基于BP神经网络的齿轮齿向载荷分布系数计算

The Calculation of Longitudinal Load Distribution of Gear Based on BP Neural Network

介绍了关于载荷参数计算的方法,提出了齿轮优化设计的基本思路。由于齿轮安装以及支承变形等因素,齿向载荷分布系数Kβ取值范围大且计算复杂。应用BP神经网络对计算接触强度用的参数KHβ进行计算,预测结果表明使用该方法是可行和高效的,该方法简便,计算量较小,容易与齿轮优化设计集成。

Some researches and methods on load coefficient were reviewed, and a basic thought on the optimization design of gear was proposed. The range of longitudinal load distribution factor Kβ is great because of mounting errors, support deformation etc. A Back-Propagation Neural Network was used to calculate the parameter Kβ for eontact strength analysis. The predicted value can satisfy calculating requirement with high precision, low calculation amount and great efficiency. The model can be easily integrated in the optimization design of gear mechanism system.