利用软计算技术改善圆锥滚子轴承的工作性能

Using soft computing technology to improve performance of tapered roller bearing

双列圆锥滚子轴承主要用于支承具有预紧力的组合式负载,以确保合理的运行状态。因为轴承滚道会出现高接触应力、疲劳剥落或表面点蚀,所以,开发一种工具,以期提前了解预设载荷是否会引发故障,就显得非常重要。为此提出一种基于四类荷载表达值的工况分类方法,首先建立了一种三维有限元(FE:Finite Element)模型;其次,设计了一个实验程序,在降低基于有限元模型的模拟计算成本的同时,提供最大的信息量;随后,将四种载荷条件分别进行实验,用统计分析和机器学习技术创建分类模型,再采用特征变换和约简、算法参数调整以及验证等方法,完成了鲁棒分类模型;最后,在柔性判别与分析的基础上,得到了最佳结果。该方法和最终模型都得到了验证,其准确性也得到公认。

Double row-TRB is a mechanical device used to support a combined load with the best preloading force to ensure reasonable operation state.Because of the high contact stress,fatigue spalling,or surface pitting of the raceway,it is important to develop a tool to understand in advance whether a preset load will cause failure.For this purpose,a method of working condition classification based on the expressed values of four kinds of loads is proposed.Firstly,a three-dimensional finite element(FE)model is established;Secondly,an experimental program is designed to reduce the cost of simulation based on finite element model,and at the same time to provide the maximum amount of information;Thirdly,each experiment was carried out under four loading conditions,and the classification model was created by statistical analysis and machine learning techniques;Then,the robust classification model was completed by using the methods of feature transformation and reduction,algorithm parameter adjustment and verification;Finally,the optimal results are obtained on the basis of the discrimination and analysis of flexibility.Both the method and the final model have been verified,and its accuracy has been recognized.