A robotic polishing parameter optimization method considering time-varying wear

In the automatic polishing process, the wear of polishing tools and the change of polishing parameters will affect the Preston coefficient, which makes it difficult to establish an accurate material removal model to achieve stable and excellent polishing quality. In this paper, a robotic polishing parameter optimization method considering time-varying wear is proposed to address these issues. First, combining the rich information in the theoretical modeling method with the data-driven regression method, a material removal regression model incorporating prior knowledge is proposed, which greatly reduces the large amount of experimental data required by the original regression model. The proposed model is able to track the wear variation of the sandpaper as well as the effect of polishing parameters. Then, based on the proposed prediction model, the genetic algorithm is used to optimize the polishing parameters in order to achieve better machining quality and less energy consumption. Finally, the experimental verification is carried out on the hybrid robot polishing test bench. The results show that the proposed material removal regression model incorporating prior knowledge has higher prediction accuracy and less required experimental data than existing models. The proposed robot polishing parameter optimization method can effectively compensate for tool wear and ensure the consistency of material removal during polishing while reducing energy consumption.