A Gas Path Fault Diagnosis Method for Aero-engine Based on TCN-LGBM Model

doi:10.12382/bgxb.2022.0615

Abstract

Abstract:

With the obvious characteristics of poor temporal logic in fault diagnosis and the strongly coupled feature parameters, the aero-engines working in the hostile gas path conditions of high temperature, pressure and strong vibration face with the degradation performance and structure defect problems such as fatigue and corrosion. And an aero-engine gas path fault diagnosis method based on temporal convolutional networks(TCN) and light gradient boosting machine(LGBM) is proposed to provide a feasible solution to the problems above. The diagnosis process can be divided into feature extraction and classification: TCN is introduced to guarantee the fault diagnosis training temporal logic and achieve the features fusion of distant layers and current layers, which is also strengthened by channel attention mechanism; the features are quickly classified based on LGBM model, and the Bayesian method is used to quickly optimize the model hyperparameters. Based on the aero-engine performance modelled by PROOSIS software, six types of fault mode are diagnosed and identified by taking a military low-bypass ratio turbofan engine as an example. The results indicate that the proposed model is effective for fault diagnosis and shows the superiority by comparing with other models.

Key words: aero-engine, fault diagnosis, temporal convolutional network, light gradient boosting machine, attention mechanism

CLC Number:

V267

LÜ Weimin, SUN Chenfeng, REN Likun, ZHAO Jie, LI Yongqiang. A Gas Path Fault Diagnosis Method for Aero-engine Based on TCN-LGBM Model[J]. Acta Armamentarii, 2024, 45(1): 253-263.

Figures/Tables 14

Fig.1 Structure graph of 1D FCN model

Fig.2 A dilated casual convolution with dilated factor D=1,2,4 and filter size k=3

Fig.3 Causal convolution process

Fig.4 SE-Net structure graph

Fig.5 Structure graph of TCN-LGBM model

Fig.6 Flow chart of fault diagnosis

Fig.7 Aero-engine performance simulation model

Table 1 Aero-engine fault mode

Table 2 TCN model and T-SNE parameters

TCN模型	参数	T-SNE	参数
Batch-size	256	init	random
Epoch	300	random_state	33
Dropout	0.1	learning_rate	200
Hidden Layers	[64,16,16]	metric	cosine

Fig.8 Aero-engine feature visualization

Table 3 Bayesian model hyperparameters

贝叶斯参数	数值	贝叶斯参数	数值
num_leaves	84	bagging_fraction	0.6
max_depth	40	bagging_freq	5
lambda_l1	0.1	min_data_in_leaf	30
lambda_l2	0.1	n_estimators	cosine

