基于证据推理和置信规则库的涡扇发动机故障预测

doi:10.16182/j.issn1004731x.joss.21-0396

系统仿真学报 ›› 2022, Vol. 34 ›› Issue (09): 2074-2086.doi: 10.16182/j.issn1004731x.joss.21-0396

• 仿真模型/系统置信度评估技术 • 上一篇下一篇

基于证据推理和置信规则库的涡扇发动机故障预测

朱海龙¹(), 贾如侠¹, 张亮², 贺维¹^,³()

^1.哈尔滨师范大学　计算机科学与信息工程学院，黑龙江　哈尔滨　150025
^2.战略支援部队　航天系统部装备部装备保障队，北京　100094
^3.中国人民解放军火箭军工程大学，陕西　西安　710025

收稿日期:2021-05-06 修回日期:2021-08-15 出版日期:2022-09-18 发布日期:2022-09-23
通讯作者: 贺维 E-mail:zhuhailong2018@vip.163.com;he_w_1980@163.com
作者简介:朱海龙（1972-），男，博士，副教授，研究方向为模式识别、数字图像处理。E-mail：zhuhailong2018@vip.163.com
基金资助:
中国博士后科学基金(2020M683736);黑龙江省自然科学基金(LH2021F038);黑龙江省大学生创新实践项目(202010231009);哈尔滨师范大学博士科研启动金项目(XKB201905);哈尔滨师范大学研究生质量培养提升工程项目(1504120015);哈尔滨师范大学研究生学术创新项目(HSDSSCX2021-29);哈尔滨师范大学计算机科学与信息工程学院自然科学基金(JKYKYZ202102)

Turbofan Engine Fault Prediction Based on Evidential Reasoning and Belief Rule Base

Hailong Zhu¹(), Ruxia Jia¹, Liang Zhang², Wei He¹^,³()

^1.College of Computer Science and Information Engineering, Harbin Normal University, Harbin 150025, China
^2.Equipment support team of equipment department of Space Systems Department, Strategic Support Force, Beijing 100094, China
^3.Rocket Force University of Engineering, Xi’an 710025, China

Received:2021-05-06 Revised:2021-08-15 Online:2022-09-18 Published:2022-09-23
Contact: Wei He E-mail:zhuhailong2018@vip.163.com;he_w_1980@163.com

摘要/Abstract

摘要：

针对某型涡扇发动机故障预测的问题，提出一种基于证据推理和置信规则库的涡扇发动机故障预测模型。为描述涡扇发动机的健康状态，利用证据推理算法融合发动机系统状态信息；结合先验知识建立混合驱动的置信规则库仿真预测模型；采用投影协方差自适应进化策略用于优化模型参数；通过实验验证了模型的有效性。研究结果表明：该方法不仅准确预测涡扇发动机故障风险概率，而且为故障诊断和维修保障提供了有力的支撑。

关键词: 涡扇发动机, 证据推理, 置信规则库, 投影协方差自适应进化, 故障预测

Abstract:

Aiming at the fault prediction problem of a turbofan engine, a fault prediction model based on evidential reasoning (ER) and belief rule base (BRB) is proposed. In order to describe the health state of turbofan engine, ER algorithm is adopted to fuse the state information. Combined with prior knowledge, a hybrid driven simulation prediction of BRB model is established. Projection covariance matrix adaptive evolution strategy (P-CMA-ES) is used to optimize the model parameters. The validity of the model is verified by experiments. Experimental results show that the proposed method not only accurately predicts the probability of failure risk of the turbofan engine, but also provides strong support for fault diagnosis and maintenance support.

Key words: turbofan, evidential reasoning (ER), belief rule base (BRB), projection covariance matrix adaptive evolution strategy (P-CMA-ES), fault prediction

中图分类号:

朱海龙, 贾如侠, 张亮, 贺维. 基于证据推理和置信规则库的涡扇发动机故障预测[J]. 系统仿真学报, 2022, 34(09): 2074-2086.

Hailong Zhu, Ruxia Jia, Liang Zhang, Wei He. Turbofan Engine Fault Prediction Based on Evidential Reasoning and Belief Rule Base[J]. Journal of System Simulation, 2022, 34(09): 2074-2086.

