Abstract: Aiming at the problem that it is difficult to predict the surface roughness of the machined object effectively and accurately due to the comprehensive influence of the surface state of grinding tools and process parameters, an intelligent prediction method of surface roughness based on the integration of abrasive characteristics and processing parameters is proposed. Based on the collected diamond abrasive images on the tool surface, three key characteristic information of abrasive number per unit area, abrasive distribution uniformity and abrasive protrusion height on the tool surface are extracted by using image processing technology, and the effectiveness of the extracted characteristic information is verified. On this basis, a surface roughness intelligent prediction algorithm integrating abrasive characteristics and process parameters is proposed, the milling experiment of YG8 cemented carbide is carried out, and the machining results are compared with the predicted results. The results show that the proposed intelligent algorithm can greatly improve the accuracy and stability of surface roughness prediction, and its prediction accuracy can reach 95.6%, while the accuracy of regression model based on traditional process parameters is only 86.3%, which provides a reference for the intelligent prediction of surface roughness in abrasive machining.

Key words: abrasive machining, image processing, characteristic information, intelligent algorithm, surface roughness