Intelligent fault diagnosis method of spacecraft control system based on sequence data-image mapping

Hanyu Liang; Chengrui Liu; Wenjing Liu; Wenbo Li; Heyu Xu

doi:10.3934/mfc.2022020

Article Contents

2024, Volume 7, Issue 2: 171-194. Doi: 10.3934/mfc.2022020

This issue Previous Article Prescribed performance tracking control of multi-link robotic manipulator with uncertainties Next Article Paint surface estimation and trajectory planning for automated painting systems

Intelligent fault diagnosis method of spacecraft control system based on sequence data-image mapping

Science and Technology on Space Intelligent Control Laboratory, Beijing Institute of Control Engineering, Beijing 100190, China

^*Corresponding author: Chengrui Liu
^*Corresponding author: Chengrui Liu

Received: December 2021

Revised: March 2022

Early access: July 6, 2022

Published online: July 6, 2022

This work was supported by the National Natural Science Funds for Excellent Young Scholars of China under Grant 62022013.

Abstract / Introduction Full Text(HTML) Figure(13) / Table(7) Related Papers Cited by

Abstract

Satellite networking, as the future development direction of aero-space, requires high-precision autonomous fault diagnosis capability for a single satellite. In this paper, aiming at the characteristics of closed-loop fault propagation and high data dimensionality of spacecraft control system, neural network algorithms are conducted to study the fault diagnosis of spacecraft high-dimensional coupled data. Based on the ground test data of a certain spacecraft, this paper converts the high-dimensional sequence data into grayscale images, and then uses Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) to diagnose them respectively. The effectiveness of the methods in this paper is illustrated by comparing and validating them with three non-image-based machine learning algorithms, namely, K-NearestNeighbor, Bayesian classifier, and K-NearestNeighbor based on Principal Component Analysis.

Keywords:

Mathematics Subject Classification: Primary: 93C35, 94C12; Secondary: 68T07.

Citation:

FullText(HTML)

Figure 1. The AOCS of spacecraft

Download: Full-size image PowerPoint slide

Figure 2. The closed-loop characteristic of spacecraft fault

Download: Full-size image PowerPoint slide

Figure 3. Mapping of high-dimensional data to pattern data

Download: Full-size image PowerPoint slide

Figure 4. The architecture of CNN

Download: Full-size image PowerPoint slide

Figure 5. The structure diagram of Long Short-Term Memory network gate unit

Download: Full-size image PowerPoint slide

Figure 6. The image processing principle of LSTM

Download: Full-size image PowerPoint slide

Figure 7. The architecture of LSTM

Download: Full-size image PowerPoint slide

Figure 8. 2D visualization image of Dataset1

Download: Full-size image PowerPoint slide

Figure 9. 2D visualization image of Dataset2

Download: Full-size image PowerPoint slide

Figure 10. The experimental fault diagnosis results on Dataset 1

Download: Full-size image PowerPoint slide

Figure 11. Confusion matrixes of each model

Download: Full-size image PowerPoint slide

Figure 12. The fault diagnosis results for mixed faults

Download: Full-size image PowerPoint slide

Figure 13. Confusion matrixes of each model for mixed faults

Download: Full-size image PowerPoint slide

Table 1. Hyperparameters of CNN

Notation	Description	Kernel size	Stride	Kernel number
Input	Input data	9$ \times $9	$ \backslash $	$ \backslash $
Conv1	Convolution	3$ \times $3	1$ \times $1	32
P1	Max pooling	1$ \times $1	1$ \times $1	32
Conv2	Convolution	3$ \times $3	1$ \times $1	64
P2	Max pooling	3$ \times $3	3$ \times $3	64
F	Fully connected	576$ \times $1	$ \backslash $	$ \backslash $

| Show Table

DownLoad: CSV

Table 2. Fault mode

Number	Fault mode	Fault device
1	Constant deviance fault of CMG	The gimbal angular velocity of the 2nd CMG
2	Noise increase fault of gyroscope	The 1st gyroscope
3	Saturation fault of gyroscope	The 3rd gyroscope
4	Constant deviance fault of gyroscope output	The 4th gyroscope
5	Constant fault of gyroscope output	The 4th gyroscope

| Show Table

DownLoad: CSV

Table 3. Fault type

Number in Dataset 2	Fault mode	Fault device and its amplitude
6	Chord width over-tolerance of infrared earth sensor	The 1st infrared earth sensor and its chord width is 4.538.
7	Chord width over-tolerance of infrared earth sensor	The 1st infrared earth sensor and its chord width is 3.
8	Ground entry angle over-tolerance of infrared earth sensor	The 2nd infrared earth sensor and its ground entry angle is 2.793°.
9	Ground entry angle over-tolerance of infrared earth sensor	The 2nd infrared earth sensor and its ground entry angle is 1.5°.