图/表 15

图1

图2

图3

图4

表1

表2

表3

表4

图5

表5

图6

表6

优化后的BRB置信度

θ_l	ϑ_rϑ_s	{(S,β_1,l),(A,β_2,l),(B,β_3,l),(C,β_4,l),(D,β_5,l),(E,β_6,l),}
1	SS	{(S,0.160 8),(A, 0.056 9),(B,0.221 7),(C,0.0181),(D,0.378 7),(E,0.163 7)}
1	AS	{(S,0.390 8),(A,0.138 7),(B,0.147 8),(C,0.039 9),(D,0.045 0),(E,0.237 6)}
1	BS	{(S,0.430 1),(A,0.006 7),(B,0.042 6),(C,0.054 6),(D,0.094 4),(E,0.371 6)}
1	CS	{(S,0.092 0),(A,0.276 2),(B,0.108 9),(C,0.189 7),(D,0.083 5),(E,0.249 7)}
1	DS	{(S,0.181 3),(A,0.172 1),(B,0.360 2),(C,0.244 8),(D,0.017 9),(E,0.023 6)}
1	ES	{(S,0.227 7),(A,0.126 7),(B,0.180 4),(C,0.245 4),(D,0.110 4),(E,0.109 4)}
1	SA	{(S,0.232 9),(A,0.094 5),(B,0.3718),(C,0.076 1),(D,0.224 5),(E,0.000 1)}
1	AA	{(S,0.192 1),(A,0.1381),(B,0.436 3),(C,0.036 1),(D,0.135 2),(E,0.062 2)}
1	BA	{(S,0.161 9),(A,0.034 0),(B,0.054 6),(C,0.214 9),(D,0.084 5),(E,0.450 1)}
1	CA	{(S,0.5281),(A,0.0431),(B,0.114 8),(C,0.192 1),(D,0.021 4),(E,0.100 4)}
1	DA	{(S,0.081 0),(A,0.064 4),(B,0.425 5),(C,0.082 6),(D,0.306 2),(E,0.040 3)}
1	EA	{(S,0.1051),(A,0.292 4),(B,0.050 1),(C,0.087 0),(D,0.383 9),(E,0.081 5)}
1	SB	{(S,0.377 6),(A,0.105 6),(B,0.241 0),(C,0.064 2),(D,0.181 9),(E,0.029 8)}
1	AB	{(S,0.119 9),(A,0.171 0),(B,0.037 5),(C,0.1951),(D,0.115 7),(E,0.360 7)}
1	BB	{(S,0.046 2),(A,0.215 6),(B,0.097 4),(C,0.442 7),(D,0.137 0),(E,0.0611)}
1	CB	{(S,0.037 4),(A,0.104 6),(B,0.069 7),(C,0.074 9),(D,0.127 6),(E,0.585 9)}
1	DB	{(S,0.323 4),(A,0.036 4),(B,0.134 4),(C,0.081 7),(D,0.177 0),(E,0.247 1)}
1	EB	{(S,0.147 5),(A,0.103 6),(B,0.040 5),(C,0.315 8),(D,0.100 0),(E,0.292 5)}
1	SC	{(S,0.131 4),(A,0.278 3),(B,0.315 9),(C,0.157 4),(D,0.089 5),(E,0.027 5)}
1	AC	{(S,0.236 3),(A,0.051 7),(B,0.121 6),(C,0.231 0),(D,0.344 8),(E,0.014 6)}
1	BC	{(S,0.019 1),(A,0.076 3),(B,0.029 8),(C,0.072 0),(D,0.053 8),(E,0.749 0)}
1	CC	{(S,0.227 8),(A,0.0418),(B,0.009 3),(C,0.021 5),(D,0.000 0),(E,0.699 5)}
1	DC	{(S,0),(A,0.005 4),(B,0.016 8),(C,0),(D,0.000 4),(E,0.979 2)}
1	EC	{(S,0.303 6),(A,0.142 7),(B,0.087 6),(C,0.023 9),(D,0.088 6),(E,0.353 6)}
1	SD	{(S,0.2470),(A,0.1759),(B, 0.2034),(C,0.2412),(D,0.0949),(E,0.0375)}

