摘要
为实现航空发动机高压转子零件快速且精确装配,通过智能算法对转子零件装配偏心进行预测进而进行相位优化。首先,运用前30阶傅里叶级数的方法模拟形貌误差并生成误差数据。其次,添加误差数据到有限元模型中计算出装配偏心。再次,搭建BP人工神经网络模型,提取傅里叶级数的幅值和相位作为神经网络的输入,装配偏心作为网络输出,在神经网络中加入衰减学习率、正则化、滑动平均算法,使计算装配偏心更加精确、稳定,使用200组数据完成神经网络训练,并用训练好的网络对3组测试数据进行验证。最后,分别使用神经网络计算得到的不同装配零件每个相位的偏心作为粒子群算法需要优化的目标,通过误差传递计算,得到优化后的零件装配相位。研究结果表明:使用神经网络模型计算装配偏心可以充分考虑止口形貌特征和装配变形,并明显提高计算效率,再运用粒子群算法对不同相位进行最优选择,达到满足航空发动机高压转子装配同轴度要求,提高服役性能。
To achieve rapid and accurate assembly of aero-engine high-pressure rotor parts,we attempt to predict the assembly eccentricity of the rotor parts via intelligent algorithms and then optimize the phase.The first 30 orders of Fourier series are adopted to simulate the shape error and generate error data.Adding the error data to the finite element model to calculate the assembly eccentricity,BP artificial neural network model is established.The amplitude and phase of the Fourier series are extracted as the input of the neural network and the assembly eccentricity as the network output.The attenuation learning rate,regularization,and moving average algorithm participate in the neural network to calculate the assembly eccentricity more accurately and stably.200 sets of data are used to complete the neural network training and the trained network verifies three sets of test data.The eccentricity of each phase of different assembly parts is calculated with this neural network.Taking the phase as the objective of particle swarm optimization,the optimized assembly phase of the parts is obtained by error transfer calculation.This approach shows that this neural network model fully considers the morphology of the flange and assembly deformation,and significantly improves the calculation efficiency.Then particle swarm optimization algorithm is used to optimally select different phases to meet the requirements of aero-engine rotor assembly and promote service performance.
作者
张子豪
郭俊康
洪军
孙岩辉
ZHANG Zihao;GUO Junkang;HONG Jun;SUN Yanhui(School of Mechanical Engineering,Xi’an Jiaotong University,Xi’an 710049,China;Key Laboratory of Education Ministry for Modern Design&Rotor-Bearing System,Xi’an Jiaotong University,Xi’an 710049,China;School of Construction Machinery,Chang’an University,Xi’an 710064,China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2021年第2期47-54,共8页
Journal of Xi'an Jiaotong University
基金
航空发动机及燃气轮机重大专项基础研究项目(2017-VII-0010-0105)
国家自然科学基金资助项目(51805419)
中国博士后科学基金资助项目(2019T120899)。
关键词
航空发动机高压转子
人工神经网络
偏心预测
粒子群算法
相位优化
aero-engine high-pressure rotor
artificial neural network
eccentric prediction
particle swarm optimization algorithm
phase optimization
作者简介
张子豪(1996-),男,硕士生;通信作者:郭俊康,男,助理研究员。
