Research On Grinding Related Parameters In Curved Surface Polishing By Ball-end Abrasive Tool

In modern manufacturing process,polishing process usually occupies 37% to 50% of the total manufacturing time during the machining process for curved surface parts especially complex curved parts.Besides,polishing mainly relies on the manual operation of the worker,resulting in low production efficiency and unstable processing quality,which is difficult to meet the short-term and high-quality modern manufacturing requirements.The ball-end polishing technique has the characteristics of high precision and processing efficiency with low cost and good surface flexibility.The removal function of the ball abrasive tool has small beam diameter that is beneficial to modifying areas.In this paper,surface polishing process of the ball type abrasive tool was mainly researched.The effects of relevant process parameters on the polishing performance index were studied.The optimization of the process parameters was performed.The main research work and achievements is summarized as the following aspects.Firstly,material removal profile model was established based on Hertz contact theory and Preston equation combined with the analysis of the material removal mechanism.The modified Preston coefficient was determined by spot polishing experiments under the convex contact condition and the relevant material removal profile equation was obtained.The results showed that the model has high reliability.Secondly,polishing process and parameter optimization of ball type abrasive tools were researched.Single factor tests on curved surface polishing of SKD-11 steel work pieces were carried out by using elastic abrasive tool with different grain size,separately aiming at the process parameters of the abrasive tool such as the particle size,the abrasive tool rotation speed and the setting depth.The influences of setting depth and other process parameters on material removal rate(MRR),abrasive wear and surface roughness(Ra)were researched through the polishing tests.Then MRR,wear ratio and Ra were used as the evaluation indexes of polishing performance.The orthogonal experiment was designed to study the impact of each parameters on the evaluation index.The Signal to Noise Ratio(SNR)analysis and Weighted Grey Relation Degree analysis based on Analytic Hierarchy Process were used to gain the multi-index optimized parameter combination which aims at evaluation indexes.Afterwards,the power function regression model of MRR and surface roughness Ra was established.The optimized parameter range of the process parameters was determined through the interval sensitivity analysis.At last,the results of the experiments revealed that the polishing quality and MRR were improved under the condition in optimized parameter range.Thirdly,multi-objective neural network prediction model was established that MRR and Ra were designed as the output vectors,with four main process parameters set as the input vectors.Particle Swarm Optimization(PSO)algorithm was used to optimize network weights and thresholds.The network convergence effect was improved by adaptive learning factor algorithm.The average error of the predicted values from the neural network was less than 10% after training.The results showed that the model could be used for process prediction analysis in curved surface polishing by ball type abrasive tool.