航空发动机叶片精锻残余应力数学模型对比分析  

Comparative analysis of mathematical models of residual stress in precision forging of aero engine blades

作　　者：吴捍疆 张丰收[1] 燕根鹏 WU Han-jiang;ZHANG Feng-shou;YAN Gen-peng(School of Mechanical Engineering,Henan University of Science and Technology,Luoyang 471003,China;CITIC Heavy Industries Co.,Ltd.,Luoyang 471039,China)

机构地区：[1]河南科技大学机电工程学院,河南洛阳471003 [2]中信重工机械股份有限公司,河南洛阳471039

出　　处：《塑性工程学报》2020年第6期135-140,共6页Journal of Plasticity Engineering

基　　金：国家自然科学基金资助项目(51475146,51475366)。

摘　　要：为了预测航空发动机叶片精锻过程中所产生的最大残余应力值,在已建立好的叶片精锻有限元模型基础上,基于simufact.forming平台对TC4钛合金航空发动机叶片精锻过程进行了仿真研究,分别获得了上模速度、坯料温度、模具温度和摩擦系数等不同工艺参数组合下的叶片最大精锻残余应力值。以收集的仿真数据作为训练样本,先后建立了响应曲面模型、BP神经网络模型和GA-BP神经网络模型,并将这3种模型的预测值与仿真值进行比较。结果表明:所建立预测模型的最大相对误差不超过6%,其中GA-BP神经网络模型预测精度最高,可对航空发动机叶片精锻最大残余应力进行有效的预测。To predict the maximum residual stress produced in the precision forging process of the aero engine blade,based on the established precision forging finite element model of the blade,the process of precision forging of TC4 titanium alloy aeronautical blade was simulated based on the simufact.forming platform.The maximum residual stress values of the blade in precision forging under different process parameter combinations of the upper die speed,blank temperature,die temperature and friction coefficient were obtained.Taking the collected simulation data as the training samples,the response surface model,BP neural network model and GA-BP neural network model were established,and the predicted values of the three models were compared with the simulation values.The results show that the maximum relative error of the prediction model is less than 6%.The prediction accuracy of GA-BP neural network model is the highest and it can effectively predict the maximum residual stress in the precision forging of aero engine blade.

关 键 词：航空发动机叶片 精锻成形 残余应力 响应曲面 BP神经网络 遗传算法 

分 类 号：V263.1[航空宇航科学与技术—航空宇航制造工程]