Aiming at the problems of low prediction accuracy of remaining useful life and difficulty in extracting degenerative features, a remaining useful life prediction method based on spatio-temporal feature fusion is proposed. Firstly, a dual-channel deep learning network is constructed, the long short-term memory neural network based on channel attention mechanism is used to extract the degraded information in the time dimension in channel 1, and the semantic segmentation network is used to mine the degraded information in the spatial dimension in channel 2. Secondly, the cross attention mechanism is used to fuse the spatiotemporal features. Finally, the regression layer is used to realize the remaining useful life prediction. Experimental verification on the simulation dataset of commercial modular aviation propulsion system shows that compared with other deep learning methods, the value of root mean square error and score in the FD004 test set are reduced by 16.21% and 14.97%, and the remaining useful life is predicted with high accuracy.
对此,提出一种基于时空特征融合的剩余寿命预测方法。首先,将通道注意力机制(Squeeze and Excitation, SE)与LSTM网络相结合,提取不同输入数据的时间维度退化特征,同时采用SegNet提取空间维度特征;其次,通过交叉注意力(Cross Attention)机制实现时空特征融合;最后,在商用模块化航空推进系统仿真数据集(Commercial Modular Aero-Propulsion System Simulation, C-MAPSS)验证模型的有效性。
XUD, XIAOX Q, LIUJ, et al. Spatio-temporal degradation modeling and remaining useful life prediction under multiple operating conditions based on attention mechanism and deep learning[J]. Reliability Engineering & System Safety, 2023,229:No.108886.
[2]
XUD Y, QIUH B, GAOL, et al. A novel dual-stream self-attention neural network for remaining useful life estimation of mechanical systems[J]. Reliability Engineering & System Safety, 2022,222:No.108444.
YANGD, CHHATREN, CAMPIF, et al. Feasibility of support vector machine gesture classification on a wearable embedded device[C]∥Proceedings of the 2016 IEEE Canadian Conference on Electrical and Computer Engineering. Piscataway, USA: IEEE, 2016. DOI: 10.1109/CCECE.2016.7726800 .
LIX, DINGQ, SUNJ Q. Remaining useful life estimation in prognostics using deep convolution neural networks[J]. Reliability Engineering & System Safety, 2018,172:1-11.
SHANGZ W, ZHANGB R, LIW X, et al. Machine remaining life prediction based on multi-layer self-attention and temporal convolution network[J]. Complex & Intelligent Systems, 2022,8(2):1409-1424.
[10]
CHENZ H, WUM, ZHAOR, et al. Machine remaining useful life prediction via an attention-based deep learning approach[J]. IEEE Transactions on Industrial Electronics, 2021,68(3):2521-2531.
FREDERICKD K, DECASTROJ A, LITTJ S. Users guide for the commercial modular aero-propulsion system simulation (C-MAPSS)[R/OL]. (2007-10-01)[2024-11-29].
[17]
LIUL, SONGX, ZHOUZ T. Aircraft engine remaining useful life estimation via a double attention-based data-driven architecture[J]. Reliability Engineering & System Safety, 2022,221:108330.
[18]
YOUK S, QIUG Q, GUY K. A 3-D attention-enhanced hybrid neural network for turbofan engine remaining life prediction using CNN and BiLSTM models[J]. IEEE Sensors Journal, 2024,24(14):21893-21905.