A non-intrusive real-time fault monitoring method for train bogies

Abstract

The invention discloses a non-intrusive real-time fault monitoring method for a train bogie, comprising: obtaining pre-collected vibration data M corresponding to each key component and pre-collected vibration data C corresponding to a total signal measurement point; extracting a signal of the bogie vibration data feature, establish a bogie vibration signal feature library; obtain a multi-label fault identification model S(F C ); collect the real-time vibration data CR at the total signal measuring point, and extract the signal feature F in the CR CR ; Execute the multi-objective optimization algorithm to obtain the non-dominated solution set NS; obtain the N groups of independent variables with the smallest corresponding optimization target values ​​in NS as the initial results of the N groups of fault identification; use the trained model S(F C ) identifies the collected signal CR, outputs the multi-label classification and identification results, and takes the group of fault identification preliminary results with the largest intersection with the multi-label classification and identification results among the N groups of preliminary fault identification results as the final fault identification result. The invention provides a non-intrusive real-time fault monitoring method for a train bogie, which does not require a large number of sensors and has high precision and stability.

Description

technical field [0001] The invention belongs to the field of train component identification and fault monitoring, in particular to a non-invasive real-time fault monitoring method for a train bogie. Background technique [0002] In recent years, my country's high-speed trains have been developing continuously. As the core component of high-speed trains, the stability of train bogies is related to the operation safety of high-speed trains. The component detection and fault identification of bogies have also received extensive attention. [0003] There are two main methods of bogie fault monitoring: real-time monitoring and parking maintenance. The frequency of parking and maintenance is low, and the real-time stability and safety of train operation cannot be guaranteed. Therefore, real-time monitoring is generally used to monitor bogie faults. [0004] The existing real-time fault monitoring of bogies is mainly realized by intrusive installation of a large number of sensors....

AI Technical Summary

Problems solved by technology

[0005] All of the above methods can realize the fault monitoring of key components of the bogie, but they are all intrusive methods, requiring the installation of a large number of sensors to achieve real-time measurement at the same time, with high cost and large redundancy, and the arrangement of sensors will affect the Structural optimization and lightweight design of bogie

Images