故障机理与领域自适应混合驱动的机械故障智能迁移诊断

doi:10.7527/S1000-6893.2021.26800

Abstract

Abstract:

The health and stability of aircraft engines play an important role in ensuring the safe operation of aircraft， and the establishment of intelligent diagnostic models with high accuracy for different engines is the key to the stable operation of aircraft. The existing fault diagnosis methods can achieve good results when fault data are available， but the actual application cannot always realize the construction of the diagnosis model because it only contains normal data. To address this problem， a hybrid fault mechanism and domain adaptive driven intelligent migration diagnosis method for mechanical faults is proposed. This method firstly establishes a virtual sample generation model for rotating mechanical faults based on the fault mechanism and source domain data， then the normal data of the target domain is used to realize the adaption of the generation model to the target domain， and finally， the target domain fault diagnosis model is obtained by virtual sample training. The proposed method is validated using standard data sets and laboratory bearing data， and the results show that the proposed method achieves an average accuracy of 88.61% when diagnosing different types of bearings， which is 41.22% higher compared with the comparison method.

Key words: fault diagnosis, mechanism model, personalized model, transfer learning, domain adaptation

CLC Number:

Gongye YU, Weidong CAI, Minghui HU, Wencai LIU, Bo MA. Intelligent migration diagnosis of mechanical faults driven by hybrid fault mechanism and domain adaptation[J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(2): 426800-426800.

Figures/Tables 16

Fig. 1

Fig. 2

Algorithm 1　Pseudo code of calculation method for G(⋅)

算法名称： $G (⋅)$
输入：频率分辨率 $τ$ 、 $f$ 、 $f'$ 、 $Ψ s$
输出：单个虚拟样本的 $G (⋅)$
1. 从 $Ψ s$ 随机采样得到一组迁移参数向量 $(k, θ, φ)$
2. $H ← 0 ∈ R 1 × L$ /零矩阵/
3. For $i ← 1$ to $L / (f / τ)$ /* $⋅$ 表示向下取整*/
4. $j = < i × (f / τ) >$ /* $< ⋅ >$ 表示四舍五入*/
5. 从正态分布 $N (1, 0.1)$ 随机采样得到 $λ$
6. $H [j] ← k × θ i - 1 × λ$ /为故障缺陷频率及其倍频处赋值/
7. $q l = < j - (f' / τ) >, q r = < j - (f' / τ) >$
8. If $q l > 0$ Then
9. 从正态分布 $N (1, 0.1)$ 随机采样得到 $λ$
10. $H [q l] = H [j] × φ × λ$ /为左侧边带赋值/
11. If $q r ≤ L$ Then
12. 从正态分布 $N (1, 0.1)$ 随机采样得到 $λ$
13. $H [q r] = H [j] × φ × λ$ /为右侧边带赋值/
14. Return $G (⋅) = H$

Fig. 3

Table 1

Structural parameters of fault diagnosis model

层名称	参数尺寸	激活函数	是否添加BN层	是否添加Dropout
I0	$1 024 × 1$
C1	$64 × 20 × 1 × 1 × 4$	ReLU^［24］	是	否
P2	$2 × 1 × 2$		否
C3	$128 × 3 × 1 × 64 × 2$	ReLU^［24］	是	否
P4	$2 × 1 × 2$		否
C5	$192 × 3 × 1 × 128 × 2$	ReLU^［24］	是	否
P6	$2 × 1 × 2$		否
C7	$256 × 3 × 1 × 192 × 1$	ReLU^［24］	是	否
P8	$2 × 1 × 2$		否
F9	$M × 512$	ReLU^［24］	是	是（0.5）
F10	$512 × (V + 1)$		是	是（0.5）
S11	$V + 1$		否
O12

