摘要
受重力等外部载荷的影响,飞机装配定位工装的结构不可避免地产生变形,直接影响飞机结构的装配精度。然而,由于产品外形遮挡,传统基于视觉的结构变形测量方法无法直接获取工装结构变形。对此,本文提出了一种激光测量数据反演的定位结构终端变形精准快速预测方法。通过激光扫描仪对定位结构进行扫描测量,获取结构实测数模;基于定位结构数字化仿真模型,自动批量化获取仿真数据集,并采用多层感知机网络建立了待测终端变形与可视区域变形的映射模型;在此基础上,构建了工装结构终端变形反演差分优化方法,实现了定位工装结构终端变形的反演求解。通过实例分析发现,定位工装终端反演预测值与实测值最大误差为8.25%,验证了所提出的定位结构变形反演方法的有效性。
Due to the influence of external loads such as gravity,the structure of aircraft assembly positioning tool is inevitably deformed,which directly affects the assembly accuracy of aircraft structure.However,due to the occlusion of product shape,the traditional vision-based structural deformation measurement method can't directly obtain the structural deformation of tooling.In this paper,an accurate and fast prediction method of terminal deformation of positioning structure based on laser data inversion is proposed.The positioning structure is scanned and measured by laser scanner,and the measured numerical model is obtained.Based on the digital simulation model of the positioning structure,the simulation data set is obtained automatically,and the mapping model between the deformation of the terminal and the deformation of the visible region is established by using the multi-layer perceptron network.On this basis,the inverse differential optimization method of the terminal deformation of the tool structure is constructed,and the inverse solution of the terminal deformation of the positioning tool structure is realized.The results show that the maximum error between the predicted value and the measured value is 8.25%,which verifies the validity of the proposed method.
作者
李宣慧
石文雄
郭文娟
胡文龙
骆彬
梁彪
LI Xuanhui;SHI Wenxiong;GUO Wenjuan;HU Wenlong;LUO Bin;LIANG Biao(Northwestern Polytechnical University,Xi'an 710072,China;AVIC Xi'an Aircraft Industry Group Company Ltd.,Xi'an 710089,China;Key Laboratory of Aircraft High Performance Assembly,Ministry of Industry and Information Technology Xi'an 710072,China)
出处
《航空制造技术》
北大核心
2025年第17期82-88,98,共8页
Aeronautical Manufacturing Technology
关键词
定位结构
激光测量
神经网络
差分优化
变形预测
Positioning structures
Laser measurement
Neural networks
Differential optimization
Deformation prediction
