摘要
将BP和RBF神经网络的理论和算法应用于预测超高压容器爆破压力的研究中。选用MATLAB神经网络工具箱建立预测爆破压力的神经网络模型,研究模型中影响爆破压力的主要参数,内外径比值和材料的强度极限,屈服极限,屈服强度与强度极限的比值;选用Faupel、Crossland和Bones等文献中的爆破实验数据对神经网络模型进行训练,用训练好的神经网络模型对爆破压力进行预测。预测结果表明,用BP和RBF神经网络方法建立的模型能够对超高压筒形容器的爆破压力进行较为准确的预测。
The theory and the algorithm of BP and RBF neural network are applied in the research for predicting the bursting pressure of ultra-high pressure vessel.First,the neural network model has been established for predicting bursting pressure by using MATLAB Neural Network Tools in consideration of the main factors of influencing bursting pressure.The factors include ultimate strength,yield stress,ratio of outer radius to inner radius of the pressure vessel cylinders and yield ratio.Then the established neural network model is trained by choosing a large amount of bursting experimental data from Faupel,Crossland and Bones,and some references.Finally,the trained neural network model is used to predict the bursting pressure.The prediction results show that the bursting pressure model using the BP and RBF neural network method can predict bursting pressure exactly.
出处
《兵器材料科学与工程》
CAS
CSCD
2010年第2期31-34,共4页
Ordnance Material Science and Engineering
基金
宝鸡文理学院重点科研项目(ZK0727)资助
关键词
超高压容器
爆破压力
BP神经网络
RBF神经网络
预测
ultra-high pressure vessel bursting pressure BP neural network RBF neural network prediction