Table 1

Table 2

Fig. 4

Table 3

List of diagnostic tasks

源域	目标域	诊断任务
A	B	$A → B$
A	C	$A → C$
A	D	$A → D$
B	A	$B → A$
B	C	$B → C$
B	D	$B → D$
C	D	$C → D$
D	C	$D → C$

Table 3

Table 4

Common parameters of bearing failure

故障类型	ORF	IRF	BF
共性参数	$Κ 、 Θ$	$Κ 、 Θ 、 Φ$	$Κ 、 Θ 、 Φ$

Table 4

Fig. 5

Fig. 6

Table 5

Diagnostic accuracy of ten-fold crossover experiment （%）

实验次数	$A → B$	$A → C$	$A → D$	$B → A$	$B → C$	$B → D$	$C → D$	$D → C$
1	62.83	99.91	94.40	72.45	99.23	93.13	94.98	99.43
2	61.92	99.69	99.36	70.93	98.21	98.02	92.37	96.74
3	59.32	99.20	82.29	69.91	99.67	86.69	93.66	99.81
4	62.87	99.65	98.62	68.32	98.25	81.04	91.87	99.96
5	58.95	99.58	99.89	71.38	98.36	98.13	91.35	99.16
6	64.22	99.85	98.24	70.16	97.91	96.47	92.39	97.99
7	58.79	99.72	98.64	69.94	98.28	82.69	90.98	99.33
8	63.74	99.99	98.71	68.33	99.16	93.40	92.51	98.91
9	61.18	99.85	95.82	68.37	99.02	97.33	91.35	99.16
10	60.06	99.58	96.38	70.31	99.06	89.04	92.94	97.31
平均	61.39	99.70	96.24	70.01	98.72	91.60	92.44	98.78

Table 5

Table 6

Diagnostic accuracy under different common parameter distributions （%）

单变量 $k$ $(θ, φ) ∈ (Θ, Φ) k ∈ N (r k, 0.1 × r k 2)$	$r k$	0.01	0.1	1	10	100	1 000	10 000	100 000	0.01	0.1
单变量 $k$ $(θ, φ) ∈ (Θ, Φ) k ∈ N (r k, 0.1 × r k 2)$	准确率	36.97	32.29	40.79	85.64	74.29	46.78	45.58	43.30	36.97	32.29
单变量 $θ$ $(k, φ) ∈ (Κ, Φ) θ ∈ N (r θ, 0.1 × r θ 2)$	$r θ$	0.000 1	0.001	0.01	0.1	0.5	1	5	10	0.000 1	0.001
单变量 $θ$ $(k, φ) ∈ (Κ, Φ) θ ∈ N (r θ, 0.1 × r θ 2)$	准确率	71.43	75.72	77.74	69.18	77.58	69.91	46.66	38.04	71.43	75.72
单变量 $φ$ $(k, θ) ∈ (Κ, Θ) φ ∈ N (r φ, 0.1 × r φ 2)$	$r φ$	0.000 1	0.001	0.01	0.1	0.5	1	5	10	0.000 1	0.001
单变量 $φ$ $(k, θ) ∈ (Κ, Θ) φ ∈ N (r φ, 0.1 × r φ 2)$	准确率	83.86	82.53	84.11	81.58	77.71	73.71	73.55	68.84	83.86	82.53
三变量 $k 、 θ 、 φ$ $k ∈ N (r k, 0.1 × r k 2) 、 θ ∈ N (r θ, 0.1 × r θ 2) 、 φ ∈ N (r φ, 0.1 × r φ 2)$						$r k = 10$ 、 $r θ = 0.01$ 、 $r φ = 0.01$
						准确率	72.17

