Study on generation of abrasive protrusion height based on projection information–driven intelligent algorithm

Abrasive protrusion height (APH) is the core parameter of abrasive tools, and it is also a principal parameter for modeling and simulation of the surface topography of abrasive tools. Because of the difficulty in obtaining the 3D feature information of abrasives and the limited measured data; this study proposes an APH model based on the spatial projection relationship. Combined with the gradient boosting decision tree (GBDT) intelligent algorithm, the APH characteristics of the abrasive tool are quickly generated by the data-fusion learning model. Through the corresponding experiments, we acquired the 2D image of the abrasive tool and extracted the area information of the abrasive projection area. The result shows that the fitting degree R² of the algorithm model reaches 0.911. Comparing the APH generated based on the algorithm model with the actual one, the average accuracy is about 94.69% and the mean absolute error of generated protrusion height MAE is 5.31 μm, which validates the proposed model. The results of this study demonstrate that the data-driven approach can effectively establish the generation model of the protrusion height of abrasive grains, which not only enable the measurement of the protrusion height of abrasive but also provide a reference for accurate modeling of abrasive tools.