Since it is difficult to predict the mechanical property parameters of tube, the parameter prediction method is proposed, which is based on RBF neural network and tube tensile tests. The stress-strain curves of partial tube are investigated by tensile tests. Then, the sample space of a neural network is established. On this basis, the neural network input parameters and output parameters are determined, and the tube is classified according to the sizes and materials to build a layered neural network model. The comparison of Network prediction and experimental results shows that the RBF neural network can effectively predict the mechanical performance parameters of tube.

In order to improve the mechanics properties of the rotor components at ultra-high load steady acceleration conditions, the significant factors to the rotor mechanics properties, which include the ellipsoid axis length, minor axis length, rotor speed and gyration radius of holes, are taken as the optimizing objects, the extreme stress, strain extremes, moment of inertia of rotor are optimized as the targets. Then optimization models of the rotor mechanics parameters are established, and the multi-objective optimization method of rotor mechanics properties is proposed based on particle swarm. The results show that the mechanics properties of the rotor were significantly improved using the optimize method.