Table 6

Table 7

Method description

方法名称	描述
DA-MCGM	模型输入为基于希尔伯特变换的包络频谱
TCA	模型输入为文献［27］的42维特征，核心代码来自文献［28］
GFK	模型输入同TCA方法，核心代码来自文献［29］
DaNN	模型输入和结构与DA-MCGM方法一致，模型训练时的损失函数添加了最大均值差异^［30］
Fine-tune	模型输入和结构与DA-MCGM方法一致，先采用 $D s$ 训练模型，然后冻结C1~P8参数，采用 $D 0 t$ 对F9~F10的参数微调
CNN	模型输入和结构与DA-MCGM方法一致，采用 $D s$ 训练模型，然后直接用训练完成的模型对 $D t$ 测试，可作为基准模型

Table 7

Table 8

Comparison of test results accuracy （%）

诊断任务	$A → B$	$A → C$	$A → D$	$B → A$	$B → C$	$B → D$	$C → D$	$D → C$	平均
DA-MCGM	61.39	99.70	96.24	70.01	98.72	91.60	92.44	98.78	88.61
TCA	42.84	20.86	39.99	44.63	24.46	32.11	51.53	59.51	39.49
GFK	44.22	31.42	42.27	45.97	37.21	38.83	53.69	60.28	44.24
DaNN	57.96	35.67	41.66	59.36	27.74	36.32	60.97	63.43	47.39
FIne-tune	40.14	33.33	37.25	41.79	33.33	37.17	53.64	57.72	41.80
CNN	52.04	18.61	28.86	56.43	22.92	29.66	56.32	63.09	40.99

Table 8

Table 9

Diagnostic accuracy of DA-MCGM method for various types of failures （%）

诊断任务	ORF	IRF	BF
$A → B$	78.08	37.96	40.00
$A → C$	100.00	98.58
$A → D$	100.00	91.80
$B → A$	81.36	54.60	44.71
$B → C$	99.92	100.00
$B → D$	93.81	89.51
$C → D$	100.00	99.33
$D → C$	96.33	100