| Show Table

DownLoad: CSV

Table 4. Model parameter setting of non-image-based algorithms

Name	Setting
KNN	$k$=4 Distance formula: Euclidean distance
NB	Gaussian Bayes
PCA+KNN	PCA: Remain 99.9% of features KNN: $k$=4, Distance formula: Euclidean distance

| Show Table

DownLoad: CSV

Table 5. Results of Dataset 1

Name	Accuracy $ \pm $ standard deviation
KNN	98.61$ \pm $0.32
NB	92.83$ \pm $0.68
PCA+KNN	88.10$ \pm $0.79
CNN	100.00$ \pm $0.01
LSTM	99.96$ \pm $0.69

| Show Table

DownLoad: CSV

Table 6. Model parameter updated setting of non-image-based algorithms

Name	Setting
KNN	$k$=3 Distance formula: Euclidean distance
NB	Gaussian Bayes
PCA+KNN	PCA: Remain 99.99999% of features KNN: $k$=3, Distance formula: Euclidean distance

| Show Table

DownLoad: CSV

Table 7. Mixed fault results

Name	Accuracy $ \pm $ standard deviation
KNN	91.59$ \pm $0.43
NB	98.10$ \pm $0.21
PCA+KNN	84.47$ \pm $0.53
CNN	100.00$ \pm $0.06
LSTM	99.86$ \pm $1.87

