改进的BP神经网络在多裂纹柱体扭转中的应用  

Application of progressed Back-Propagation neural network to multi-cracked prismatic shaft torsion

作　　者：赵军华[1] 郭万林[1] 孟波[1] 

机构地区：[1]南京航空航天大学纳米科学研究所,南京中国210016

出　　处：《计算力学学报》2007年第3期289-293,共5页Chinese Journal of Computational Mechanics

基　　金：国家自然科学基金(50275073);航空科学基金(03B52011)资助项目

摘　　要：提出一种基于BP神经网络的多裂纹柱体扭转问题的数据新处理方法。以多裂纹柱体扭转问题为例,以MATLAB中的神经网络工具箱为工具,采用了改进的BP神经网络,并对其设计方案进行了详细的分析说明,发现动量参数对训练次数影响很大,而学习率对它的影响很小;采用双隐含层比单隐含层训练更稳定,收敛的也更快,同时给出了理想的学习方案。最后对柱体的抗扭刚度实验值进行快速拟合,得到了裂纹尖端的应力强度因子K3。结果证明这种设计方案计算的更精确、收敛速度更快。Based on the Back-Propagation neural network theory, a new method is presented to process the data of the multi-cracked prismatic shaft torsion problem. For a multi-cracked prismatic shaft torsion problem example, an optimized project of Back-propagation training is introduced by using Neural Network toolbox in MATLAB software, and the project is explained in detail and the good learning scheme is given by simulating the experimental results of the torsion rigidity. In the method, the momentum parameter a has intensive influence on the training times, while the learning ratio η has little effect on them. In addition, the training is more effective with the couple hidden layers than that with the single hidden layer. Finally, the stress intensity factor Ks at the crack tip can be obtained by the project. It is proved that the method is accurate and converged quickly by the example of the experiment.

关 键 词：应力强度因子 BP神经网络 MATLAB 抗扭刚度 

分 类 号：O345[理学—固体力学]