一种合成残差式的反作用轮故障检测方法

doi:10.16183/j.cnki.jsjtu.2019.254

摘要/Abstract

摘要：

为了在地面根据卫星遥测数据有效把握在轨卫星反作用轮的健康状态,设计了一种合成残差式的反作用轮故障检测方法.该方法针对在轨遥测数据的特点,根据实际可获得的闭环控制反作用轮遥测数据,利用极端梯度提升(XGBoost)回归模型进行转速预测生成残差,并结合黏性摩擦因数对摩擦力矩突变的敏感性进行故障检测.此外,还针对在轨遥测数据不完备的情况,对检测方法进行了验证.结果表明,该方法对于常见的在轨卫星反作用轮故障均有较好的检测效果.合成残差式故障检测方法不依赖于故障样本,对于数据样本的要求也较低,因此在实际地面故障检测系统中具有一定的应用价值.

关键词: 闭环系统, 反作用轮, 故障检测, 不完备数据, 机器学习

Abstract:

A fault detection method of combined residual is proposed to effectively master the health state of the reaction wheels of in-orbit satellite according to the telemetry data. Based on the characteristics of in-orbit telemetry data, in the proposed method, the XGBoost regression model is used to predict the rotation speed and generate residual with available data of the closed-loop controlled reaction wheel. The sensitivity of viscous friction coefficient to the sudden change of friction torque is also combined with the method. In addition, the fault detection method is verified in view of the incomplete telemetry data in orbit. The result shows that the proposed method has an excellent detection effect on common reaction wheel faults. The combined residual fault detection method does not rely on fault samples and has low requirements for samples, so it has a certain application value in practical fault detection system.

Key words: closed-loop system, reaction wheel, fault detection, incomplete data, machine learning

中图分类号:

V474

何夏维, 蔡云泽, 严玲玲. 一种合成残差式的反作用轮故障检测方法[J]. 上海交通大学学报, 2021, 55(6): 716-728.

HE Xiawei, CAI Yunze, YAN Lingling. A Combined Residual Detection Method of Reaction Wheel for Fault Detection[J]. Journal of Shanghai Jiao Tong University, 2021, 55(6): 716-728.

