线性液动压抛光波纹度加工特性研究

目的 探究线性液动压抛光波纹度特性，建立抛光波纹度预测模型，获取最佳加工工艺参数组合。方法结合线性液动压抛光原理，分析抛光波纹度产生机理，探究流场力分布特性，并获得影响波纹度的相关工艺参数。设计单因素试验，探究各工艺参数对波纹度的影响规律，并进行显著性分析，选取显著参数为试验因子，设计正交试验，以试验结果作为训练集，建立基于支持向量回归机（Support Vector Regression,SVR）的波纹度预测模型。以该预测模型为适应度函数，进行遗传算法寻优，以获取最佳工艺参数。结果 线性液动压抛光波纹度由流场力分布特性及工件进给运动共同作用产生，其大小受抛光间隙、抛光速度、进给速度和抛光液黏度影响。单因素试验分析结果显示，抛光波纹度Wa随进给速度和抛光间隙的增大而增大，随抛光速度的增大而减小，随抛光液黏度的增大而先减小、再增大。其中进给速度的影响最显著，抛光间隙和抛光速度次之，而抛光液黏度的影响具有分段差异性，显著性最弱。以正交试验数据所建立的抛光波纹度预测模型的回归相关系数R2为0.992 0。随机验证实验结果显示，各组预测值与真实值的误差均在10%以内，遗传算法寻优得到最佳工艺参...

Processing Characteristics of Linear Hydrodynamic Polishing Waviness

This paper aims to explore the characteristics of linear hydrodynamic polishing waviness, establish a polishing waviness prediction model, and obtain the best combination of processing parameters. The waviness generation mechanism was analyzed according to the principle of linear hydraulic pressure polishing, the relevant process parameters that affect the waviness was obtained according to the force distribution characteristics of the flow field. A single factor experiment was designed to explore the influence of each process parameter on the waviness and a significant analysis was conducted, the significant parameter was selected as the experimental factor to design an orthogonal experiment. A Waviness prediction model based on Support Vector Regression (SVR) was established with a training set composed of experimental result, the prediction model was used as the fitness function to perform genetic algorithm optimization to obtain the best process parameters. Linear hydrodynamic polishing waviness is produced by the combined action of the force distribution characteristics of the flow field and the feed motion of the workpiece. Its size is affected by the polishing gap, polishing speed, feed speed and polishing fluid viscosity. The single factor experimental analysis results show:The Wa value of polishing waviness increases with the increase of feed speed and polishing gap, decreases with the increase of polishing speed, and first decreases and then increases with the increase of polishing liquid viscosity. The influence of feed speed is the most significant, followed by polishing gap and polishing speed, while the influence of polishing liquid viscosity has segmental differences, and the significance is the weakest. The regression correlation coefficient R2 of the polishing waviness prediction model established by orthogonal experimental data is 0.992 0. The results of random verification experiments show that the errors between the predicted values and the true values of each group are within 10%. The optimal process parameters obtained by the genetic algorithm optimization are:(h0, u0, vf)=(50, 8, 200), the surface waviness of the workpiece after polishing for one hour is 5.23 nm. Reasonable selection of process parameters can optimize polishing waviness, and the SVR-based linear hydrodynamic polishing waviness prediction model has reliable predictive ability and can realize controlled processing of polishing waviness.