基于无监督迁移学习的电梯制动器剩余寿命预测

doi:10.16183/j.cnki.jsjtu.2020.175

摘要/Abstract

摘要：

为了改善电梯制动器在真实工作环境下的寿命预测效果,提出一种基于长短期记忆网络自编码器(LSTM-ED)的无监督深度迁移学习方法,利用仿真数据实现制动器在工作时的健康状态分析.利用源领域数据初步训练LSTM-ED和全连接网络;以LSTM-ED为特征提取器,将仿真和实际数据映射到特征空间并利用最大平均差异实现数据对齐;利用全连接网络回归特征空间中的目标领域数据,从而实现对制动器在真实工作环境下的剩余生命周期预测.在训练阶段中,采用分步训练法替代传统的联合训练法,以保证单个模块的准确性.对比试验仿真数据与电梯塔中的实际工作数据,以验证方法的有效性.结果表明:通过引入迁移学习和分步训练法,所提方法可以将剩余生命周期预测的均方误差降低至 0.0016,能够实现电梯制动器在真实工作环境下的剩余生命周期精准预测.

关键词: 电梯制动器, 无监督深度迁移学习, 长短期记忆网络自编码器, 剩余生命周期, 分步训练

Abstract:

In order to improve the life prediction effect of elevator brake in the real working environment, an unsupervised deep transfer learning (UDTL) method based on long short-term memory encoder-decoder (LSTM-ED) was proposed. The simulation data were used to analyze the health status of brake when it was working. First, the LSTM-ED and the fully connected network were initially trained through the source domain data. Then, the LSTM-ED was used as a feature extractor to map the simulated and actual data to the feature space, and the maximum mean discrepancy was adopted to achieve data alignment. Finally, the target domain data in the feature space was regressed through the fully connected network to predict the remaining useful life (RUL) of the real brake. In the training phase, a step-by-step training method was used to ensure the accuracy of a single module. The validity was verified by comparing the experimental simulation data with the real working data in the elevator tower. The results show that by introducing the transfer learning and step-by-step training methods, the proposed method can reduce the mean square error of RUL prediction to 0.0016, and can achieve accurate RUL prediction of elevator brakes in real working environment.

Key words: elevator brake, unsupervised deep transfer learning (UDTL), long short-term memory encoder-decoder (LSTM-ED), remaining useful life (RUL), step training

中图分类号:

TP242

姜宇迪, 胡晖, 殷跃红. 基于无监督迁移学习的电梯制动器剩余寿命预测[J]. 上海交通大学学报, 2021, 55(11): 1408-1416.

JIANG Yudi, HU Hui, YIN Yuehong. Unsupervised Transfer Learning for Remaining Useful Life Prediction of Elevator Brake[J]. Journal of Shanghai Jiao Tong University, 2021, 55(11): 1408-1416.

