Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2021, Vol. 55 Issue (3): 548-554    DOI: 10.3785/j.issn.1008-973X.2021.03.015
计算机与控制工程     
FDM型增材制造中送丝机构动态监测与识别
刘晓伟1(),陈赟1,2,*(),张思1,陈康1
1. 江苏科技大学 机械工程学院，江苏 镇江 212003
2. 江苏省船海机械装备先进制造重点实验室，江苏 镇江 212003
Dynamic monitoring and identification of wire feeder in FDM-based additive manufacturing
Xiao-wei LIU1(),Yun CHEN1,2,*(),Si ZHANG1,Kang CHEN1
1. School of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, China
2. Jiangsu Key Laboratory of Advance Manufacturing of Ship and Ocean Machinery Equipment, Zhenjiang 212003, China
 全文: PDF(1180 KB)   HTML
摘要：

为了研究在增材制造过程中，流量比（打印机打印速度与丝材挤出速度之间的比值）的异常状态对喷头阻塞或打印产品分层现象的影响情况，采用加速度振动传感器监测送丝机构中电机的工作状态. 采集打印过程中送丝机构电机不同运动状态的振动信号，利用傅里叶变换方法将时域信号转换成频域信号. 基于频域数据提取表征每组信号间差异的特征值，通过KNN分类算法并引入K折交叉验证，研究特征量以明确故障模式与信号的关系，识别送丝机构的不同运动状态. 实验结果表明，以信号频域数据差异为特征量提出的监测方法对异常流量比的识别准确率达到92.73%.
关键词： 增材制造;  熔融沉积成型（FDM）;  送丝机构;  流量比;  过程监测    
Abstract:

An acceleration vibration sensor was used to monitor the working state of the motor in the wire feeder, in order to study the effect of abnormal flow ratio (the ratio between the printing speed of the printer and the extrusion speed of the wire material) on nozzle blocking or lamination of printed products in the additive manufacturing process. The vibration signal of the wire feeder motor in different motion state during the printing process were collected, and the Fourier transform method was used to convert the time domain signal into frequency domain signal. Based on a frequency domain data, the characteristic value that characterizes the difference between each group of signals was extracted, KNN classification algorithm and K-fold cross-validation were introduced, the characteristic quantity was studied to clarity the relationship between the failure mode and the signal, to identify the different motion states of the wire feeder. Experimental results show that the proposed monitoring method has an accuracy of 92.73% for the identification of abnormal flow ratio, by using the signal frequency domain data difference as the characteristic quantity.
Key words: additive manufacturing    fused deposition modeling (FDM)    filament feeding    flow ratio    process monitoring
收稿日期: 2020-01-16 出版日期: 2021-04-25
CLC:  TP 181  
基金资助: 国家自然科学基金资助项目（51705214，51875003）；江苏省自然科学基金资助项目（BK20170582）
通讯作者: 陈赟     E-mail: issue_ge@foxmail.com;yunchen.just@foxmail.com
作者简介: 刘晓伟（1993—），男，硕士生，从事状态监测研究. orcid.org/0000-0003-0191-2231. E-mail： issue_ge@foxmail.com
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　　
刘晓伟
陈赟
张思
陈康

引用本文:

刘晓伟,陈赟,张思,陈康. FDM型增材制造中送丝机构动态监测与识别[J]. 浙江大学学报(工学版), 2021, 55(3): 548-554.

Xiao-wei LIU,Yun CHEN,Si ZHANG,Kang CHEN. Dynamic monitoring and identification of wire feeder in FDM-based additive manufacturing. Journal of ZheJiang University (Engineering Science), 2021, 55(3): 548-554.

链接本文:

http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2021.03.015    或    http://www.zjujournals.com/eng/CN/Y2021/V55/I3/548

