Abstract:Railway turnout systems (RTS), as one of the most critical ground infrastructures of the high-speed rail, which directly impact the railway operation safety, are also susceptible to frequent faults. Hence, efficient fault prediction and intelligent maintenance decision support of RTS are significant to avoid unexpected outages and guarantee the operation safety of trains. A key challenge for fault prediction of RTS is to extract and select effective feature combinations that describe the degradation status of the system and then generate health indicators (HI) with high monotonicity, correlation, and robustness. In this paper, we focus on the jam fault caused by insufficient lubrication, which is one of the most frequent faults of RTS, and we present a novel unsupervised adaptive latent feature extraction method based on the improved sparse auto-encoder (ISAE). The locally weighted regression (LOESS) is performed to smooth the extracted latent features to overcome the influence of the noise. Improved auto-associative kernel regression (IAAKR) is utilized to fuse the feature combination and generate HI, which can intuitively describe the degradation status of RTS. Finally, the gated recurrent unit (GRU) network is introduced to enhance the correlation data process and construct the fault prediction models. The parameters of GRU are optimized by the improved particle swarm optimization (IPSO) algorithm. Extensive experiments are conducted on a real-world RTS degradation dataset collected from the Changsha Railway Station in China, which is related to the jam fault caused by the lubrication reduction, and the results show the proposed method achieves better prediction performance compared to state-of-the-art methods in terms of different evaluation metrics.

Fault Prediction Method of Railway Braking System Based on LSTM Neural Network

Research on Fault Prediction of High-Speed Train Auxiliary Power Supply System Based on LSTM

Data-Based Line Trip Fault Prediction in Power Systems Using LSTM Networks and SVM.

Spatio-Temporal Data Prediction of Braking System Based on Residual Error

A Deep-LSTM-Based Fault Detection Method for Railway Vehicle Suspensions

Multi-sensor Data Mining of Long-grouped Train Braking System Based on an LSTM Network

Real-Time Crash Risk Prediction using Long Short-Term Memory Recurrent Neural Network

Prediction of Feature Points in Train Braking Curve Based on EC-LSTM Multiple Regression Network

Accuracy and Efficiency Enhancement for Air Pressure of Train Braking Based on Multidimensional LSTM-Attention Model

Traffic Accident Prediction Based on LSTM Neural Network Model

Fault Prediction of Railway Turnout Systems Based on Improved Sparse Auto Encoder and Gated Recurrent Unit Network

Research on the Fault Prediction Method of an On-Board Subsystem in High-Speed Train Control Systems

Sparse Temporal Data-Driven SSA-CNN-LSTM-Based Fault Prediction of Electromechanical Equipment in Rail Transit Stations

Fault Prediction Method for On-Board Equipment of CTCS Based on Seq2Seq Model Using LSTM

Seismic Response Prediction of a Train-Bridge Coupled System Based on a LSTM Neural Network

Safety Prediction Of Rail Transit System Based On Deep Learning

Short-Term Interval Prediction of Inbound Passenger Flow of Subway Station under Failure Events

A Modified Bidirectional Long Short-Term Memory Neural Network for Rail Vehicle Suspension Fault Detection

Multi-sensor Data Mining of Single Train Braking System Based on Long Short-term Memory Network

Safety analysis of high-speed trains on bridges under earthquakes using a LSTM-RNN-based surrogate model

Research on Prediction of Time Between Failures for Onboard Subsystem of Train Control System.