基于人工神经网络的螺旋锥齿轮磨削加工表面粗糙度预测

Surface Roughness Forecasting of Spiral Bevel Gear Based on Artificial Neural Network

影响螺旋锥齿轮磨削加工表面粗糙度Ra的因素众多且很不明确,Ra值的预测属于典型的模糊非线性问题.根据神经网络原理,建立了预测Ra的BP模型,此模型可精确地描述砂轮进给速度、齿深进给量对螺旋锥齿轮磨削加工表面粗糙度的影响.实验证明,用BP模型预测螺旋锥齿轮磨削加工表面粗糙度可获得平均相对误差为3.78%的高精度预测结果.

Because the value of Ra is affected by a lot of factors and some of them are undefined,the surface roughness forecasting of spiral bevel gears is a typical fuzzy,non-linear system.In this paper,based on the priority principle of BP artificial neural network,surface roughness forecasting is set up.This model BP can accurately describe the effect of wheel's feed velocity and deep tooth feed on surface roughness of spiral bevel gears.The experiment data proves that the model BP used in forecasting the surface ...