Please wait a minute...
浙江大学学报(工学版)  2019, Vol. 53 Issue (7): 1274-1281    DOI: 10.3785/j.issn.1008-973X.2019.07.005
机械与能源工程     
基于梯度提升树的飞机机身对接状态识别
蔡畅1(),黄亦翔1,*(),邢宏文2
1. 上海交通大学 机械系统与振动国家重点实验室,上海 200240
2. 上海飞机制造有限公司 航空制造技术研究所,上海 200436
State recognition for fuselage join based on gradient boosting tree
Chang CAI1(),Yi-xiang HUANG1,*(),Hong-wen XING2
1. State Key Laboratory of Mechanical System and Vibration, Shanghai Jiao Tong University, Shanghai 200240, China
2. Institute of Aeronautical Manufacturing Technology, Shanghai Aircraft Manufacturing Limited Company, Shanghai 200436, China
 全文: PDF(1317 KB)   HTML
摘要:

为了实时监控飞机机身的对接过程,针对机身对接数据没有标注和样本不平衡的特点,提出基于梯度提升树(GBDT)的机身对接状态识别方法. 通过定位器及定位器上的载荷传感器,实时获取机身对接过程中的位移和载荷数据. 结合飞机部件对接的工艺流程对历史对接数据进行状态标注,提出准确、高效的对接状态自动标注方法. 在经过标注的对接数据上训练基于GBDT的机身对接状态识别模型,通过该模型可以获得各个特征的重要性. 与长短期记忆网络(LSTM)、卷积神经网络(CNN)以及一些传统机器学习方法相比,该方法对接状态识别的宏F1(macro_F1)指标高达0.998,能够精准地识别每一种对接状态且训练速度较快.

关键词: 机身对接状态标注状态识别数据驱动梯度提升树(GBDT)不平衡多分类    
Abstract:

A state recognition method for fuselage join based on gradient boosting decision tree (GBDT) was proposed by considering the practical conditions of the lack of label and sample imbalance of fuselage joining data in order to monitor the process of fuselage join in real time. The displacement and the load data were acquired in real time through the positioners and the load sensors during the process of fuselage join. The joining state of historical data was labeled based on the process of airliner component join, and an accurate and efficient automatic labeling method for fuselage joining state was proposed. The state recognition model for fuselage join based on GBDT was trained through the labeled data, from which the importance of each feature was obtained. The macro_F1 for joining state recognition of the proposed method was as high as 0.998, compared with the latest deep learning methods such as long short-term memory (LSTM), convolutional neural network (CNN) and some traditional machine learning methods. Each joining state was accurately recognized, and the model training process was more efficient.

Key words: fuselage join    state labeling    state recognition    data driven    gradient boosting decision tree (GBDT)    unbalanced multi-classification
收稿日期: 2018-07-23 出版日期: 2019-06-25
CLC:  TP 274  
通讯作者: 黄亦翔     E-mail: ccuniquelife@sjtu.edu.cn;huang.yixiang@sjtu.edu.cn
作者简介: 蔡畅(1993—),男,硕士生,从事工业数据挖掘的研究. orcid.org/0000-0002-1772-051X. E-mail: ccuniquelife@sjtu.edu.cn
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章  
蔡畅
黄亦翔
邢宏文

引用本文:

蔡畅,黄亦翔,邢宏文. 基于梯度提升树的飞机机身对接状态识别[J]. 浙江大学学报(工学版), 2019, 53(7): 1274-1281.

Chang CAI,Yi-xiang HUANG,Hong-wen XING. State recognition for fuselage join based on gradient boosting tree. Journal of ZheJiang University (Engineering Science), 2019, 53(7): 1274-1281.

链接本文:

http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2019.07.005        http://www.zjujournals.com/eng/CN/Y2019/V53/I7/1274

