Abstract
High-speed train bogies are essential for the safety and comfort of train operation. The performance of the bogie usually degrades before it fails, so it is necessary to detect the performance degradation of a high-speed train bogie in advance. In this paper, with two key dampers on the bogie taken as experimental objects (lateral damper and yaw damper), a novel 1D-ConvLSTM time-distributed convolutional neural network (CLTD-CNN) is proposed to estimate the performance degradation of a high-speed train bogie. The proposed CLTD-CNN is an encoder-decoder structure. Specifically, the encoder part of the proposed structure consists of a time-distributed 1D-CNN module and a 1D-ConvLSTM. The decoder part consists of a 1D-ConvLSTM and a simple time-CNN with residual connections. In addition, an auxiliary training part is introduced into the structure to support CLTD-CNN in learning the performance degradation trend characteristic, and a special input format is designed for this structure. The whole structure is end-to-end and does not require expert knowledge or engineering experience. The effectiveness of the proposed CLTD-CNN is tested by the high-speed train CRH380A under different performance states. The experimental results demonstrate the superiority of CLTD-CNN. Compared to other methods, the estimation error of CLTD-CNN is the smallest.
Highlights
With the rapid development of high-speed trains, the safety of the train operation is widely concerned [1–3]
CLTD-CNN contains three parts: an encoder part, a decoder part, and an auxiliary training part. e encoder part consists of a time-distributed 1D-CNN module [21] and a 1DConvLSTM, which encodes the input data in the order of performance states from high to low. e decoder part consists of a 1D-ConvLSTM and a time-CNN [22], which decodes the results obtained by the encoder part and outputs the estimation results
The proposed structure is well investigated by focusing on two key bogie components, the lateral damper and the yaw damper, and the effectiveness and superiority of the proposed structure are demonstrated and proved through experiments. e experimental results show that the proposed structure can be utilized to estimate unknown further degraded performance states by adopting historical data of early degradation. e experimental data adopted in the experiments come from high-speed train vibration signal datasets [14]. e mean absolute error (MAE) and root mean square error (RMSE) have been employed as metrics to evaluate the performance of the structures
Summary
With the rapid development of high-speed trains, the safety of the train operation is widely concerned [1–3]. Is paper fully considers the above issues and proposes a 1D-ConvLSTM time-distributed CNN (CLTDCNN), which is an encoder-decoder structure [20] to realize performance degradation estimation of high-speed train bogie, while proposing a new input format for this structure. With this input format, the proposed structure CLTD-CNN is able to learn the characteristics of the performance degradation trend and estimate the unknown postdegradation performance states (the performance states of test samples are not within the range of the performance states of training samples). (1) is paper proposes a novel 1D-ConvLSTM timedistributed CNN (CLTD-CNN) to achieve performance degradation estimation of a high-speed train bogie.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
