For a high speed moving mechanism driven by fluid, abrupt stop resulted from barriers will lead to its large deformation, even fracture. To achieve optimum structural style and parametric matching of buffer members, nine test samples were selected out of twenty-seven design styles through orthogonal experiment method in advance. Their high-speed moving and impacting process were simulated by nonlinear finite element method(FEM) considering fluid structure interaction. On the basis of the above results, a modified error back-propagation(BP) network method was applied to train these samples, and obtained the nonlinear mapping relation between parameters of tubes for energy absorption and strain energy of crucial parts. The optimum structural parameters of buffer tube were determined, at the same time, the efficiency of schemes selection was improved obviously.
Zhang G P, Huang Y M, Shi W H, et al. Predicting dynamic behaviours of a whole machine tool structure based on computer-aided engineering[J].International Journal of Machine Tools & Manufacture.2003, 43(6):699-706
����̩,���ˣ�����������������[M]������:��ѧ������,2001 Fang Kaitai, Ma Changxing. Orthogonal and uniform experiments design [M]. Beijing: Science Press, 2001(in Chinese)
��Ϊ,����Ⱥ��������Ԫ����ģ�͵�ϵͳ�����Ż���Ʒ���[J]����Ϣ�����,2001,30(6):562-565 He Wei, Han Liqun. Neural network modeling based method for optimizing product parameters [J]. Information and Control, 2001, 30(6): 562-565(in Chinese)
Karatasou S, Santamouris M, Geros V. Modeling and predicting building’s energy use with artificial neural networks: methods and results [J].Energy and Buildings.2006, 38:949-958
�����磮Matlab��������Ӧ��[M]������:������ҵ������,2005 Dong Changhong. Matlab neural network and application [M]. Beijing: Defense Industry Press, 2005 (in Chinese)