表6

表7

图7

图8

参考文献 24

[1]	韦祥, 李本威, 杨欣毅, 等. 某型涡扇发动机燃调故障联合仿真[J]. 系统仿真学报, 2018, 30(10): 3923-3932.
	Wei Xiang, Li Benwei, Yang Xinyi, et al. Fault Co-simulation of Fuel Regulator in a Certain Type of Turbofan Engine[J]. Journal of System Simulation, 2018, 30(10): 3923-3932.
[2]	蔡景, 胡维, 陈曦. 全修复策略下FADEC系统多故障TLD仿真分析[J]. 航空动力学报, 2020, 35(4): 823-831.
	Cai Jing, Hu Wei, Chen Xi. TLD Simulation Analysis with Multiple Faults for FADEC System under Full Repair Policy[J]. Journal of Aerospace Power, 2020, 35(4): 823-831.
[3]	年夫顺. 关于故障预测与健康管理技术的几点认识[J].仪器仪表学报, 2018, 39(8): 1-14.
	Nian Fushun. Viewpoints About the Prognostic and Health Management[J]. Chinese Journal of Scientific Instrument, 2018, 39(8): 1-14.
[4]	Samir Khan, Takehisa Yairi. A Review on the Application of Deep Learning in System Health Management[J].Mechanical System and Signal Processing (S0888-3270), 2018, 107: 241-265.
[5]	Farzaneh Ahmadzadeh, Jan Lundberg. Remaining Useful Life Estimation: Review[J]. International Journal of System Assurance Engineering and Management (S0975-6809), 2014, 5(4): 461-474.
[6]	焦李成, 杨淑媛, 刘芳, 等. 神经网络七十年：回顾与展望[J]. 计算机学报, 2016, 39(8): 1697-1716.
	Jiao Licheng, Yang Shuyuan, Liu Fang, et al. Seventy Years beyond Neural Networks: Retrospect and Prospect[J]. Chinese Journal of Computers, 2016, 39(8): 1697-1716.
[7]	Li Guofa, Wang Yanbo, He Jialong, et al. Tool Wear State Recognition Based on Gradient Boosting Decision Tree and Hybrid Classification RBM[J]. International Journal of Advanced Manufacturing Technology (S0268-3768), 2020, 110(18): 511-522.
[8]	余嘉熹, 李奇, 陈维荣, 等. 基于随机森林算法的大功率质子交换膜燃料电池系统故障分类方法[J]. 中国电机工程学报, 2020, 40(17): 5591-5599.
	Yu Jiaxi, Li Qi, Chen Weirong, et al. A Fault Classification Method of High-power Proton Exchange Membrane Fuel Cell Systems Based on the Random Forest[J]. CSEE, 2020, 40 (17): 5591-5599.
[9]	郑玉巧, 魏剑峰, 朱凯, 等. 风力机主轴承故障监测方法[J]. 振动.测试与诊断, 2021, 41(2): 341-347,415.
	Zheng Yuqiao, Wei Jianfeng, Zhu Kai, et al. Fault Monitoring Method of Wind Turbine Main Bearing[J].Journal of Vibration, Measurement & Diagnosis, 2021, 41(2): 341-347,415.
[10]	胡爱军, 南冰. 基于自适应概率主成分分析的滚动轴承故障特征增强方法[J]. 振动与冲击, 2017, 36(19): 145-150.
	Hu Aijun, Nan Bing. Fault Feature Enhancement Method for Rolling Bearing Based on Adaptive Probabilistic Principal Component Analysis[J]. Journal of Vibration and Shock, 2017, 36(19): 145-150.
[11]	David Siegel, Jay Lee. An Auto-associative Residual Processing and K-means Clustering Approach for Anemometer Health Assessment[J]. International Journal of Prognostics and Health Management (S2153-2648), 2011, 2(2): 50-61.
[12]	陈志强, 陈旭东, José Valente de Olivira, 等. 深度学习在设备故障预测与健康管理中的应用[J]. 仪器仪表学报, 2019, 40(9): 206-226.
	Chen Zhiqiang, Chen Xudong, Valente de Olivira José, et al. Application of Deep Learning in Equipment Prognostics and Health Management[J]. Chinese Journal of Scientific Instrument, 2019, 40(9): 206-226.
[13]	闫理跃, 王厚军, 刘震. 一种结合噪声辅助技术和现场数据的模拟电路实时可靠度预测方法[J]. 兵工学报, 2019, 40(2): 326-333.
	Yan Liyue, Wang Houjun, Liu Zhen. A Real-time Reliability Prediction Approach for Analog Circuits Based on Noise-Assisted Technique and On-site Data Update[J]. Acta Armamentarii, 2019, 40(2): 326-333.
[14]	郭济鸣, 齐金平, 李兴运. 基于模糊动态故障树的动车制动系统可靠性分析[J]. 中国机械工程, 2019, 30(13): 1585-1589,1599.
	Guo Jiming, Qi Jinping, Li Xingyun. Reliability Analysis of EMUs Braking Systems with Fuzzy Dynamic Fault Tree[J]. China Mechanical Engineering, 2019, 30(13): 1585-1589,1599.
[15]	陈玉昆, 高崎, 陈健, 等. 初始保障期内装备维修器材保障决策方法[J].火力与指挥控制, 2020, 45(2): 23-27.
	Chen Yukun, Gao Qi, Chen Jian, et al. Support Decision Method for Maintenance Supply in Initial Equipment Support Period[J]. Fire Control & Command Control, 2020, 45(2): 23-27.
[16]	谷梦瑶, 陈友玲, 王新龙. 多退化变量下基于实时健康度的相似性寿命预测方法[J]. 计算机集成制造系统, 2017, 23(2): 362-372.
	Gu Mengyao, Chen Youling, Wang Xinlong. Multi-index Modeling for Similarity-Based Residual Life Estimation Based on Real-time Health Degree[J]. Computer Integrated Manufacturing Systems. 2017, 23(2): 362-372.
[17]	黄金泉, 王启航, 鲁峰. 航空发动机气路故障诊断研究现状与展望[J]. 南京航空航天大学学报, 2020, 52(4): 507-522.
	Huang Jinquan, Wang Qihang, Lu Feng. Research Status and Prospect of Gas Path Fault Diagnosis for Aeroengine[J]. Journal of Nanjing University of Aeronautics & Astronautics, 2020, 52(4): 507-522.
[18]	Yang Jianbo, Liu Jun, Wang Jin, et al. Belief Rule-Base Inference Methodology Using the Evidential Reasoning Approach-RIMER[J]. IEEE Transactions on Systems, Man & Cybernetics (S2168-2267): Part A, 2006, 36(2): 266-285.
[19]	Yang Jianbo, Xu Dongling. Evidential Reasoning Rule for Evidence Combination[J]. Artificial Intelligence (S0004-3702), 2013, 205: 1-29.
[20]	He Wei, Yu Chuanqiang, Zhou Guohui, et al. A New Belief-Rule-Based Method for Fault Diagnosis of Wireless Sensor Network[J]. IEEE Access (S2169-3536), 2018, 6: 9404-9419.
[21]	Li Gailing, Zhou Zhijie, Hu Changhua, et al. A New Safety Assessment Model for Complex System Based on the Conditional Generalized Minimum Variance and the Belief Rule Base[J]. Safety Science (S0925-7535), 2017, 93: 108-120.
[22]	Sonmez M, Holt G, Yang J, et al. Applying Evidential Reasoning to Prequalifying Construction Contractors[J].Journal of Management in Engineering (S0742-597X), 2002, 18(3): 111-119.
[23]	Kong Guilan. Belief Rule-Based Inference for Predicting Trauma Outcome[J]. Knowledge-Based Systems (S0950-7051), 2016, 95: 35-44.
[24]	Abhinav Saxena, Kai Goebel, Don Simon, et al. Damage Propagation Modeling for Aircraft Engine Run-to-Failure Simulation[C]// 2008 International Conference on Prognostics and Health Management. Denver, CO, USA:IEEE, 2008: 508.

