Partial Subsidence Detection And Recognition Technology For High Speed Railway

Track regularity is an important factor to ensure the comfort and safety of the train running, and high speed railways must have good track regularity. China’s high speed railways are generally faced with the partial subsidence. The partial subsidence will inevitably lead to the track deformation, which will bring a serious security risks. Therefore, it is necessary and urgent to detect and prevent the occurrence of partial subsidence of high speed railways. However, the existing methods of subsidence measurement are difficult to effectively implement under the current environment of the high speed railways. In view of the above problems, this paper proposed a dynamic subsidence detection method based on high speed comprehensive inspection train, and related studies were carried out.In this paper, firstly, the subsidence detection methods both at home and abroad have been studied, the advantages and disadvantages were analyzed, as well as the applicability of the method. Then, from the point of view of the application scenario, system output and system modeling, the sensor installation mode was selected. According to the simulation analysis, the designed precision index of the sensor was determined. On the basis of above, the selection of the sensor was completed, and the system hardware was constructed based on the high speed comprehensive inspection train platform. Taking into account the poor electromagnetic compatibility of the high speed comprehensive train, this system was focused on strengthening the EMC ability of the equipment. By using high speed comprehensive inspection train to execute debugging commissioning and daily detection tasks, the feasibility and reliability of dynamic subsidence detection system has been verified.For dynamic subsidence detection algorithm, this paper proposed two schemes, based on single sensor and based on multi sensors. By comparing the experimental data, the advantages、disadvantages and applicability of two schemes were analyzed. The first scheme used bogie pitch gyroscope as single sensor, whose structure and algorithm is simple. However, because the axis of the gyro is rotated with the carrier, it is difficult to reflect the actual subsidence of the line, and the scope of application is limited. By multi sensor data fusion and compensation, the second scheme can get the curvature of plane and vertical section. Accordingly to the curvature and line document, the line subsidence can be distinguished.At the same time, according with experimental data of ZhengXu and JingJin, this paper summarized the subsidence feature recognition, and pointed out the future research direction.