图 1  打印机结构图
图 2  送丝机构监测平台
图 3  KNN分类算法原理图
图 4  K折交叉验证图
真实值 预测值
正例 负例
正例 A B
负例 C D
表 1  二分类混淆矩阵
真实类 预测类
1 2 $ \cdots $ n
1 ${a_{11}}$ ${a_{12}}$ $ \cdots $ ${a_{1n}}$
2 ${a_{21}}$ ${a_{22}}$ $ \cdots $ ${a_{2n}}$
$ \vdots $ $ \vdots $ $ \vdots $ $\vdots$ $ \vdots $
n ${a_{n1}}$ ${a_{n2}}$ $ \cdots $ ${a_{nn}}$
表 2  多分类混淆矩阵
图 5  送丝机构不同运动状态振动信号时域图
图 6  送丝机构不同运动状态振动信号幅频图
图 7  不同K对应的准确率
图 8  10次5折训练后平均结果
图 9  送丝机构不同运动状态信号分类结果
真实值 预测值
类1 类2 类3
类1 54 3 3
类2 2 95 3
类3 0 5 55
表 3  送丝机构不同运动状态信号三分类混淆矩阵
类别 SE/% SP/% AN/%
类1 90.00 98.75 92.73
类2 95.00 93.33 92.73
类3 91.76 96.25 92.73
表 4  三分类后数值统计
1 VYAVAHARE S, TERAIYA S, PANGHAL D, et al Fused deposition modelling: a review[J]. Rapid Prototyping Journal, 2020, 26 (1): 176- 201
doi: 10.1108/RPJ-04-2019-0106
2 鹿芳芳, 朱峰, 陈晓旭, 等 3D打印在汽车行业的应用[J]. 汽车实用技术, 2020, (6): 152- 154
LU Fang-fang, ZHU Feng, CHEN Xiao-xu, et al Appli-cation of 3D printing in automobile industry[J]. Utomo-bile Technology, 2020, (6): 152- 154
3 吴海曦. 面向增材制造的声发射监测技术及应用研究[D]. 杭州: 浙江大学, 2017.
WU Hai-xi. Research on acoustic emission monitoring technology and application for additive manufacturing [D]. Hangzhou: Zhejiang University, 2017.
4 LI Jia-qiang, XIE Hui-min, MA Kang In-situ monitoring of the deformation during Fused Deposition Modeling process using CGS method[J]. Polymer Testing, 2019, 76: 166- 172
doi: 10.1016/j.polymertesting.2019.03.030
5 LIU J, HU Y M, WU B, et al An improved fault diagnosis approach for FDM process with acoustic emission[J]. Journal of Manufacturing Processes, 2018, 35: 570- 579
doi: 10.1016/j.jmapro.2018.08.038
6 KOUSIATZA C, KARALEKAS D In-situ monitoring of strain and temperature distributions during fused deposition modeling process[J]. Materials and Design, 2016, 97: 400- 406
doi: 10.1016/j.matdes.2016.02.099
7 吴海曦, 余忠华, 张浩, 等 面向熔融沉积成型的3D打印机故障声发射监控方法[J]. 浙江大学学报:工学版, 2016, (50): 78- 84
WU Hai-xi, YU Zhong-hua, ZHANG Hao 3D printer fault acoustic emission monitoring method for fused dep-osition molding[J]. Journal of Zhejiang University: Engineering Science, 2016, (50): 78- 84
8 WANG H Y, JING Y C, SONG A L, et al. The design of wire feeding mechanism for desktop 3D printer [P]. Proceedings of the 2018 8th International Conference on Applied Science, Engineering and Technology (ICASET 2018), 2018.
9 LI Z Y, ZHANG D W, SHAO L C, et al Experimental investigation using vibration testing method to optimize feed parameters of color mixing nozzle for fused deposition modeling color 3D printer[J]. Advances in Mechanical Engineering, 2019, 11 (12): 1- 12
10 PENG A H, WANG Z M Researches into influence of process parameters on FDM parts precision[J]. Applied Mechanics and Materials, 2010, 1021: 338- 343
11 RAO P, LIU J, ROBERSON D, et al Online real-time quality monitoring in additive manufacturing processes using heterogeneous sensors[J]. Journal of Manufacturing Science and Engineering, 2015, 137 (6): 1- 12
12 ARDI W, EKRAR W The implementation of fast fourier transform and coherence function to detect the millimetric hole on strip iron plate[J]. Key Engineering Materials, 2020, 5977: 430- 437
13 JAFARIAN K, MOBIN M, JAFARI-MARANDI R, et al Misfire and valve clearance faults detection in the combustion engines based on a multi-Sensor vibration signal monitoring[J]. Measurement, 2018, 128: 527- 536
doi: 10.1016/j.measurement.2018.04.062
14 杨毅明. 数字信号处理: 第2版[M]. 北京: 机械工业出版社, 2017: 69-77.
15 姜皓月, 王晟旻 基于Matlab的FFT算法研究[J]. 电子制作, 2020, (1): 52- 54
JIANG Hao-yue, WANG Sheng-min Research on FFT algorithm based on Matlab[J]. Pratical Eletronics, 2020, (1): 52- 54
doi: 10.3969/j.issn.1006-5059.2020.01.023
16 张浩. 基于HSMM和EEMD的熔融沉积成型3D打印过程故障诊断研究[D]. 杭州: 浙江大学, 2017.
ZHANG Hao. Research on fault diagnosis of fused deposition molding 3D printing process based on HSMM and EEMD [D]. Hangzhou: Zhejiang University, 2017.
17 LIAO T W, KUO R J Five discrete symbiotic organisms search algorithms for simultaneous optimization of feature subset and neighborhood size of KNN classification models[J]. Applied Soft Computing, 2018, 64: 581- 595
doi: 10.1016/j.asoc.2017.12.039
18 CHEN Y, YANG H Multiscale recurrence analysis of long-term nonlinear and nonstationary time series[J]. Chaos, Solitons and Fractals, 2012, 45 (7): 978- 987
doi: 10.1016/j.chaos.2012.03.013
19 李俏, 苏世杰, 陈赟 基于动态时空规整的系泊链闪光焊接在线质量评估[J]. 焊接学报, 2019, 40 (3): 52- 58
LI Qiao, SU Shi-jie, CHEN Yun, et al On-line quality evaluation of flash welding of mooring chain based on dynamic space-time regularization[J]. Transactions of the China Welding Institution, 2019, 40 (3): 52- 58
doi: 10.12073/j.hjxb.2019400071
[1] 李斌,傅建中. 基于三维T样条的异质材料实体建模与切片[J]. 浙江大学学报(工学版), 2021, 55(1): 135-144.
[2] 杜军,马琛,魏正英. 基于视觉传感的铝合金电弧增材沉积层形貌动态响应[J]. 浙江大学学报(工学版), 2020, 54(8): 1481-1489.
[3] 邵惠锋, 贺永, 傅建中. 增材制造可降解人工骨的研究进展——从外形定制到性能定制[J]. 浙江大学学报(工学版), 2018, 52(6): 1035-1057.
[4] 杜军, 王鑫, 任传奇, 白浩. 金属熔融涂覆成形薄壁件表面形貌与工艺[J]. 浙江大学学报(工学版), 2018, 52(1): 24-28.
[5] 吴海曦,余忠华,张浩,杨振生,WANG Yan. 面向熔融沉积成型的3D打印机故障声发射监控方法[J]. 浙江大学学报(工学版), 2016, 50(1): 78-84.
[6] 葛志强 宋执环 杨春节. 基于MCUSUM-ICA-PCA的微小故障检测[J]. J4, 2008, 42(3): 373-377.