Fig.9 Classification accuracy and loss curve

Fig.10 Model confusion matrix

Fig.11 Aero-engine fault diagnosis accuracy

References 33

[1]	朱之丽, 陈敏, 唐海龙, 等. 航空燃气涡轮发动机工作原理及性能[M]. 上海: 上海交通大学出版社, 2018:257-261.
	ZHU Z L, CHEN M, TANG H L, et al. Working principle and performance of aircraft gas turbine engines[M]. Shanghai: Shanghai Jiao Tong University Press, 2018:257-261. (in Chinese)
[2]	周忠宝, 马超群, 周经伦. 基于动态贝叶斯网络的动态故障树分析[J]. 系统工程理论与实践, 2008, 28(2):35-42. doi: 10.12011/1000-6788(2008)2-35
	ZHOU Z B, MA C Q, ZHOU J L. Dynamic fault tree analysis based on dynamic bayesiannetworks[J]. System Engineering-Theory & Practice, 2008, 8(2):35-42. (in Chinese)
[3]	蔡鼎. 融合先验知识的燃气轮机欠定气路故障诊断方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2105: 40-42.
	CAI D. Method forunderdetermined fault diagnosis with the prior knowledge of gas turbine[D]. Harbin: Harbin Institute of Technology, 2015:40-42. (in Chinese)
[4]	金山. 复杂系统可靠性建模与分析[M]. 北京: 国防工业出版社, 2015:5-17.
	JIN S. Complexsystem reliability modeling and analysis[M]. Beijing: National Defense Industry Press, 2015:5-17. (in Chinese)
[5]	杨洁, 万平安, 王景霖, 等. 基于多传感器融合卷积神经网络的航空发动机轴承故障诊断[J]. 中国电机工程学报, 2022, 42(13):4933-4941.
	YANG J, WAN P A, WANG J L, et al. Aeroengine bearing fault diagnosis based on convolutional neural network for multi-sensor information fusion[J]. Proceedings of the CSEE, 2022, 42(13): 4933-4941. (in Chinese)
[6]	魏晓良, 潮群, 陶建峰, 等. 基于LSTM和CNN的高速柱塞泵故障诊断[J]. 航空学报, 2021, 42(3):435-445.
	WEI X L, CHAO Q, TAO J F, et al. Cavitation fault diagnosis method for high-speed plunger pumps based on LSTM and CNN[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(3): 435-445. (in Chinese)
[7]	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]// Proceedings of the 15th European Conference on Computer Vision.Munich, Germany: Springer, 2018: 3-19.
[8]	ZHAO M H, ZHONG S S, FU X Y, et al. Deep residual shrinkage networks for fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2020, 16(7):4681-4690. doi: 10.1109/TII.9424 URL
[9]	DING Y, MA L, MA J, et al. Intelligent fault diagnosis for rotating machinery using deep Q-network based health state classification[J]. Advanced Engineering Informatics, 2019, 42: 100977. doi: 10.1016/j.aei.2019.100977 URL
[10]	邓飞跃, 丁浩, 吕浩洋, 等. 一种基于轻量级神经网络的高铁轮对轴承故障诊断方法[J]. 工程科学学报, 2021, 43(11): 1482-1490.
	DENG F Y, DING H, LÜ H Y, et al. Fault diagnosis of high-speed train wheelset bearing based on a lightweight neural network[J]. Chinese Journal of Engineering, 2021, 43(11): 1482-1490. (in Chinese)
[11]	潘宏侠, 张玉学. 基于SST时频图纹理特征的供输弹系统故障诊断[J]. 振动与冲击, 2020, 39(6): 132-137.
	PAN H X, ZHANG Y X. Fault diagnosis of the ammunition supply system based on the texture features of SST time-frequency distribution image[J]. Journal of Vibration and Shock, 2020, 39(6): 132-137. (in Chinese)
[12]	SHOKROLAHI S M, KARIMIZIARANI M. A deep network solution for intelligent fault detection in analog circuit[J]. Analog Integrated Circuits and Signal Processing, 2021, 107(3):597-604. doi: 10.1007/s10470-020-01732-8
[13]	HOWARD A G, ZHU M L, CHEN B, et al. Mobilenets: Efficientconvolutional neural networks for mobile vision Applications:arXiv:1704.04861[R/OL]. Ithaca, NY, US: Cornell University.(2017-04-17) [2020-12-09]. https://arxiv.org/abs/1704.04861.
[14]	BAI S J, ZICO K, VLADLEN K. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling:arXiv:1803.01271[R/OL]. Ithaca, NY, US: Cornell University.(2018-03-04). https://arxiv.org/abs/1803.01271.
[15]	皮骏, 黄江博. 基于IPSO-Elman神经网络的航空发动机故障诊断[J]. 航空动力学报, 2017, 32(12):3031-3038.
	PI J, HUANG J B. Aero-engine fault diagnosis based on IPSO-Elman neural network[J]. Journal of Aerospace Power, 2017, 32(12):3031-3038. (in Chinese)
[16]	谭海旺, 杨启亮, 邢建春, 等. 基于XGBoost-LSTM组合模型的光伏发电功率预测[J]. 太阳能学报, 2022, 43(8):75-81. doi: 10.19912/j.0254-0096.tynxb.2021-0005
	TAN H W, YANG Q L, XING J C, et al. Photovoltaic output prediction based on combined XGBoost-LSTM model[J]. Acta Energiae Solaris Sinica, 2022, 43(8):75-81. (in Chinese)
[17]	马怀祥, 冯旭威, 李东升, 等. 基于CNN和XGBoost的滚动轴承故障诊断方法[J]. 中国工程机械学报, 2019, 19(3):254-259.
