基于神经网络优化的正交试验内齿轮齿面偏载矫正研究  

Research on Correcting Eccentric Load of the Internal Gear Tooth Surface by Orthogonal Tests Based on Neural Network Optimization

作　　者：王振博 郑鹏[1] 刘逸飞 Wang Zhenbo;Zheng Peng;Liu Yifei(School of Mechanical Engineering,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学机械工程学院,辽宁沈阳110870

出　　处：《机械传动》2023年第8期16-23,共8页Journal of Mechanical Transmission

基　　金：中国科学院战略性先导科技专项(XDC04030503)。

摘　　要：为解决盾构机大型内啮合齿轮工作时由轴偏角造成的齿面偏载问题,给出了综合修形齿面的齿面方程,提出了采用神经网络联合正交试验的方法对齿面修形参数进行筛选预测。该种方法不再将齿廓修形和螺旋线修形看作相互独立的调整方法,能够将齿面偏载转变为均匀分布在齿面上的载荷。通过有限元方法对修形后的齿轮进行接触分析发现,该方法能够有效降低因轴偏角造成的齿面局部应力,并减小了传动误差;可以一次性找到较优的齿面修形参数值,避免了对大型盾构机内啮合齿轮进行啮合试验成本过高等问题。In order to solve the problem of tooth surface eccentric load caused by shaft deflection angle of large internal meshing gears of the shield machine,the tooth surface equation of comprehensively modified tooth surface is given,and the method of neural network combined with orthogonal test is proposed to screen and pre⁃dict the tooth surface modification parameters.In this method,the tooth profile modification and tooth direction modification are no longer regarded as independent variables,and the eccentric load on the tooth surface can be transformed into the load evenly distributed on the tooth surface.Through the contact analysis of the modified gear by the finite element method,it is found that this method can effectively reduce the excessive local stress on the tooth surface caused by the shaft deflection angle and reduce the transmission error.The optimal tooth sur⁃face modification parameter value can be found at one time,which avoids the high cost of meshing experiment for the internal meshing gear of large shield machines.

关 键 词：齿轮修形 正交试验 神经网络 大型内齿轮 

分 类 号：TH132.41[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]