摘要
以动压润滑理论为基础,推导了线性液动压抛光的流场剪切力数学模型。借助FLUENT软件研究了流场底面的剪切力分布,通过单因素控制变量试验探究了各参数对剪切力的影响规律,灵敏度从高到低依次为:抛光间隙、抛光液黏度、抛光辊子转速、抛光辊子半径。以正交试验结果为训练集,建立了基于支持向量回归(SVR)的剪切特性预测模型,相关系数为98.35%、均方误差为3.44×10^(-3)。最后计算了剪切力理论值。对比发现,数值模拟和理论计算误差在15%以内;不同参数组合下剪切力分布趋势相同;SVR预测模型可信度高。
Based on the dynamic pressure lubrication theory,a mathematical model of the flow field shear forces of LHP was derived.With the help of FLUENT software,the shear force distribution on the bottom of flow fields was studied,and the influences of each parameters on shear force was explored through a single-factor control variable test.The sensitivity sequence from high to low is as follows:polishing gap,polishing fluid viscosity,polishing roller speed and polishing roller radius.Using orthogonal test results as training set,a shear characteristic prediction model was established based on support vector regression(SVR),and the correlation coefficient is as 98.35%and the mean squared error is as 3.44×10^(-3).Finally,the theoretical values of shear forces were calculated.Comparison results show that the errors between numerical simulations and theoretical calculations are within 15%.Shear force distribution trends are the same under different parameter combinations.SVR prediction model has a high credibility.
作者
文东辉
许鑫祺
郑子军
WEN Donghui;XU Xinqi;ZHENG Zijun(School of Mechanical Engineering,Zhejiang University of Technology,Hangzhou,310023)
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2021年第18期2203-2210,2216,共9页
China Mechanical Engineering
基金
国家自然科学基金(51775509)
浙江省自然科学基金(LZ17E050003)。
关键词
线性液动压抛光
剪切特性
灵敏度
预测模型
linear hydrodynamic polishing(LHP)
shear characteristic
sensitivity
prediction model