The robust guarantee of train control on-board equipment is inextricably linked to the safe functioning of a high-speed train. A fault diagnostic model of on-board equipment is built utilizing the integrated learning XGBoost (eXtreme Gradient Boosting) algorithm to help technicians assess the malfunction category of high-speed train control on-board equipment accurately and rapidly. XGBoost algorithm iterates multiple decision tree models to improve the accuracy of fault diagnosis by lifting the predicted residual and adding regular terms. To begin, the text features were extracted using the improved TF-IDF (Term Frequency–Inverse Document Frequency) approach, and 24 fault feature words were chosen and converted into weight word vectors. Secondly, considering the imbalanced fault categories in the data set, ADASYN (Adaptive Synthetic sampling) adaptive synthetically oversampling technique was used to synthesize a few category fault samples. Finally, the data samples were split into training and test sets based on the fault text data of CTCS-3 train control on-board equipment recorded by Guangzhou Railway Group maintenance personnel. The XGBoost model was utilized to realize the automatic fault location of the test set after optimized parameter tuning through grid search. Compared with other methods, the evaluation index of the XGBoost model was significantly improved. The diagnostic accuracy reached 95.43%, which verifies the effectiveness of the method in text fault diagnosis.

To assess the operational safety risk of long-term evolution for metro (LTE-M) communication system more accurately, guide maintenance strategy, the improved evidence theory and multi-attribute ideal reality comparative analysis (MAIRCA) approaches are proposed respectively. According to the features of the LTE-M system, the risk evaluation system is established. The enhanced structural entropy weight method is used to count the weight. Furthermore, combined with nine-element fuzzy mathematics to transform the degree of membership. Modifying the conflict and fusion rules to solve the confidence degree clashed problem of evidence theory. Then get the system risk grade assessment result. For the purpose of forming the ranking of indicator importance, the MAIRCA is introduced and the sort is based on three-dimensional. The operational state of the metro line is used as the data source in various ways based on the test and calculation. The results show that the method is effective, compared with the others, the confidence degree in the obtained risk grade increased by 7.12%. It is verified that MAIRCA can be applied to the field of urban rail transit because excellent stability and the ranking result of risk factors is reasonable. The influencing indicator with the highest importance is’ the equipment failure rate’.

Debris flow causes huge casualties and economic losses to railway construction and transportation every year, so it is of great significance to analyze the severity of debris flow to reduce the loss. For the debris flow hazards severity analysis, an integrated approach based on G1-ANP was proposed. Firstly, under the condition of the environmental risk, induced conditional risk, and vulnerability risk, the 16 index factors have been selected, for example, the degree of slope, land use type, flow accumulation, and annual mean rainfall. Then, considering the interaction among risk factors, a multi-level G1-ANP risk factor structure model has been established based on the criteria of buried capacity, impact velocity, and scouring distance (or area) of debris flow and the solution process of the model was described. Finally, the risk severity and the proportion of the various risks for each section were calculated using the weighted method. The analysis results are shown in the improved radar chart. The results show that the overall severity of debris flow in the Chengkang railway is not very high. The results can provide a reference for the management of debris flow hazards prevention and reduce the losses caused by hazards in railway operation.

Aiming at the current problems of high failure rate and low diagnostic efficiency of Railway Point Machines (RPMs) in railway industry, a short-time method of fault diagnosis is proposed. Considering the effect of noise on power signals in the data acquisition process of railway Centralized Signaling Monitoring (CSM) System, this study utilizes wavelet threshold denoising to eliminate the interference of it. The consequences show that the accuracy of fault diagnosis can be improved by 4.4% after denoising the power signals. Then in order to attain lightweight and shorten running time of diagnosis model, Mallat wavelet decomposition and artificial immune algorithm are applied to RPMs fault diagnosis. Finally, voluminous experiments using veritable power signals collected from CSM are introduced, which manifest that combining these methods can procure higher precision of RPMs and curtail fault diagnosis time. It substantiates the validity and feasibility of the presented approach.

No abstract