	MA H X, FENG X W, LI D S, et al. Fault diagnosis method of rolling bearing based on CNN and XGBoost[J]. Chinese Journal of Construction Machinery, 2019, 19(3):254-259. (in Chinese)
[18]	AARON V D O, SANDER D, HEIGA Z. Wavenet:a generative model for raw audio:arXiv :1609.03499[R/OL]. Ithaca,NY, US: Cornell University, 2016:1609.03499.(2016-09-19). https://arxiv.org/abs/1609.03499.
[19]	HE Y D, ZHAO J B. Temporal convolutional networks for anomaly detection in time series[J]. Journal of Physics: Conference Series, 2013, 1213(4):042050. doi: 10.1088/1742-6596/1213/4/042050
[20]	HU J, SHEN L, SUN G. Squeeze-and-eitation networks[C]// Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, US:IEEE, 2019:7132-7141.
[21]	华校专. AI算法工程师手册[R/OL].(2022-04-03)[2022-04-08]. http://www.huaxiaozhuan.com/.
	HUA X Z. AI algorithms engineer handbook[R/OL]. (2022-04-03)[2022-04-08]. http://www.huaxiaozhuan.com/. (in Chinese)
[22]	杜先君, 巩彬, 余萍, 等. 基于CBAM-CNN的模拟电路故障诊断[J]. 控制与决策, 2022, 37(10):2609-2618.
	DU X J, GONG B, YU P, et al. Research on CBAM-CNN based analog circuit fault diagnosis[J]. Control and Decision, 2022, 37(10):2609-2618. (in Chinese)
[23]	田彦云. 涡轴发动机建模与扭矩加载控制技术研究[D]. 南京: 南京航空航天大学, 2016.
	TIAN Y Y. Research on modeling and torque loading control technology of turbo-shaft[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2016. (in Chinese)
[24]	张浩, 李密, 赵海刚. 基于试飞数据的涡轴发动机性能建模技术[J]. 推进技术, 2021, 42(10):2177-2186.
	ZHANG H, LI M, ZHAO H G. Performance modeling technique of turboshaft engine based on flight experimental data[J]. Journal of Propulsion Technology, 2021, 42(10): 2177-2186. (in Chinese)
[25]	潘慕绚, 陈强龙, 周永权, 等. 涡扇发动机多动力学建模方法[J]. 航空学报, 2019, 40(5):99-110.
	PAN M X, CHEN Q L, ZHOU Y Q, et al. Amulti-dynamicsapproachto turbofan engine modeling[J]. Acta Aeronautica et Astronautica Sinica, 2019, 40(5): 99-110. (in Chinese)
[26]	LU F, JU H F, HUANG J Q. An improved extended Kalman filter with inequality constraints for gas turbine engine health monitoring[J]. Aerospace Science and Technology, 2016, 58(1):36-47. doi: 10.1016/j.ast.2016.08.008 URL
[27]	楚娜娜, 张曙光, 高艳蕾, 等. 基于Simscape模型的航空发动机系统安全性分析方法[J]. 航空动力学报, 2021, 36(4): 885-896.
	CHU N N, ZHANG S G, GAO Y L, et al. Safety analysis method of aero-engine systems based on simscape model[J]. Journal of Aerospace Power, 2021, 36(4): 885-896. (in Chinese)
[28]	刘星怡. 基于深度信念网络的航空发动机气路故障诊断方法研究[D]. 南京: 南京航空航天大学, 2020.
	LIU X Y. Research ongas path fault diagnosis method for aeroengine based on deep belief network[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2020. (in Chinese)
[29]	葛怡. 航空发动机几种典型故障建模仿真及预测[D]. 西安: 西安理工大学, 2020.
	GE Y. Modeling, simulation and prediction of several typical faults of aeroengine[D]. Xi’an: Xi’an University of Technology, 2020. (in Chinese)
[30]	WEI T T, VAN BEEK A, HAO J R, et al. Bayesian calibration of performance degradation in a gas turbine-driven compressor unit for prognosis health management[J]. Journal of Engineering for Gas Turbines and Power, 2022, 144(5):51014. doi: 10.1115/1.4053564 URL
[31]	ZHAO Y P, CHEN Y B. Extreme learning machine based transfer learning for aero engine fault diagnosis[J]. Aerospace Scienece and Technology, 2022, 121:107311.
[32]	石重托, 姚伟, 黄彦浩, 等. 基于SE-CNN和仿真数据电力系统主导失稳模式智能识别[J/OL]. 中国电机工程学报.(2021-11-08). https://kns.cnki.net/kcms/detail/11.2107.TM.2021110-5.1639.046.html.
	SHI Z T, YAO W, HUANG Y H, et al. Power system dominant instability model identification based on convolutional neural networks with squeeze and excitation block and simulation data[J]. Proceedings of the CSEE. (2021-11-08). https://kns.cnki.net/kcms/detail/11.2107.TM.20211105.1639.046.html. (in Chinese)
[33]	张浩, 胡昌华, 杜党波, 等. 多状态影响下基于Bi-LSTM网络的锂电池剩余寿命预测方法[J]. 电子学报, 2022, 50(3):619-624. doi: 10.12263/DZXB.20210207
	ZHANG H, HU C H, DU D B, et al. Remaining useful life prediction method of lithium-Ion battery based on Bi-LSTM network under multi-state influence[J]. Acta Electronica Sinica, 2022, 50(3):619-624. (in Chinese) doi: 10.12263/DZXB.20210207