GA-BP maglev train fault detection method based on threshold determination

Abstract

The invention discloses a GA-BP maglev train fault detection method based on threshold determination. Characteristic parameters of data obtained by a sensor are normalized and then used as BP input, avariance threshold and a change rate determination fault are used as output, and an optimal network weight and a threshold are optimized by adopting a genetic algorithm, so that the detection precision of the BP neural network is improved. It can be accurately diagnosed whether a maglev train has a vibration fault or not, and the problem of false alarm is avoided.

Description

technical field

[0001] The invention relates to rail transit fault diagnosis technology, in particular to a GA-BP maglev train fault detection method based on threshold value judgment. Background technique

[0002] my country's railways have developed rapidly over the years. By the end of 2017, the operating mileage of railways had reached 127,000 kilometers, including 25,000 kilometers of high-speed rail, ranking first in the world. Magnetic levitation has also ushered in rapid development since the 21st century. It satisfies the diverse travel modes of the people. With the improvement of people's living standards, the comfort of the riding experience is also constantly improving.

[0003] The safety and reliability of vehicle operation and ride comfort are all very important indicators of ride. Fault diagnosis technology can improve the safety and reliability of train operation and ride comfort by judging the abnormal state caused by the vibration of the train in operat...

Examples

Embodiment Construction

[0083] The present invention is further analyzed below in conjunction with specific accompanying drawings.

[0084] Such as figure 1 As shown, the GA-BP fault detection method for maglev trains based on threshold judgment, the steps are as follows:

[0085] Step 1. Obtain the required data by installing multiple sets of acceleration, current, and gap sensors on the train, and filter and preprocess the original data. Screen 40,000+ data, select 18,000+ data, and use the five-point average method to process the data for missing detection data and over-detection data for the selected data. Then slice (t time period) for the processed data.

[0086] Step 2, using signal processing technology and statistical learning method to extract the characteristic parameters within the time period t for each type of data processed in step 1. The characteristic parameters include time-domain indicators, frequency-domain indicators and time-frequency features.

[0087] 2.1 Time Domain Index...

technical field

[0001] The invention relates to rail transit fault diagnosis technology, in particular to a GA-BP maglev train fault detection method based on threshold value judgment. Background technique

[0002] my country's railways have developed rapidly over the years. By the end of 2017, the operating mileage of railways had reached 127,000 kilometers, including 25,000 kilometers of high-speed rail, ranking first in the world. Magnetic levitation has also ushered in rapid development since the 21st century. It satisfies the diverse travel modes of the people. With the improvement of people's living standards, the comfort of the riding experience is also constantly improving.

[0003] The safety and reliability of vehicle operation and ride comfort are all very important indicators of ride. Fault diagnosis technology can improve the safety and reliability of train operation and ride comfort by judging the abnormal state caused by the vibration of the train in operat...

Examples

Embodiment Construction

[0083] The present invention is further analyzed below in conjunction with specific accompanying drawings.

[0084] Such as figure 1 As shown, the GA-BP fault detection method for maglev trains based on threshold judgment, the steps are as follows:

[0085] Step 1. Obtain the required data by installing multiple sets of acceleration, current, and gap sensors on the train, and filter and preprocess the original data. Screen 40,000+ data, select 18,000+ data, and use the five-point average method to process the data for missing detection data and over-detection data for the selected data. Then slice (t time period) for the processed data.

[0086] Step 2, using signal processing technology and statistical learning method to extract the characteristic parameters within the time period t for each type of data processed in step 1. The characteristic parameters include time-domain indicators, frequency-domain indicators and time-frequency features.

[0087] 2.1 Time Domain Index...

Images