主要参数	具体信息
收录时间	2008-05-18
收录库名称	Prognostics CoE at NASA
数据收集系统	Commercial Modular Aero-Propulsion System Simulation
发动机推力	90 000磅
环境条件	海平面
数据集大小	4.01 MB

名称	权重	参考值1	参考值2	参考值3	参考值4	参考值5	参考值6
低压压气机出口总温度/ºR	0.10	645	644	643	642	641	640
高压压气机出口总温度/ºR	0.05	1 616	1 600	1 592	1 584	1 573	1 562
低压涡轮出口总温度/ºR	0.10	1 442	1 433	1 421	1 409	1 397	1 377
高压压气机出口静压/psia	0.04	571	568	564	559	554	548
物理风扇转速/(rpm)	0.10	2 389.80	2 389.40	2 389.00	2 388.20	2 387.00	2 386.00
高压涡轮冷气流量/(lbm/s)	0.10	9 235	9 200	9 162	9 120	9 052	9 017
低压涡轮冷气流量/(lbm/s)	0.04	48.5	48.2	47.9	47.5	47.1	46.6
风扇换算转速/(rpm)	0.10	538	535	532	527	523	518
核心机换算转速/(rpm)	0.04	2 389.00	2 388.50	2 388.00	2 387.40	2 386.60	2 386.00
需求风扇转速/(rpm)	0.04	8 290	8 257	8 216	8 183	8 155	8 099
需求风扇换算转速/(rpm)	0.04	8.58	8.50	8.43	8.33	8.21	8.15
燃油油量与高压压气机出口静压的比率/(pps/psi)	0.10	400	398	396	394	391	388
涵道比	0.10	39.90	39.80	39.60	39.40	38.80	38.10
抽气焓	0.05	24.00	23.80	23.70	23.50	23.20	22.80