Table 9

References 32

1	SUN H B, WANG J, CHEN K. A tip clearance prediction model for multistage rotors and stators in aero-engines [J]. Chinese Journal of Aeronautics, 2021, 34(2): 343-357.
2	王辰, 左彦飞, 江志农, 等. 全转速系数矩阵降维重构的燃机不平衡量逆推方法[J]. 航空学报, 2020, 41(11): 223670.
	WANG C, ZUO Y F, JIANG Z N, et al. A backstepping method of gas turbine unbalance vector based on dimension reduction and reconstruction of full speed coefficient matrix[J]. Acta Aeronautica et Astronautica Sinica, 2020, 41(11): 223670 (in Chinese).
3	孙灿飞, 王友仁. 直升机行星传动轮系故障诊断研究进展[J]. 航空学报, 2017, 38(7): 111-124.
	SUN C F, WANG Y R. Advance in study of fault diagnosis of helicopter planetary gears[J]. Acta Aeronautica et Astronautica Sinica, 2017, 38(7): 111-124 (in Chinese).
4	MA S J, CHENG B, SHANG Z W. Scattering transform and LSPTSVM based fault diagnosis of rotating machinery[J]. Mechanical Systems and Signal Processing, 2018, 104: 155-170.
5	ZHANG Y H, ZHOU T T, HUANG X F, et al. Fault diagnosis of rotating machinery based on recurrent neural networks[J]. Measurement, 2020, 171: 108774.
6	王庆锋, 刘家赫, 卫炳坤, 等. 数据驱动的聚类分析故障识别方法研究[J]. 机械工程学报, 2020, 56(18):7-14.
	WANG Q F, LIU J H, WEI B K, et al. Research on data-driven clustering analysis fault identification method[J]. Journal of Mechanical Engineering, 2020, 56(18): 7-14 (in Chinese).
7	沈长青, 汤盛浩, 江星星, 等. 独立自适应学习率优化深度信念网络在轴承故障诊断中的应用研究[J]. 机械工程学报, 2019, 55(7): 81-88.
	SHEN C Q, TANG S H, JIANG X X, et al. Bearings fault diagnosis based on improved deep belief network by self-individual adaptive learning rate[J]. Journal of Mechanical Engineering, 2019, 55(7): 81-88 (in Chinese).
8	沈飞, 陈超, 徐佳文, 等. 基于时间迁移模型的旋转机械实时故障诊断[J]. 仪器仪表学报, 2019, 40(10): 84-94.
	SHEN F, CHEN C, XU J W, et al. Time transfer model based rotating machine real-time fault diagnosis[J]. Chinese Journal of Scientific Instrument, 2019, 40(10): 84-94 (in Chinese).
9	LEI Y G, YANG B, JIANG X W, et al. Applications of machine learning to machine fault diagnosis: a review and roadmap[J]. Mechanical Systems and Signal Processing, 2020, 138: 106587.
10	LI C, ZHANG S H, QIN Y, et al. A systematic review of deep transfer learning for machinery fault diagnosis[J]. Neurocomputing, 2020, 407: 121-135.
11	徐颖强, 陈仙亮, 曹栋波. 样本量为2的极小样本相容性检验方法[J]. 航空学报, 2018, 39(5): 144-151.
	XU Y Q, CHEN X L, CAO D B. Compatibility test method in minimal samples situation with two samples[J]. Acta Aeronautica et Astronautica Sinica, 2018, 39(5): 144-151 (in Chinese).
12	CHEN C, LI Z H, YANG J, et al. A cross domain feature extraction method based on transfer component analysis for rolling bearing fault diagnosis[C]∥2017 29th Chinese Control and Decision Conference (CCDC), 2017.
13	雷亚国, 杨彬, 杜兆钧, 等. 大数据下机械装备故障的深度迁移诊断方法[J]. 机械工程学报, 2019, 55(7): 1-8.
	LEI Y G, YANG B, DU Z J, et al. Deep transfer diagnosis method for machinery in big data era[J]. Journal of Mechanical Engineering, 2019, 55(7): 1-8 (in Chinese).
14	康守强, 邹佳悦, 王玉静, 等. 基于无监督特征对齐的变负载下滚动轴承故障诊断方法[J]. 中国电机工程学报, 2020, 40(1): 274-281.
	KANG S Q, ZOU J Y, WANG Y J, et al. Fault diagnosis method of a rolling bearing under varying loads based on unsupervised feature alignment[J]. Proceedings of the CSEE, 2020, 40(1): 274-281 (in Chinese).
15	LIAO Y X, HUANG R Y, LI J P, et al. Deep semisupervised domain generalization network for rotary machinery fault diagnosis under variable speed[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(10): 8064-8075.
16	SHEN F, CHEN C, YAN R Q, et al. Bearing fault diagnosis based on SVD feature extraction and transfer learning classification[C]∥2015 Prognostics and System Health Management Conference, 2015.
