基于SSAE和相似性匹配的航空发动机剩余寿命预测

王昆; 郭迎清; 赵万里; 周启凡; 郭鹏飞

doi:10.13700/j.bh.1001-5965.2021.0741

基于SSAE和相似性匹配的航空发动机剩余寿命预测

doi: 10.13700/j.bh.1001-5965.2021.0741

西北工业大学动力与能源学院，西安 710129

基金项目: 国家科技重大专项(J2019-V-0003-0094)

详细信息

通讯作者:
E-mail：yqguo@nwpu.edu.cn

中图分类号: V233
计量
- 文章访问数: 584
- HTML全文浏览量: 69
- PDF下载量: 37
- 被引次数: 0
出版历程
- 收稿日期: 2021-12-08
- 录用日期: 2022-03-04
- 网络出版日期: 2022-03-22
- 整期出版日期: 2023-10-31

Remaining useful life prediction of aeroengine based on SSAE and similarity matching

School of Power and Energy，Northwestern Polytechnical University，Xi’an 710129，China

Funds: National Science and Technology Major Project (J2019-V-0003-0094)

More Information

Corresponding author: E-mail：yqguo@nwpu.edu.cn

摘要

摘要:
航空发动机作为高度复杂的热力机械，其剩余寿命(RUL)预测往往作为提高安全性和经济性的重要保障。为了提高航空发动机剩余寿命预测精度，提出一种基于堆栈稀疏自编码器(SSAE)及相似性匹配的剩余寿命预测方法。以Spearman秩相关系数(SRCC)作为适应度函数，利用遗传算法(GA)对融合参数候选集进行寻优；采用SSAE的结构融合最优参数集，生成特征融合指标；采用相似性匹配的方法在历史数据库内全局搜索最优匹配的历史轨迹，得到寿命预测结果；采用美国国家航空航天局(NASA)公布的C-MAPSS数据集验证该融合指标和方法的有效性。
- 航空发动机 /
- 剩余寿命 /
- 堆栈稀疏自编码器 /
- Spearman秩相关系数 /
- 相似性匹配
Abstract:
As a highly complex thermal machinery, the prognosis of the remaining useful life (RUL) of an aero-engine is often used as an important guarantee to improve safety and economy. In order to increase the engine’s remaining usable life prediction accuracy, a strategy based on stacked sparse autoencoders (SSAE) and similarity matching is proposed in this study. Firstly, Spearman’s rank correlation coefficient (SRCC) is utilized as a fitness function and optimizes the candidate set of fusion parameters through a genetic algorithm (GA). The SSAE fuses the optimal parameter set in order to generate the feature comprehensive index. The results of the life prediction are then obtained by using the similarity matching approach to search the history database worldwide for the best matching trajectory. Finally, the C-MAPSS dataset published by the National Aeronautics and Space Administration (NASA) is obtained to verify the validity of the fusion index and method.
- aeroengine /
- remaining useful life /
- stacked sparse autoencoders /
- Spearman’s rank correlation coefficient /
- similarity matching

HTML全文

图 1 RUL预测模型框架

Figure 1. Schematic diagram of RUL predicting model

下载: 全尺寸图片幻灯片

图 2 标准AE基本结构

Figure 2. Basic structure of standard AE

下载: 全尺寸图片幻灯片

图 3 SSAE建立过程

Figure 3. Process of building SSAE

下载: 全尺寸图片幻灯片

图 4 dropout建立过程

Figure 4. Process of building dropout

下载: 全尺寸图片幻灯片

图 5 样本特征提取结果

Figure 5. Results of extracting sample features

下载: 全尺寸图片幻灯片

图 6 相似性匹配结果

Figure 6. Similarity matching results

下载: 全尺寸图片幻灯片

图 7 69号发动机的T24、P30标准化结果

Figure 7. Training set No.69 Engine’s T24, P30 standardization results

下载: 全尺寸图片幻灯片

图 8 GA寻优总流程

Figure 8. GA optimization processing

下载: 全尺寸图片幻灯片

图 9 各代适应度函数变化

Figure 9. Fitness function changes from generation to generation

下载: 全尺寸图片幻灯片

图 10 特征融合指标

Figure 10. Feature synthesis indexes

下载: 全尺寸图片幻灯片

图 11 不同预测样本长度的RMSE比较

Figure 11. RMSE comparison of different predicting sample length

下载: 全尺寸图片幻灯片

图 12 不同预测样本长度的MAE比较

Figure 12. MAE comparison of different predicting sample length

下载: 全尺寸图片幻灯片

图 13 数据库容量为40、60、80、100时的预测结果

Figure 13. Prediction results when size of database is 40, 60, 80 and 100

下载: 全尺寸图片幻灯片

表 1 GA参数设定

Table 1. Parameter setting of GA

种群规模	交叉概率	变异概率	进化代数	编码长度
13	0.88	0.01	40	13

下载: 导出CSV

表 2 SSAE模型参数设定

Table 2. Parameter setting of SSAE model

参数	数值
输入层神经元个数	9
第1隐藏层神经元个数	4
第2隐藏层神经元个数	1
第3隐藏层神经元个数	4
输出层神经元个数	9
迭代轮次	100
编码器激活函数	sigmoid
解码器激活函数	purelin
稀疏参数	0.1

