基于神经网络超弹性多楔带有限元分析

Finite Element Analysis of Hyperelastic V-ribbed Belt based on Neural Network

建立神经网络模型对多楔带材料特性进行预测,并将预测的材料特性应用于有限元分析模型中,对多楔带轮接触问题进行分析。研究结果表明:建立的神经网络模型,可以有效地预测肋复合橡胶材料的特性;在不同接触点带肋接触压力不同,且带肋接触压力从肋下到肋上逐渐增加;随着角度的增加,带肋的接触压力呈现先急剧增加后平稳增加后急剧减小的趋势;多楔带在高温和深深进入滑轮槽的条件下,其速度损失较大,传动效率较低。

Through establishing the neural network model to predict v-ribbed belt material property,and the material property is applied to the finite element analysis model,the V-ribbed belt pulley contact problems are analyzed.The results show that the neural network model can effectively predict the characteristic of rib composite rubber material.The ribbed contact pressure is different in different contact points,and the ribbed contact pressure gradually increase from under the floor to the rib.With the increase of the angle,the contact pressure of ribbed is increased dramatically after the first decreases sharply after smooth increase.The V-ribbed belt under the condition of high temperature and deep into the pulley groove,the velocity loss is bigger,the transmission efficiency is lower.