17	邵海东, 张笑阳, 程军圣, 等. 基于提升深度迁移自动编码器的轴承智能故障诊断[J]. 机械工程学报, 2020, 56(9): 84-90.
	SHAO H D, ZHANG X Y, CHENG J S, et al. Intelligent fault diagnosis of bearing using enhanced deep transfer auto-encoder[J]. Journal of Mechanical Engineering, 2020, 56(9): 84-90 (in Chinese).
18	罗嘉宁. 实验与仿真数据驱动的滚动轴承故障严重性评估研究[D]. 哈尔滨: 哈尔滨工业大学, 2019.
	LUO J N. Research on experimental and simulation data-driven fault severity assessment of rolling element bearings[D]. Harbin: Harbin Institute of Technology, 2019 (in Chinese).
19	董韵佳. 基于动力学仿真和迁移学习的滚动轴承故障诊断方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2019.
	DONG Y J. Investigation of dynamic simulation and transfer learning based fault diagnosis method for rolling element bearings[D]. Harbin: Harbin Institute of Technology, 2019 (in Chinese).
20	马波, 蔡伟东, 赵大力. 基于GAN样本生成技术的智能诊断方法[J]. 振动与冲击, 2020, 39(18): 153-160.
	MA B, CAI W D, ZHAO D L. Intelligent diagnosis method based on GAN sample generation technology[J]. Journal of Vibration and Shock, 2020, 39(18): 153-160 (in Chinese).
21	袁壮, 董瑞, 张来斌, 等. 深度领域自适应及其在跨工况故障诊断中的应用[J]. 振动与冲击, 2020, 39(12): 286-293.
	YUAN Z, DONG R, ZHANG L B, et al. Deep domain adaptation and its application in fault diagnosis across working conditions[J]. Journal of Vibration and Shock, 2020, 39(12): 286-293 (in Chinese).
22	柏壮壮, 卢一相, 高清维, 等. 基于自适应紧框架学习的轴承故障诊断[J]. 振动与冲击, 2021, 40(12): 296-303.
	BAI Z Z, LU Y X, GAO Q W, et al. Bearing fault diagnosis based on adaptive tight frame learning[J]. Journal of Vibration and Shock, 2021, 40(12): 296-303 (in Chinese).
23	董光玲, 姚郁, 贺风华, 等. 制导精度一体化试验的Bayesian样本量计算方法[J]. 航空学报, 2015, 36(2): 575-584.
	DONG G L, YAO Y, HE F H, et al. Bayesian sample size determination for integrated test of missile hit accuracy[J]. Acta Aeronautica et Astronautica Sinica, 2015, 36(2): 575-584 (in Chinese).
24	GLOROT X, BORDES A, BENGIO Y. Deep sparse rectifier neural networks[J]. Journal of Machine Learning Research, 2011, 15: 315-323.
25	SMITH W A, RANDALL R B, Randall. Rolling element bearing diagnostics using the case western reserve university data: a benchmark study[J]. Mechanical Systems ＆ Signal Processing, 2015, 64-65: 100 -131.
26	LEE D, SIU V, CRUZ R, et al. Convolutional neural net and bearing fault analysis[C]∥Proceedings of the International Conference on Data Mining, 2016: 194-200.
27	刘海宁, 宋方臻, 窦仁杰, 等. 小数据条件下基于测地流核函数的域自适应故障诊断方法研究[J]. 振动与冲击, 2018, 37(18): 36-42.
	LIU H N, SONG F Z, DOU R J, et al. Domain adaptive fault diagnosis based on the geodesic flow kernel under small data condition[J]. Journal of Vibration and Shock, 2018, 37(18): 36-42 (in Chinese).
28	PAN S J, TSANG I W, KWOK J T, et al. Domain adaptation via transfer component analysis[J]. IEEE Transactions on Neural Networks, 2011, 22 (2): 199-210.
29	GONG B Q, SHI Y, SHA F, et al. Geodesic flow kernel for unsupervised domain adaptation[C]∥2012 IEEE Conference on Computer Vision and Pattern Recognition, 2015.
30	CHEN C, SHEN F, XU J W, et al. Domain adaptation-based transfer learning for gear fault diagnosis under varying working conditions[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 3500510.
31	CHEN X H, ZHANG B, GAO D, et al. Bearing fault diagnosis base on multi-scale CNN and LSTM model[J]. Journal of Intelligent Manufacturing, 2021, 32: 971-987.
32	杨国安. 滚动轴承故障诊断实用技术[M]. 北京: 中国石化出版社, 2012.
	YANG G A. Techniques for rolling bearing fault diagnosis[M]. Beijing: China Petrochemical Press, 2012 (in Chinese).

