood fiber direction and measurement direction on the measurement value of wood surface topography is studied and analyzed through the sawing experiment of diamond circular saw blade,and the measurement method considering the measurement direction is put forward;The single factor experiment of saw cutting shows that the speed and feed speed of saw blade can cause the change of Ra in different measuring directions.The fluctuation amplitude decreases with the increase of saw blade speed,and increases with the increase of feed speed;Based on six kinds of wood orthogonal experiments,it is found that there are differences in the range of Ra value fluctuation in different fiber direction and measurement direction,and it is proved that the change of Ra will affect the accuracy and application range of multiple linear regression model.The difference box dimension method is used to demonstrate the "gray-scale roughness" of the wood sawing surface image can truly reflect the roughness changes of micro-topography such as irregularities,irregularities of the wood sawing surface and verify the feasibility of fractal dimension to characterize the three-dimensional morphology of wood sawing surface through theoretical analysis and calculation;with the aid of the gray value algorithm of the reward window to improve the gray value utilization in the image window and optimize the accuracy of the difference box dimension algorithm;on this basis,the improved differential box dimension algorithm is used to calculate the fractal dimension of six kinds of wood sawing surfaces,and variance analysis and Duncan's test are carried out on the experimental results,which shows that the fractal dimension can effectively represent the three-dimensional shape of wood sawing surface,and the fractal dimension has a high correlation with the sawing surface quality.According to the regression principle of support vector machine,the fractal dimension of different cutting speed,feed speed,cutting direction and wood density is obtained by PCD saw blade cutting wood test,writing an algorithm program,normalized processing of test data,training the model with training set data,and optimize the model parameters by grid search method(penalty factor C and kernel function parameter g);Using the prediction optimization model to predict the fractal dimension,the results show that the wood sawing surface roughness prediction model after parameter optimization can more accurately reflect the non-linear problems between various factors in the wood sawing process,and achieve accurate prediction of the surface roughness.Model regression performance evaluation indicators mean square error(EMS)and determination coefficient(R2)meet the model accuracy requirements,which can meet production needs.