摘要
基于曲轴强化的残余应力理论,将人工神经网络引入发动机曲轴圆角的残余应力预测中,首先利用DEFORM有限元软件对480Q曲轴进行滚压试验,得到数组不同滚压参数对应的残余应力,然后根据此数据建立了比较稳定的神经网络,并利用此网络预测曲轴圆角滚压后的残余应力。该神经网络与有限元分析结果比较接近,为曲轴滚压中残余应力预测提供了一种新方法。
Based on the theory of crankshaft strengthening residual stress, an artificial neural networks（ANN） was introduced into residual stress prediction of the engine crankshaft round corner. Firstly, thefillet rolling experiments were made using finite element DEFORM software, a set of the residual stress which was corresponded to the different rolling parameters could be obtained, then a stable ANN was established using these data, and the residual stress of crankshaft round corner after rolling could be predicted using the ANN. The results of ANN prediction were similar to those of finite element analysis, a new method for predicting residual stress of the engine crankshaft round is proposed, orner residual stress.
出处
《铸造技术》
CAS
北大核心
2007年第5期686-689,共4页
Foundry Technology
基金
国家自然科学基金项目(50675060)
河北省科技攻关计划项目立项资助(04212156)