数据集	轴承型号	数据来源	数据描述	健康状态	标签	样本数
A	SKF6205	CWRU	每种健康状态包含3种故障尺寸、4种负载	N	0	2 871
				ORF	1	2 871
				IRF	2	2 871
				BF	3	2 871
B	SKF6203	CWRU	每种健康状态包含3种故障尺寸、4种负载	N	0	1 429
				ORF	1	1 429
				IRF	2	1 429
				BF	3	1 429
C	未知	MFPT	故障状态包含7种负载	N	0	2 016
				ORF	1	2 016
				IRF	2	2 016
D	N205EM	实验台	每种健康状态转速固定	N	0	300
				ORF	1	300
				IRF	2	300

[1]	Minghui HU, Jinji GAO, Zhinong JIANG, Weimin WANG, Limin ZOU, Tao ZHOU, Yunfeng FAN, Yue WANG, Jiaxin FENG, Chenyang LI. Research progress on vibration monitoring and fault diagnosis for aero-engine [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(4): 630194-630194.
[2]	Zhanjun HUANG, Xin DONG, Muyu LU, Ruitao ZHANG, Zhaoyang YAN, An ZHANG. Fault diagnosis of inverter of aviation HVDC sysytem based on DRSN and voltage amplitude analysis [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(3): 328685-328685.
[3]	Tianqing LI, Weimin WANG, Xulong ZHANG, Shuhui WANG, Zhenyu FU. Identification method of rotor blade axial displacement based on blade tip timing [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(2): 228682-228682.
[4]	Jie LI, Wenxin HUANG, Yiming CAI, Siyuan WANG, Yufei GAO, Xuefeng JIANG. Fault diagnosis and fault tolerant control of position sensor based on DFPMM [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(10): 329307-329307.
[5]	Baohui JIA, Fan JIANG, Yuxin WANG, Du WANG. Fault diagnosis method based on civil aircraft maintenance text data [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(5): 326598-326598.
[6]	Jinrui WANG, Shanshan JI, Zongzhen ZHANG, Zhenyun CHU, Baokun HAN, Huaiqian BAO. Parallel sparse filtering for fault diagnosis under bearing acoustic signal [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(4): 426887-426887.
[7]	Zehao CHEN, Hui CHEN, Yushan GAO, Hang ZHANG. Review and prospect of model-based fault diagnosis technology for liquid rocket engines [J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(23): 629016-629016.
[8]	Haowen LUO, Shaoming HE, Tianyu JIN, Zichao LIU. Impact-angle-constrained with time-minimum guidance algorithm based on transfer learning [J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(19): 328400-328400.
[9]	Lianzi WANG, Ling WANG, Daiyin ZHU. An ISAR autofocus imaging algorithm based on FCN and transfer learing [J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(17): 328172-328172.
[10]	Chenghao GUO, Jinsong YU, Yue SONG, Qi YIN, Jiaxuan LI. Application of digital twin⁃based aircraft landing gear health management technology [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(11): 227629-227629.
[11]	Fangli WANG, Kai LIU, Wei PAN, Mingbo TONG. Application and development of green structure maintenance for civil aircraft [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(11): 25851-025851.
[12]	Wanli ZHAO, Yingqing GUO, Kejie XU, Cansen WANG, Haojie YING, Xinxin TAO. Review of key technologies for fault diagnosis and accommodation for multi⁃electric distributed engine control system [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(10): 27519-027519.
[13]	ZHANG Zhongqiang, ZHANG Xin, WANG Jiaxu, LIU Zhiwen. Reweighted kurtogram for aero-engine fault diagnosis [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2022, 43(9): 625445-625445.
[14]	HAN Songyu, SHAO Haidong, JIANG Hongkai, ZHANG Xiaoyang. Intelligent fault diagnosis of aero-engine high-speed bearings using enhanced CNN [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2022, 43(9): 625479-625479.
[15]	ZHANG Shuo, LIU Zhiwen. Structured Bayesian sparse representation of aero-engine bearing failures [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2022, 43(9): 625056-625056.

Intelligent migration diagnosis of mechanical faults driven by hybrid fault mechanism and domain adaptation

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 16

References 32

Related Articles 15

Recommended Articles

Metrics

Comments