图/表 17

图1

表1

图2

表2

图3

图4

图5

表3

图6

图7

图8

图9

图10

图11

表4

图12

表5

参考文献 23

[1]	江山. 美“开普勒”望远镜出现重大故障[J]. 太空探索, 2013(7):46-46.
	JIANG Shan. Major failure of Kepler in USA[J]. Space Exploration, 2013(7):46-46.
[2]	曲春娣, 张洪钺. 卫星动量轮闭环系统的UIO双观测器故障诊断[J]. 航天控制, 2005, 23(6):66-71.
	QU Chundi, ZHANG Hongyue. The fault diagnosis of momentum close-loop system used in satellite based on the UIO double observers[J]. Aerospace Control, 2005, 23(6):66-71.
[3]	BALDI P, BLANKE M, CASTALDI P, et al. Fault diagnosis for satellite sensors and actuators using nonlinear geometric approach and adaptive observers[J]. International Journal of Robust and Nonlinear Control, 2019, 29(16):5429-5455. doi: 10.1002/rnc.v29.16 URL
[4]	李知周, 张锐, 朱振才等. 基于扩展卡尔曼滤波的动量轮故障检测方法[J]. 航空学报, 2010, 31(8):1614-1621.
	LI Zhizhou, ZHANG Rui, ZHU Zhencai, et al. Extended Kalman filter-based fault detection for momentum wheel[J]. Acta Aeronautica ET Astronautica Sinica, 2010, 31(8):1614-1621.
[5]	吴红芳, 王志. 基于扩展卡尔曼滤波器的航天器姿控飞轮故障诊断[J]. 航天控制, 2012, 30(3):34-37.
	WU Hongfang, WANG Zhi. The fault diagnosis of spacecraft attitude control flywheel based on extended Kalman filter[J]. Aerospace Control, 2012, 30(3):34-37.
[6]	LIM J K, PARK C G. Satellite fault detection and isolation scheme with modified adaptive fading EKF[J]. Journal of Electrical Engineering and Technology, 2014, 9(4):1401-1410. doi: 10.5370/JEET.2014.9.4.1401 URL
[7]	RAHIMI A, KUMAR K D, ALIGHANBARI H. Fault estimation of satellite reaction wheels using covariance based adaptive unscented Kalman filter[J]. Acta Astronautica, 2017, 134:159-169. doi: 10.1016/j.actaastro.2017.02.003 URL
[8]	余鑫. 基于小波包分析的小飞轮轴承故障诊断方法研究[J]. 洛阳理工学院学报(自然科学版), 2013, 23(1):43-46.
	YU Xin. Fault diagnosis of flywheel bearings based on wavelet package analysis[J]. Journal of Luoyang Institute of Science and Technology (Natural Science Edition) , 2013, 23(1):43-46.
[9]	陆文高, 苏振华, 齐晶等. 基于小波变换的卫星故障诊断方法[J]. 国外电子测量技术, 2018, 37(2):30-33.
	LU Wengao, SU Zhenhua, QI Jing, et al. Method of satellite fault diagnosis based on wavelet translate[J]. Foreign Electronic Measurement Technology, 2018, 37(2):30-33.
[10]	BALDI P, BLANKE M, CASTALDI P, et al. Combined geometric and neural network approach to generic fault diagnosis in satellite reaction wheels[J]. IFAC-PapersOnline, 2015, 48(21):194-199.
[11]	龚学兵, 王日新, 徐敏强. 飞轮传感器的高斯混合模型故障检测方法[J]. 宇航学报, 2015, 36(6):699-705.
	GONG Xuebing, WANG Rixin, XU Minqiang. Gaussian mixed model-based fault detection method for flywheel sensor[J]. Journal of Astronautics, 2015, 36(6):699-705.
[12]	王嘉轶, 闻新, 张婉怡. 卫星群反作用轮的故障诊断方法研究[J]. 中国空间科学技术, 2016, 36(4):24-32.
	WANG Jiayi, WEN Xin, ZHANG Wanyi. Fault diagnosis of reaction wheel in satellite formation[J]. Chinese Space Science and Technology, 2016, 36(4):24-32.
[13]	MOUSAVI S, KHORASANI K. Fault detection of reaction wheels in attitude control subsystem of formation flying satellites: A dynamic neural network-based approach[J]. International Journal of Intelligent Unmanned Systems, 2014, 2(1):2-26. doi: 10.1108/IJIUS-02-2013-0011 URL
[14]	王日新, 龚学兵, 徐敏强, 等. 飞轮系统的符号动力学故障检测方法[J]. 哈尔滨工业大学学报, 2016, 48(10):31-38.
	WANG Rixin, GONG Xuebing, XU Minqiang, et al. A symbolic dynamic analysis of flywheel system for fault detection[J]. Journal of Harbin Institute of Technology, 2016, 48(10):31-38.
[15]	WANG R X, GONG X B, XU M Q, et al. Fault detection of flywheel system based on clustering and principal component analysis[J]. Chinese Journal of Aeronautics, 2015, 28(6):1676-1688. doi: 10.1016/j.cja.2015.10.003 URL
[16]	李磊, 高永明, 吴止锾. 基于混沌吸引子的飞轮故障检测[J]. 北京航空航天大学学报, 2018, 44(9):1894-1902.
	LI Lei, GAO Yongming, WU Zhihuan. Fault detection for flywheels based on chaotic attractor[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(9):1894-1902.
[17]	赵琳, 王艺鹏, 郝勇. 在轨飞轮故障诊断混合框架设计[J]. 光学精密仪器, 2018, 26(7):1728-1740.
	ZHAO Lin, WANG Yipeng, HAO Yong. Design of hybrid frame for on-orbit flywheel fault diagnosis[J]. Optics and Precision Engineering, 2018, 26(7):1728-1740.
[18]	BIALKE B. High fidelity mathematical modeling of reaction wheel performance[J]. Advances in the Astronautical Sciences, 1998, 98:483-496.
[19]	TALBI H A, PATEL R V, KHORASANI K. Fault detection and isolation for uncertain nonlinear systems with application to a satellite reaction wheel actuator[C]// IEEE International Conference on Systems. Montreal, Canada: IEEE, 2007: 3140-3145.
[20]	刘翔. 卫星姿态控制系统故障模式分析与故障诊断研究[D]. 哈尔滨: 哈尔滨工业大学, 2015.
	LIU Xiang. Study on fault mode analysis and fault diagnosis of satellite attitude control system[D]. Harbin: Harbin Institute of Technology, 2015.
[21]	LIU R N, YANG B Y, ZIO E, et al. Artificial intelligence for fault diagnosis of rotating machinery: A review[J]. Mechanical Systems and Signal Processing, 2018, 108:33-47. doi: 10.1016/j.ymssp.2018.02.016 URL
[22]	CHEN T Q, GUESTRIN C. XGBoost: A scalable tree boosting system[C]// Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: ACM, 2016: 785-794.
[23]	IZADI V, ABEDI M, BOLANDI H. Supervisory algorithm based on reaction wheel modelling and spectrum analysis for detection and classification of electromechanical faults[J]. IET Science, Measurement & Technology, 2017, 11(8):1085-1093. doi: 10.1049/smt2.v11.8 URL