下载: 导出CSV

表 3 SRCC计算结果

Table 3. SRCC calculation results

传感器参数	SRCC结果
LPC出口总温T₂₄/℃	−273.525
HPC出口总温T₃₀/℃	−273.482
LPT出口总温T₅₀/℃	−273.601
HPC出口总压P₃₀/kPa	0.112
风扇物理转速N_f/(r·min⁻¹)	−0.856
核心物理转速N_c/(r·min⁻¹)	0.113
HPC出口静压P_s30	−0.125
燃油量与HPC出口静压比率C_Phi	0.833
修正风扇转速N_Rf/(r·min⁻¹)	−0.855
修正核心转速N_Rc/(r·min⁻¹)	0.551
涵道比r_BPR	−0.744
HPT冷却引气流量W₃₁/(kg·s⁻¹)	1.62
LPT冷却引气流量W₃₂/(kg·s⁻¹)	1.62

下载: 导出CSV

表 4 不同数据库大小的测试集结果

Table 4. Experimental results of different database size

数据库容量	RMSE/次	MAE/次
40	23.40	25.30
60	20.85	20.32
80	18.64	19.25
100	17.08	17.36

下载: 导出CSV

表 5 各类方法比较

Table 5. Comparison of different methods

方法	RMSE/次	S
MLP^[15]	25.23	1205
SVR^[15]	20.96	1381
DLSTM^[15]	18.33	655
CNN+LSTM^[15]	16.36	443
DeepCNN^[21]	18.45
HMM^[21]	45.11
本文	17.08	391

下载: 导出CSV

参考文献(21)

