In this paper,artificial neural network(ANN)-finite element metho (FEM) model of rubber piston membrane to work out a pneumatic braking device based on Traincgf algorithm were established.In this model,the thickness of the rubber piston membrane of the pneumatic braking device was T,radius of the central ring was R,the fillet radius was R1 and fillet radius was R2,the output was stress analysis for the rubber piston membrane of the pneumatic braking device and network framework was considered to be 4-6-1.The final relative errors (RE) was 9.225%. As for optimal parameters for the rubber piston membrane of the pneumatic braking device the thickness was 8 mm,R1 was 6 mm,R2 was 18 mm and the greatest tolerant pressure was 4.83 MPa.%基于Traincgf算法建立了建立气动刹车装置橡胶活塞膜的神经网络(ANN)-有限元(FEM)设计模型。气动刹车装置橡胶活塞膜的厚度T,中心环半径R,倒圆角半径R1,倒圆角半径R2;输出层为气动刹车装置橡胶活塞膜应力分析。网络模型结构为4-6-1。最终测试相对误差(RE)为9.225%。气动刹车橡胶活塞膜设计的最佳参数为厚度T为8mm,倒角R1为6mm,倒角R2为18mm,可承受的压力最大为4.83MPa。
展开▼