图 1  集成学习原理示意图
图 2  飞机部件的对接原理
图 3  部件自动对接流程
图 4  前机身和中机身对接示意图
图 5  各特征的箱线图
图 6  调姿过程中S2LAPOS_X的变化
图 7  机身连接过程中Y方向力的变化
对接状态 编号 人工标注 自动标注 协同标注
空工位 0 32 107 32 107 32 107
对接准备 1 19 081 19 084 19 081
调姿对接 2 987 956 987
连接1 3 ? 24 881 24 853
连接2 4 ? 80 928 80 928
表 1  不同标注方法的各对接状态样本数量
图 8  当前样本与前一样本的欧氏距离
图 9  各对接状态的样本数目
真实情况 预测结果
正例 反例
正例 TP(真正例) FN(假反例)
反例 FP(假正例) TN(真反例)
表 2  分类结果的混淆矩阵
图 10  GBDT模型的混淆矩阵
图 11  GBDT模型的特征重要性
模型 macro_P macro_R macro_F1 F1(0) F1(2) t/s
GBDT 0.998 0.998 0.998 1.000 1.000 0.2
LR 0.984 0.837 0.889 0.857 0.653 0.4
SVM 0.984 0.919 0.944 1.000 0.777 0.9
RF 0.990 0.984 0.987 1.000 0.970 0.1
LSTM 0.981 0.975 0.978 1.000 0.926 11.4
CNN 0.984 0.972 0.977 1.000 0.925 18.6
表 3  各模型评价指标的对比
图 12  各模型的F1
1 邹冀华, 刘志存, 范玉青 大型飞机部件数字化对接装配技术研究[J]. 计算机集成制造系统, 2007, 13 (7): 1367- 1373
ZOU Ji-hua, LIU Zhi-cun, FAN Yu-qing Large-size airplane parts digital assembly technology[J]. Computer Integrated Manufacturing Systems, 2007, 13 (7): 1367- 1373
doi: 10.3969/j.issn.1006-5911.2007.07.019
2 戴肇鹏. 飞机部件对接调姿技术研究与软件开发[D]. 南京: 南京航空航天大学, 2015.
DAI Zhao-peng. Research on the adjustment technology of position and pose for aircraft component assembly and software development [D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2015.
3 陈伟, 李明慧, 周曈, 等 基于数字化装配偏差建模的飞机舱段对接定位方案研究[J]. 机械设计与制造, 2012, (1): 245- 247
CHEN Wei, LI Ming-hui, ZHOU Tong, et al Research of aircraft fuselage joining and positioning scheme based on digital assembly variation modeling[J]. Machinery Design and Manufacture, 2012, (1): 245- 247
doi: 10.3969/j.issn.1001-3997.2012.01.092
4 王青, 梁琴, 李江雄, 等 飞机数字化装配机翼姿态评价及调整方法[J]. 浙江大学学报: 工学版, 2014, 48 (7): 1287- 1294
WANG Qing, LIANG Qin, LI Jiang-xiong, et al Estimation and alignment method of wing position and orientation for aircraft digital assembly[J]. Journal of Zhejiang University: Engineering Science, 2014, 48 (7): 1287- 1294
5 LI Y, ZHANG L, WANG Y An optimal method of posture adjustment in aircraft fuselage joining assembly with engineering constraints[J]. Chinese Journal of Aeronautics, 2017, 30 (6): 2016- 2023
doi: 10.1016/j.cja.2017.05.006
6 BUTTERFIELD J, CROSBY S, CURRAN R, et al Optimization of aircraft fuselage assembly process using digital manufacturing[J]. Journal of Computing and Information Science in Engineering, 2007, 7 (3): 269- 275
doi: 10.1115/1.2753879
7 朱永国, 黄翔, 宋利康, 等 基于理想驱动力的中机身调姿多项式轨迹规划[J]. 计算机集成制造系统, 2015, 21 (7): 1790- 1796
ZHU Yong-guo, HUANG Xiang, SONG Li-kang, et al Polynomial trajectory planning method based on ideal drive forces for aircraft fuselage pose adjustment[J]. Computer Integrated Manufacturing Systems, 2015, 21 (7): 1790- 1796
8 刘继红, 庞英仲, 邹成 基于关键特征的飞机大部件对接位姿调整技术[J]. 计算机集成制造系统, 2013, 19 (5): 1009- 1014
LIU Ji-hong, PANG Ying-zhong, ZOU Cheng Adjusting position-orientation of large components based on key features[J]. Computer Integrated Manufacturing Systems, 2013, 19 (5): 1009- 1014
9 邱宝贵, 蒋君侠, 毕运波, 等 大型飞机机身调姿与对接试验系统[J]. 航空学报, 2011, 32 (5): 908- 919
QIU Bao-gui, JIANG Jun-xia, BI Yun-bo, et al Posture alignment and joining test system for large aircraft fuselages[J]. Acta Aeronautica Et Astronautica Sinica, 2011, 32 (5): 908- 919
10 YANG Guo-wei, ZHANG Cheng-jing, ZHANG Xiao-feng. A process quality monitoring approach of automatic aircraft component docking [M] // Computer engineering and networking. Richardson: Springer, 2014: 473–480.
11 FRIEDMAN J H Greedy function approximation: a gradient boosting machine[J]. Annals of Statistics, 2001, 29 (5): 1189- 1232
12 CHEN T, GUESTRIN C. XGBoost: a scalable tree boosting system [C] // ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 785–794.
13 罗芳, 邹方, 周万勇 飞机大部件对接中的位姿计算方法[J]. 航空制造技术, 2011, (3): 91- 94
LUO Fang, ZOU Fang, ZHOU Wan-yong Posture calculating algorithm in large aircraft component butt[J]. Aeronautical Manufacturing Technology, 2011, (3): 91- 94
doi: 10.3969/j.issn.1671-833X.2011.03.017
14 叶志飞, 文益民, 吕宝粮 不平衡分类问题研究综述[J]. 智能系统学报, 2009, 4 (2): 148- 156
YE Zhi-fei, WEN Yi-min, LV Bao-liang A survey of imbalanced pattern classification problems[J]. CAAI Transactions on Intelligent Systems, 2009, 4 (2): 148- 156
[1] 黄杰,王东,王新晴,殷勤,邵发明. 液压挖掘机作业循环状态智能识别方法[J]. 浙江大学学报(工学版), 2019, 53(9): 1663-1673.
[2] 张燕,王建宙,李威,王婕,陈玲玲,杨鹏. 基于数据驱动的膝关节外骨骼控制[J]. 浙江大学学报(工学版), 2019, 53(10): 2024-2033.
[3] 宋筱轩,冯天恒,黄平捷,侯迪波,张光新. 基于动态数据驱动的突发水污染事故仿真方法[J]. 浙江大学学报(工学版), 2015, 49(1): 63-68.