[1]	KORDESTANI M, SAIF M, ORCHARD M E, et al. Failure prognosis and applications—A survey of recent literature[J]. IEEE Transactions on Reliability, 2021, 70(2): 728-748. doi: 10.1109/TR.2019.2930195
[2]	PENG Y Z, WANG Y, ZI Y Y. Switching state-space degradation model with recursive filter/smoother for prognostics of remaining useful life[J]. IEEE Transactions on Industrial Informatics, 2019, 15(2): 822-832. doi: 10.1109/TII.2018.2810284
[3]	ELLEFSEN A L, BJØRLYKHAUG E, ÆSØY V, et al. Remaining useful life predictions for turbofan engine degradation using semi-supervised deep architecture[J]. Reliability Engineering & System Safety, 2019, 183: 240-251.
[4]	ZIO E. Prognostics and health management (PHM): Where are we and where do we (need to) go in theory and practice[J]. Reliability Engineering & System Safety, 2022, 218: 108119.
[5]	HU Y, MIAO X W, SI Y, et al. Prognostics and health management: A review from the perspectives of design, development and decision[J]. Reliability Engineering & System Safety, 2022, 217: 108063.
[6]	LI H, ZHAO W, ZHANG Y X, et al. Remaining useful life prediction using multi-scale deep convolutional neural network[J]. Applied Soft Computing, 2020, 89: 106113. doi: 10.1016/j.asoc.2020.106113
[7]	SUN W J, ZHAO R, YAN R Q, et al. Convolutional discriminative feature learning for induction motor fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2017, 13(3): 1350-1359. doi: 10.1109/TII.2017.2672988
[8]	EL-THALJI I, JANTUNEN E. A summary of fault modelling and predictive health monitoring of rolling element bearings[J]. Mechanical Systems and Signal Processing, 2015, 60-61: 252-272. doi: 10.1016/j.ymssp.2015.02.008
[9]	谭治学, 钟诗胜, 林琳. 多源数据融合的民航发动机修后性能预测[J]. 北京航空航天大学学报, 2019, 45(6): 1106-1113. doi: 10.13700/j.bh.1001-5965.2018.0557 TAN Z X, ZHONG S S, LIN L. Commercial aircraft engine post-repairing performance prediction based on fusion of multisource data[J]. Journal of Beijing University of Aeronautics and Astronautics, 2019, 45(6): 1106-1113(in Chinese). doi: 10.13700/j.bh.1001-5965.2018.0557
[10]	MUNEER A, TAIB S M, NASEER S, et al. Data-driven deep learning-based attention mechanism for remaining useful life prediction: Case study application to turbofan engine analysis[J]. Electronics, 2021, 10(20): 2453. doi: 10.3390/electronics10202453
[11]	赵洪利, 张猛. 基于随机维纳过程的航空发动机性能衰退研究[J]. 推进技术, 2021, 42(3): 488-494. doi: 10.13675/j.cnki.tjjs.190632 ZHAO H L, ZHANG M. Performance degradation of aeroengines based on stochastic Wiener process[J]. Journal of Propulsion Technology, 2021, 42(3): 488-494(in Chinese). doi: 10.13675/j.cnki.tjjs.190632
[12]	杜方舟, 孙有朝, 郭媛媛, 等. 基于数据的航空发动机排气温度裕度及剩余寿命计算方法[J]. 航空动力学报, 2020, 35(11): 2456-2464. doi: 10.13224/j.cnki.jasp.2020.11.023 DU F Z, SUN Y C, GUO Y Y, et al. Calculation method of aero-engine exhaust gas temperature margin and remaining life based on data[J]. Journal of Aerospace Power, 2020, 35(11): 2456-2464(in Chinese). doi: 10.13224/j.cnki.jasp.2020.11.023
[13]	赵洪利, 陈天铭, 郑涅. 基于特征融合多阶段相似的发动机寿命预测[J]. 系统工程与电子技术, 2021, 43(5): 1430-1436. doi: 10.12305/j.issn.1001-506X.2021.05.33 ZHAO H L, CHEN T M, ZHENG N. Engine life prediction based on multi-stage similarity of comprehensive index[J]. Systems Engineering and Electronics, 2021, 43(5): 1430-1436(in Chinese). doi: 10.12305/j.issn.1001-506X.2021.05.33
[14]	李京峰, 陈云翔, 项华春, 等. 基于LSTM-DBN的航空发动机剩余寿命预测[J]. 系统工程与电子技术, 2020, 42(7): 1637-1644. doi: 10.3969/j.issn.1001-506X.2020.07.28 LI J F, CHEN Y X, XIANG H C, et al. Remaining useful life prediction for aircraft engine based on LSTM-DBN[J]. Systems Engineering and Electronics, 2020, 42(7): 1637-1644(in Chinese). doi: 10.3969/j.issn.1001-506X.2020.07.28
[15]	李杰, 贾渊杰, 张志新, 等. 基于融合神经网络的航空发动机剩余寿命预测[J]. 推进技术, 2021, 42(8): 1725-1734. doi: 10.13675/j.cnki.tjjs.200792 LI J, JIA Y J, ZHANG Z X, et al. Remaining useful life prediction of aeroengine based on fusion neural network[J]. Journal of Propulsion Technology, 2021, 42(8): 1725-1734(in Chinese). doi: 10.13675/j.cnki.tjjs.200792
[16]	洪骥宇, 王华伟, 倪晓梅. 基于降噪自编码器的航空发动机性能退化评估[J]. 航空动力学报, 2018, 33(8): 2041-2048. doi: 10.13224/j.cnki.jasp.2018.08.028 HONG J Y, WANG H W, NI X M. Assessment of performance degradation for aero-engine based on denoising autoencoder[J]. Journal of Aerospace Power, 2018, 33(8): 2041-2048(in Chinese). doi: 10.13224/j.cnki.jasp.2018.08.028
[17]	房友龙, 贺星, 刘东风, 等. 燃气轮机健康状态组合法综合评价[J]. 推进技术, 2020, 41(8): 1903-1913. doi: 10.13675/j.cnki.tjjs.190336 FANG Y L, HE X, LIU D F, et al. Combinatorial comprehensive assessment of gas turbine health condition[J]. Journal of Propulsion Technology, 2020, 41(8): 1903-1913(in Chinese). doi: 10.13675/j.cnki.tjjs.190336
[18]	刘渊, 余映红, 田彦云, 等. 航空发动机排气温度基线建模新方法研究[J]. 推进技术, 2022, 43(4): 16-25. doi: 10.13675/j.cnki.tjjs.200511 LIU Y, YU Y H, TIAN Y Y, et al. Investigation on new method for baseline modelling of aeroengine exhaust gas temperature[J]. Journal of Propulsion Technology, 2022, 43(4): 16-25(in Chinese). doi: 10.13675/j.cnki.tjjs.200511
[19]	ZHANG Q, TSE P W T, WAN X, et al. Remaining useful life estimation for mechanical systems based on similarity of phase space trajectory[J]. Expert Systems with Applications, 2015, 42(5): 2353-2360. doi: 10.1016/j.eswa.2014.10.041
[20]	SAXENA A, GOEBEL K, SIMON D, et al. Damage propagation modeling for aircraft engine run-to-failure simulation[C]//Proceeding of the 2008 International Conference on Prognostics and Health Management. Piscataway: IEEE Press, 2008: 1-9.
[21]	彭开香, 皮彦婷, 焦瑞华, 等. 航空发动机的健康指标构建与剩余寿命预测[J]. 控制理论与应用, 2020, 37(4): 713-720. PENG K X, PI Y T, JIAO R H, et al. Health indicator construction and remaining useful life prediction for aircraft engine[J]. Control Theory & Applications, 2020, 37(4): 713-720(in Chinese).