Prediction Research On Material Removal Rate Of Blade Robot Belt Grinding

Titanium alloy blades are used as the core components of aeroengines.Their contour machining accuracy and surface quality determine the life and performance of aeroengines.With its low cost,high flexibility,flexible cutting,and consideration of grinding and polishing,robot belt grinding is gradually becoming a new method of blade precision grinding.However,the material removal of robotic abrasive belt grinding is affected by the coupling of processing trajectory,process parameters,contact wheel deformation and other factors.The mechanism is not thoroughly studied,which is the difficulty that hinders its further application.Therefore,based on the dimensional reduction contact method and XGBoost(Extreme Gradient Boosting)integrated learning method,the rule of material removal is revealed from the perspective of process parameters,and a belt grinding machine suitable for Ti-6Al-4V aviation hair blade robots is established.The material removal prediction model is used to achieve precise control of material removal during the blade grinding and polishing process,and is applied to actual production.The main research work of this paper is as follows:A material removal mechanism model based on the contact between single abrasive particle of hollow silicon carbide sand belt and dimension reduction is established.The material removal equation was established by analyzing the grinding trajectory of a single abrasive particle,and the three-dimensional pressure distribution on the contact surface between the contact wheel and the workpiece was mapped to the one-dimensional space by the dimensionless contact method.The experimental results of TC4 test block robot sandbelt grinding show that compared with the simplified Hertz model,the material removal rate model proposed in this paper reduces the root mean square error and average absolute percentage error from 2.02 to 1.67 and from 18.46% to 13.78%,respectively,showing higher prediction accuracy and robustness.A material removal rate prediction model based on integrated learning XGBoost algorithm is proposed.The model integrates a trained and optimized k-fold XGBoost algorithm to avoid overfitting and improve generalization by cross-validation and hyperparameter optimization.The experimental data showed that the predicted value of the model was basically consistent with the real value,and the root mean square error and the average absolute percentage error were reduced to 0.60 and 5.23%.Compared with other models,this model showed higher prediction accuracy on the same data set.For a certain type of aerospace blades,a robotic abrasive belt grinding and polishing platform was built.Based on the equal residual height method and the equal chord height error method,the trajectory planning and offline programming of the blade grinding and polishing processing area were designed.The blade grinding and polishing test was carried out,and good machining accuracy and surface quality were obtained,and the material removal rate prediction model was verified and applied.The above research is used to make up for the precise control of the material removal of the robot abrasive belt grinding,to achieve the quantitative removal and high automation of blade grinding and polishing;while ensuring the grinding accuracy and surface consistency,the blade rework is avoided,and it is successfully applied to the enterprise 's Mass production.