参考点	参考值	参考点	参考值
S	0	C	0.70
A	0.30	D	0.80
B	0.55	E	1

θ_l	ϑ_rϑ_s	{(S,β_1,l),(A,β_2,l),(B,β_3,l), {(C,β_4,l),(D,β_5,l),(E,β_6,l)}
1	SS	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AS	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BS	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CS	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}
1	DS	{(S,0),(A,0),(B,0),(C,0),(D,1),(E,0)}
1	ES	{(S,0),(A,0),(B,0),(C,0),(D,0),(E,1)}
1	SA	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AA	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BA	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CA	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}
1	DA	{(S,0),(A,0),(B,0),(C,0),(D,1),(E,0)}
1	EA	{(S,0),(A,0),(B,0),(C,0),(D,0),(E,1)}
1	SB	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AB	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BB	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CB	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}
1	DB	{(S,0),(A,0),(B,0),(C,0),(D,1),(E,0)}
1	EB	{(S,0),(A,0),(B,0),(C,0),(D,0),(E,1)}
1	SC	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AC	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BC	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CC	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}
1	DC	{(S,0),(A,0),(B,0),(C,0),(D,1),(E,0)}
1	EC	{(S,0),(A,0),(B,0),(C,0),(D,0),(E,1)}
1	SD	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AD	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BD	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CD	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}
1	DD	{(S,0),(A,0),(B,0),(C,0),(D,1),(E,0)}
1	ED	{(S,0),(A,0),(B,0),(C,0),(D,0),(E,1)}
1	SE	{(S,1),(A,0),(B,0),(C,0),(D,0),(E,0)}
1	AE	{(S,0),(A,1),(B,0),(C,0),(D,0),(E,0)}
1	BE	{(S,0),(A,0),(B,1),(C,0),(D,0),(E,0)}
1	CE	{(S,0),(A,0),(B,0),(C,1),(D,0),(E,0)}

规则状态	匹配度
规则状态	t=334时刻	t=335时刻
S&S	0	0
S&A	0	0.358 4
S&B	0	0.641 6
S&C	0	0
S&D	0	0
S&E	0	0
A&S	0.370 4	0
A&A	0.370 4	0.358 4
A&B	0.370 4	0.641 6
A&C	0.370 4	0
A&D	0.370 4	0
A&E	0.370 4	0
B&S	0.629 6	0
B&A	0.629 6	0.358 4
B&B	0.629 6	0.641 6
B&C	0.629 6	0
B&D	0.629 6	0
B&E	0.629 6	0
C&S	0	0
C&A	0	0.358 4
C&B	0	0.641 6
C&C	0	0
C&D	0	0
C&E	0	0
D&S	0	0
D&A	0	0.358 4
D&B	0	0.641 6
D&C	0	0
D&D	0	0
D&E	0	0
E&S	0	0
E&A	0	0.358 4
E&B	0	0.641 6
E&C	0	0
E&D	0	0
E&E	0	0

基于证据推理和置信规则库的涡扇发动机故障预测

Turbofan Engine Fault Prediction Based on Evidential Reasoning and Belief Rule Base

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 15

参考文献 24

相关文章 2

编辑推荐

Metrics

本文评价

[1]	邱俊杰, 郑红, 程云辉. 基于多尺度LSTM预测模型研究[J]. 系统仿真学报, 2022, 34(7): 1593-1604.
[2]	孟小凡, 宋华. 基于神经网络的卫星姿控系统故障预测[J]. 系统仿真学报, 2019, 31(11): 2499-2508.

方法	MSE
BRB	2.34E–04
BP	5.54E–04
Elman	2.3E–03
RBF	5.59E–04