Multi-objective prediction and optimization for soft abrasive rotary jet polishing inner surface

In aerospace, optical instrumentation, new energy, life science, and other fields, a large number of core components need ultra-precision polishing to obtain a high-quality inner surface, but polishing quality and efficiency are difficult due to mutual constraints. In order to polish the inner surface under both high quality and efficiency, a new soft abrasive rotary jet polishing (SARJP) method and system are proposed. Based on the polishing method and system proposed in this paper, an integrated multi-objective optimization algorithm model based on CatBoost and AGE-MOEA is proposed to realize the multi-objective prediction and process optimization of the roughness and material removal rate of the inner surface of a circular tube by SARJP. The lowest average absolute percentage error of surface roughness is 3.73%, and the largest average absolute percentage error is 6.49%, according to a comparison of the multi-objective optimization forecast findings with the experimental results. The minimum average absolute percentage error in the material removal rate is 3.72%, and the maximum average absolute percentage error is 8.28%. The initial surface Ra value of the inner surface of the circular tube is reduced by 94.99% from 211.16 to 10.57 nm, and the maximum material removal rate is increased by 15.83% from 5.234 to 6.218 mg/min, showing the effectiveness of the algorithm for prediction and optimization.