图/表 13

图1

图2

图3

图4

图5

图6

图7

图8

图9

表1

表2

图10

表3

参考文献 16

[1]	WOLSZCZAK P, LONKWIC P, CUNHA A, et al. Robust optimization and uncertainty quantification in the nonlinear mechanics of an elevator brake system[J]. Meccanica, 2019, 54(7):1057-1069. doi: 10.1007/s11012-019-00992-7 URL
[2]	樊朝锺. 电梯曳引机电磁制动系统故障检测及系统测试[J]. 中国设备工程, 2018(23):104-106.
	FAN Chaozhong. Fault detection and system test of electromagnetic braking system of elevator traction machine[J]. China Plant Engineering, 2018(23):104-106.
[3]	赵海文, 吴云龙, 贺鹏, 等. 电梯曳引机制动器故障检测方法研究[J]. 机床与液压, 2018, 46(1):185-188.
	ZHAO Haiwen, WU Yunlong, HE Peng, et al. Research for detection method of elevator tractor brake fault[J]. Machine Tool & Hydraulics, 2018, 46(1):185-188.
[4]	周前飞, 丁树庆, 冯月贵, 等. 基于支持向量机的电梯制动器智能监测预警系统[J]. 中国特种设备安全, 2018, 34(5):22-27.
	ZHOU Qianfei, DING Shuqing, FENG Yuegui, et al. The elevator brake intelligent monitoring and fault early warning system based on SVM[J]. China Special Equipment Safety, 2018, 34(5):22-27.
[5]	贺无名, 王培良, 沈万昌. 基于LS-SVM的电梯制动器故障诊断[J]. 工矿自动化, 2010, 36(2):44-48.
	HE Wuming, WANG Peiliang, SHEN Wanchang. Fault diagnosis of elevator brake based on LS-SVM[J]. Industry and Mine Automation, 2010, 36(2):44-48.
[6]	RAMASSO E. Investigating computational geometry for failure prognostics[J]. International Journal of Prognostics and Health Management, 2014, 5(1):1-18.
[7]	SI X S, WANG W B, HU C H, et al. Remaining useful life estimation—A review on the statistical data driven approaches[J]. European Journal of Operational Research, 2011, 213(1):1-14. doi: 10.1016/j.ejor.2010.11.018 URL
[8]	TAN C Q, SUN F C, KONG T, et al. A survey on deep transfer learning[M]// Artificial Neural Networks and Machine Learning-ICANN 2018. Cham: Springer International Publishing, 2018: 270-279.
[9]	ZHAO Z B, ZHANG Q Y, YU X L, et al. Unsupervised deep transfer learning for intelligent fault diagnosis: An open source and comparative study[EB/OL]. (2019-12-28)[2020-06-09]. https://arxiv.org/abs/1912.12528.
[10]	YANG B, LEI Y G, JIA F, et al. An intelligent fault diagnosis approach based on transfer learning from laboratory bearings to locomotive bearings[J]. Mechanical Systems and Signal Processing, 2019, 122:692-706. doi: 10.1016/j.ymssp.2018.12.051 URL
[11]	AHN E, KUMAR A, FENG D G, et al. Unsupervised deep transfer feature learning for medical image classification[C]// 2019 IEEE 16th International Symposium on Biomedical Imaging. Venice, Italy: IEEE, 2019: 1915-1918.
[12]	TAHMORESNEZHAD J, HASHEMI S. Visual domain adaptation via transfer feature learning[J]. Knowledge and Information Systems, 2017, 50(2):585-605. doi: 10.1007/s10115-016-0944-x URL
[13]	宋鹏, 郑文明, 赵力. 基于特征迁移学习方法的跨库语音情感识别[J]. 清华大学学报(自然科学版), 2016, 56(11):1179-1183.
	SONG Peng, ZHENG Wenming, ZHAO Li. Cross-corpus speech emotion recognition based on a feature transfer learning method[J]. Journal of Tsinghua University (Science and Technology), 2016, 56(11):1179-1183.
[14]	SUN C, MA M, ZHAO Z B, et al. Deep transfer learning based on sparse autoencoder for remaining useful life prediction of tool in manufacturing[J]. IEEE Transactions on Industrial Informatics, 2019, 15(4):2416-2425. doi: 10.1109/TII.9424 URL
[15]	JIA F, LEI Y G, GUO L, et al. A neural network constructed by deep learning technique and its application to intelligent fault diagnosis of machines[J]. Neurocomputing, 2018, 272:619-628. doi: 10.1016/j.neucom.2017.07.032 URL
[16]	ZHANG B, LI W, TONG Z, et al. Bearing fault diagnosis under varying working condition based on domain adaptation[EB/OL]. (2017-07-31)[2020-06-09]. https://arxiv.org/abs/1707.09890.

方法	MAE	MSE	RMSE
无迁移学习	0.053	0.0039	0.063
传统训练法	0.045	0.0035	0.059
UDTL(本文)	0.030	0.0016	0.040

方法	MSE		迁移损失
方法	源领域	目标领域	迁移损失
卷积神经网络	0.0015	0.0084	0.0069
快速傅里叶法	0.0013	0.0104	0.0091
稀疏自编码器+KL散度	0.0012	0.0034	0.0022
UDTL(本文方法)	0.0010	0.0016	0.0006

名称	r
名称	制动器1	制动器2	制动器3
制动器1	1	0.975	0.997
制动器2	-	1	0.993
制动器3	-	-	1