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	S. Ai, J. Song and G. Cai, A real-time fault diagnosis method for hypersonic air vehicle with sensor fault based on the auto temporal convolutional network, Aerospace Science and Technology, 119 (2021), 107220. doi: 10.1016/j.ast.2021.107220.
[2]	D. Bergman, B. Glass, K. Zacny and G. Paulsen, Using Distributed Transfer Function method (DTFM) for autonomous health monitoring of interplanetary drills, AIAA SPACE 2015 Conference and Exposition, Pasadena, CA, USA, (2015), 2015–4482. doi: 10.2514/6.2015-4482.
[3]	S. Belagoune, N. Bali, A. Bakdi, B. Baadji and K. Atif, Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems, Measurement, 117 (2021), 109330. doi: 10.1016/j.measurement.2021.109330.
[4]	G. J. Clark, W. Eddy, S. K. Johnson, J. Barnes and D. Brooks, Architecture for cognitive networking within NASA's future space communications infrastructure, 34th AIAA International Communications Satellite Systems Conference, (2016), 2016–5725. doi: 10.2514/6.2016-5725.
[5]	F. Cheng, X. Guo, Y. Qi, J. Xu, W. Qiu, Z. Zhang, W. Zhang and N. Qi, Research on satellite power anomaly detection method based on LSTM, 2021 IEEE International Conference on Power Electronics, Computer Applications (ICPECA), (2021), 20528754. doi: 10.1109/ICPECA51329.2021.9362601.
[6]	M. Daigle, A. Bregon and I. Roychoudhury, Qualitative event-based diagnosis with possible conflicts applied to spacecraft power distribution systems, IFAC Proceedings Volumes, 45 (2012), 265-270. doi: 10.3182/20120829-3-MX-2028.00084.
[7]	A. Falcoz, F. Boquet, M. Dinh, B. Polle, G. Flandin and E. Bornschlegl, Robust fault diagnosis strategies for spacecraft application to LISA Pathfinder experiment, IFAC Proceedings Volumes, 43 (2010), 404-409. doi: 10.3182/20100906-5-JP-2022.00069.
[8]	D. Gao, Y. Zhu, Z. Ren, K. Yan and W. Kang, A novel weak fault diagnosis method for rolling bearings based on LSTM considering quasi-periodicity, Knowledge-Based Systems, 231 (2021), 107413. doi: 10.1016/j.knosys.2021.107413.
[9]	M. Ganesan, R. Lavanya and M. Nirmala Devi, Fault detection in satellite power system using convolutional neural network, Telecommunication Systems, 76 (2021), 505-511. doi: 10.1007/s11235-020-00722-5.
[10]	Q. Hu, G. Niu and C. Wang, Spacecraft attitude fault-tolerant control based on iterative learning observer and control allocation, Aerospace Science and Technology, 75 (2018), 245-253.
[11]	D. Henry, C. Le Peuvédic, L. Strippoli and F. Ankersen, Robust model-based fault diagnosis of thruster faults in spacecraft, IFAC-PapersOnLine, 48 (2015), 1078-1083. doi: 10.1016/j.ifacol.2015.09.670.
[12]	Y. Hua, L. Mou and X. Zhu, Recurrently exploring class-wise attention in a hybrid convolutional and bidirectional LSTM network for multi-label aerial image classification, ISPRS Journal of Photogrammetry and Remote Sensing, 149 (2019), 188-199. doi: 10.1016/j.isprsjprs.2019.01.015.
[13]	M. S. Islam and A. Rahimi, Fault prognosis of satellite reaction wheels using a two-step LSTM network, 2021 IEEE International Conference on Prognostics and Health Management (ICPHM), (2021).
[14]	Y. Li, Q. Hu and X. Shao, Neural network-based fault diagnosis for spacecraft with single-gimbal control moment gyros, Chinese Journal of Aeronautics, 35 (2021), 261-273. doi: 10.1016/j.cja.2021.11.020.
[15]	J. Li, Y. Liu and Q. Li, Intelligent fault diagnosis of rolling bearings under imbalanced data conditions using attention-based deep learning method, Measurement, 189 (2021), 110500. doi: 10.1016/j.measurement.2021.110500.
[16]	Y. Li, X. Du, F. Wan, X. Wang and H. Yu, Rotating machinery fault diagnosis based on convolutional neural network and infrared thermal imaging, Chinese Journal of Aeronautics, 33 (2020), 427-438. doi: 10.1016/j.cja.2019.08.014.
[17]	Y. Liu, Q. Pan, H. Wang and T. He, Fault diagnosis of satellite flywheel bearing based on convolutional neural network, 2019 Prognostics and System Health Management Conference (PHM-Qingdao), (2019), 8942957. doi: 10.1109/PHM-Qingdao46334.2019.8942957.
[18]	D. Long, X. Wen, J. Wang and B. Wei, A data fusion fault diagnosis method based on LSTM and DWT for satellite reaction flywheel, Mathematical Problems in Engineering, 2020 (2020), 2893263. doi: 10.1155/2020/2893263.
[19]	Z. Meng, Y. Zhang, B. Zhu, Z. Pan, L. Cui, J. Li and F. Fan, Research on rolling bearing fault diagnosis method based on ARMA and optimized MOMEDA, Measurement, 189 (2021), 110465. doi: 10.1016/j.measurement.2021.110465.
[20]	H. Ruan, Y. Wang, X. Li, Y. Qin and B. Tang, An enhanced non-local weakly supervised fault diagnosis method for rotating machinery, Measurement, 189 (2021), 110433. doi: 10.1016/j.measurement.2021.110433.
[21]	J. Shi, D. Peng, Z. Peng, Z. Zhang, L. Goebol and D. Wu, Planetary gearbox fault diagnosis using bidirectional-convolutional LSTM networks, Mechanical Systems and Signal Processing, 162 (2022), 107996. doi: 10.1016/j.ymssp.2021.107996.
[22]	S. Tariq, S. Lee and Y. Shin, M. S. Lee, O. Jung, D. Chung and S. S. Woo, Detecting anomalies in space using multivariate convolutional LSTM with mixtures of probabilistic PCA, Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, (2019), 3330776. doi: 10.1145/3292500.3330776.
[23]	R. Wang, Y. Cheng and M. Xu, Analytical redundancy based fault diagnosis scheme for satellite attitude control systems, Journal of the Franklin Institute, 352 (2015), 1906-1931. doi: 10.1016/j.jfranklin.2015.02.003.
[24]	Y. Xia, J. Zhou, T. Xu and W. Gao, An improved deep convolutional neural network model with kernel loss function in image classification, Mathematical Foundations of Computing, 3 (2020), 51-64. doi: 10.3934/mfc.2020005.
[25]	J. Xu, L. Zhou, W. Zhao, Y. Fan and X. Ding, Zero-shot learning for compound fault diagnosis of bearings, Expert Systems with Applications, 190 (2022), 116197. doi: 10.1109/IJCNN52387.2021.9534279.
[26]	B. Xiao and S. Yin, A deep learning-based data-driven thruster fault diagnosis approach for satellite attitude control system, IEEE Transactions on Industrial Electronics, 68 (2020), 10162-10170. doi: 10.1109/TIE.2020.3026272.
[27]	Z. Zhang, S. Li, J. Lu, Y. Xin and H. Ma, Intrinsic component filtering for fault diagnosis of rotating machinery, Chinese Journal of Aeronautics, 34 (2021), 397-409. doi: 10.1016/j.cja.2020.07.019.
[28]	C. Zhang, Y. Liu, F. Wan, B. Chen, J. Liu and B. Hu, Multi-faults diagnosis of rolling bearings via adaptive customization of flexible analytical wavelet bases, Chinese Journal of Aeronautics, 33 (2020), 407-417. doi: 10.1016/j.cja.2019.03.014.
[29]	J. Zhang, Y. Sun, L. Guo, H. Gao, X. Hong and H. Song, A new bearing fault diagnosis method based on modified convolutional neural networks, Chinese Journal of Aeronautics, 33 (2020), 439-447. doi: 10.1016/j.cja.2019.07.011.
[30]	F. Zhou, R. Hang, Q. Liu and X. Yuan, Hyperspectral image classification using spectral-spatial LSTMs, CCCV 2017: Computer Vision, 771 (2017), 557-558. doi: 10.1007/978-981-10-7299-4_48.
[31]	Q. Zhang, J. Zhang, J. Zou and S. Fan, A novel fault diagnosis method based on stacked LSTM, IFAC-PapersOnLine, 53 (2020), 790-795. doi: 10.1016/j.ifacol.2020.12.832.
[32]	F. Zhao, Z. Zhang, M. Hu, Y. Deng and X. Shen, Exo-atmospheric infrared objects classification based on dual-channel LSTM network, Infrared Physics & Technology, 111 (2020), 103535. doi: 10.1016/j.infrared.2020.103535.