Machining parameters inversion and regularity analysis based on surface topography point cloud of GH4169 alloy

Machining parameters directly influence the formation of surface topography. However, the surface topography is susceptible to partial distortion due to the interaction of various external disturbances, making it challenging to establish the mapping relationship between surface topography and machining parameters. PointNet++ can be utilized to develop a model for inverting machining parameters from surface topography using preprocessed point cloud data, thereby providing a research foundation for intelligent manufacturing. The test results of this model can be analyzed for regularity to study the results of different machining parameters on the surface topography. The GH4169 alloy is used as the study object. The mechanisms of the cutting depth, feed per tooth, cutting speed, and their orthogonal combinations are systematically investigated to reveal the parameter interactions and characteristic regularity. The dynamic relationship among data dimensionality, data density, model accuracy, and convergence speed is investigated by comparing multiple training and testing results. The robustness of PointNet++ in capturing surface topographic features is demonstrated. The output of this research provides theoretical and technical support for machining parameter inversion and a data-driven basis for high-quality surface machining in intelligent manufacturing.