The Research Of Key Technology In Vehicle Security And Early Warning

Safety is a permanent theme of railway transportation. The major train accidents are crack in plate hole in plate, Cold-Cut of relief notch, thermal breakage of bearing. The thermal breakage of bearing is the most representative accident in the train accidents, the advanced "5T" intelligent warning system is used to monitor the train states, which including Trace hotbox detection system, truck performance detection system, trackside acoustic detection system, trouble of moving freight car detection system, train coach running diagnosis system.With the with intelligence, information and network technology, the5T system can establish a railway truck security system which have a direct and effective preventive measure, a network equipment layout, a global data share channel, and a full coverage monitoring system. the5t system achieve the dynamic monitoring, remote diagnosis and networking operation of the running train by ground equipment, which realize the historical transformation of railway security system change from man-controlled to computerized, from tradition to modernization, and from extensive management to intensive management. Meanwhile, the system provides a large number of data information for railway truck manufacturing and maintenance, which improve the overall quality and repair efficiency of rail wagon.This paper analyzed the railway security situation and discussed the existing problems in the current management. The key technology of5T system and train safety was researched. In order to understood principles and functions of the5T system and improve the early warning and management system, an5t project was built in Kunming railway administration, which give full play to role of safety and transport in the high-speed railway development process.Second, this paper achieved the bearing condition monitoring based on the key technology of5T system. a bearing monitoring test bed was built, which can simulate the bearing wear condition of train. The monitoring signals in different conditions can be got by the sensor and data acquisition system. The bearing states feature information can be obtained by the feature extraction to time domain, frequency domain and time-frequency domain. The bearing states can be identified by the pattern recognition to the obtained features. The feature information in time domain includes peak, variance, standard deviation, and kurtosis. After the FFT transform to time domain signal, the feature information in frequency domain can be got, include gravity frequency, mean-square frequency, root mean square frequency, variance frequency and standard deviation frequency. Multi-resolution analysis for signal time-frequency decomposition is effective, but due to the transformation is according to the binary conversion, so in high frequency band the frequency resolution is poorer, and in low frequency band the time resolution is poorer. Faceing the actual engineering problems, in order to more elaborate decomposition signal, that the high frequency part of signal has been further wavelet decomposition is necessary. So in the time domain and frequency domain feature extraction process, this paper adopted the wavelet packet analysis method which are widely used in the process of signal analysis now days, at the same time got the feature of different time-frequency space. Last through using artificial neural network (based on the BP neural network) for different state of pattern recognition, and effectively identify the running state of the bearing.Finally, with the aid of experimental data, the experimental results and the research of experience, and combinating with the operation process of "5t" prevention, early warning system, all of this could result that better use of "5t" prevention, early warning system, and let the railway operation become more safety.