基于神经网络的CFRP冲击失效预测研究
Study on CFRP Impact Failure Study based on Neural Network
张显俊 1龚巍俊 1李秋节 1范秋寒 1秦国锋1
作者信息
- 1. 广西师范大学 职业技术师范学院 广西 桂林 541004
- 折叠
摘要
针对CFRP材料的冲击失效预测问题,采用CFRP板开展冲击实验,基于ABAQUS建立仿真分析模型,比较仿真分析与实验测试的冲击载荷,相对误差低于 3.3%,说明了仿真分析方法的有效性.改变冲击能量、冲击角度和CFRP厚度,设计正交实验,基于仿真分析模型获得冲击载荷.将仿真冲击载荷、冲击角度、CFRP厚度抽象为神经网络输入神经元参数,基于BP神经网络理论,建立冲击载荷预测模型,调整训练参数并重新排列进行模型训练.结果表明,文章模型预测准确度较高.
Abstract
For the impact failure prediction of CFRP materials,CFRP board is used to carry out the impact experiment,establish a simulation analysis model based on ABAQUS,compare the impact load of simulation analysis and experimental test,and the relative error was less than 3.3%,which demonstrates the effectiveness of the simulation analysis method.The study changes the impact energy,impact angle and CFRP thickness,designs the orthogonal experiments,and obtains the impact load based on the simulation analysis model.The simulated impact load,impact Angle and CFRP thickness are abstracted into the input neuron parameters of the neural network.Based on the BP neural network theory,the impact load prediction model is established,and the training parameters are adjusted and rearranged for model training.The results show that the present model has a high prediction accuracy.
关键词
CFRP材料/神经网络Key words
CFRP Material/Neural Network引用本文复制引用
出版年
2024
NETL
NSTL
万方数据