基于神经网络的齿轮根切特性仿真研究

Research of Undercut Gear Characteristic Simulating Based on Neural Network

根据齿轮加工的范成原理,采用共轭齿面曲率的关系,建立齿条型刀具齿顶圆角与齿轮齿根圆角的对应关系;运用神经网络的非线性映射特性建立齿轮模数和齿数与根切量的关系模型,仿真出齿轮齿数与根切量,齿轮模数与根切量的关系曲线。仿真曲线能直观的反映齿数和模数对根切量的影响情况,为齿轮设计和加工提供了理论数据。

Based on the principle of gear wheel generation, using the relationship of conjugate tooth surface curvature, the corresponding relation between tool tip fillet and the gear root fillet was established. The nonlinear model based on the relation be- tween the gear modulus and teeth number and undercut amount by BP neural network was established, the curve of the gear modu- lus and teeth number and undercut amount in the method was simulated. The results indicated that the method of BP neural net- work simulation can intuitively reflect the undercut amount, it provided the theory data for the design and manufacture of gear.