序号	参数	数值
1	反作用轮惯量, J/(kg·m²)	0.0229
2	电机电动势反馈常数,K_e/(V·rad^-1·s)	0.029
3	输入阻抗,R_in/Ω	2
4	母线电压,V_bus/V	28
5	偏置电压,V_bias/V	6
6	电压反馈增益K_f	0.5
7	驱动增益,G_d/(A·V^-1)	0.37
8	驱动频率,ω_d/(rad·s^-1)	2000
9	电机转矩参数,K_t/(N·m·A^-1)	0.054
10	超速循环增益,K_s/(V·rad^-1·s)	95
11	限制转速,ω_s/(r·min^-1)	5000
12	转矩噪声角偏差,θ_a/rad	0.05
13	高通滤波器频率ω_a/(rad·s^-1)	0.2
14	库伦摩擦因数τ_c	0.00477

代号	故障模式	故障表现	故障原因
F1	卡死故障	输出力矩首先产生一个巨大的反向扰动,然后快速变为零	电子转子不启动、电机转子和定子抱死、转轴断裂、电路短路
F2	空转故障	无法正常响应控制指令,在摩擦力矩作用下转速减小,输出力矩几乎为零	电子线路、驱动电机或是电源供应故障
F3	摩擦故障	摩擦力矩增大导致反作用轮输出力矩小于控制力矩,影响控制效果	轴承温度增高、轴承润滑差、保持架不稳定、轮体真空密封失效、压力降低
F4	增益下降故障	输出力矩相对期望力矩比例减小	驱动电机故障、元器件老化失效
F5	缓变故障	摩擦力矩随时间缓慢增加	故障原因与F3摩擦故障一致
F6	跳变故障	转速波动,产生非预期力矩	母线电压故障或存在间歇性时变故障

模型	R²	MAE	MSE	RMSE
LASSO	0.9991	0.3625	0.2047	0.4524
RIDGE	0.9991	0.3624	0.2046	0.4524
ENet	0.9991	0.3625	0.2047	0.4524
KRR	0.9992	0.3529	0.1933	0.4397
SVR	0.9993	0.3186	0.1717	0.4144
XGBoost	0.9995	0.2554	0.1071	0.3273

故障	参数设置	SP/%	Acc/%
F1		100	100
F2		99.17	99.59
F3	K₃=1.32	79.05	88.12
F3	K₃=1.35	82.58	90.13
F3	K₃=1.43	98.31	98.68
F4	K₄=0.83	80.18	88.32
F4	K₄=0.79	98.39	98.80
F5	β₅=0.21×10^-3	88.48	88.48
F6		99.92	98.80

故障	频率是否一致	SP/%	Acc/%	故障	频率是否一致	SP/%	Acc/%
F1	是	100	100	F4(K₄=0.79)	是	98.39	98.80
F1	否	98.59	99.30	F4(K₄=0.79)	否	83.79	90.33
F2	是	99.17	99.59	F5(β₅=0.21×10^-3)	是	88.48	88.48
F2	否	99.17	99.59	F5(β₅=0.21×10^-3)	否	87.52	87.52
F3(K₃=1.43)	是	98.31	98.68	F6	是	99.92	98.80
F3(K₃=1.43)	否	81.22	89.04	F6	否	100	98.68