Railway degradation and normal operations are the common causes of track irregularities affecting the safety and comfort of railway vehicles. These track irregularities induce both track and vehicle vibrations resulting in expensive railway maintenance costs. As such, the dynamic vehicle-bridge interactions have continued to gain considerable attention in railway engineering. Previous reports have shown that vehicle-bridge interactions are likely to cause larger dynamic vehicle responses than track irregularities and should therefore be given much attention. Several methods, including rigid wheel-rail contact and elastic wheel-rail contact, have been proposed for the identification and monitoring of track irregularities. However, track irregularity identification based on the vehicle-bridge interactions has not been fully explored, possibly due to the varying mass, damping, and stiffness matrices that result in a complex time-dependent system.
Kalman filter algorithms have been extensively used for system identifications in different research fields. Considering this, researchers have identified the possibility of using Kalman filter algorithms to identify the track irregularities. Recently, Professor Xiang Xiao from the Wuhan University of Technology together with Professor Wenai Shen from the Huazhong University of Science and Technology, developed a Kalman filter algorithm for the identification of track irregularities in railway bridges. The approach employed the use of vehicle dynamic responses, taking into account real-time vehicle-bride interactions. Their work is currently published in the journal, Mechanical Systems and Signal Processing.
In their work, the authors first established a state-space model representing the time-dependent system. This system was subjected to unknown track irregularity excitations. Next, a Kalman filter algorithm was proposed to estimate the state vector of the above system and subsequently identify its track irregularities. Lastly, the feasibility of the proposed algorithm was validated using two numerical examples: a typical real Chinese railway bridge and a two-span simply supported railway bridge.
Comparison results showed that unlike the conventional approaches, the proposed algorithm was more effective and accurate, thus enabling efficient identification and monitoring of the track irregularities. These advantages were attributed to the fact that the proposed algorithm took into consideration the vehicle-bridge interactions that exhibit a significant impact on the track/vehicle vibrations. Moreover, the effects of measurement noise, parameter uncertainty, vehicle running state, and model uncertainty on the identification of track irregularities was evaluated parametrically. The differences in the estimated responses and exact values increased proportionately with the increase in the measurement noise. Nevertheless, from the two examples, the authors observed that the proposed algorithm could successfully overcome the environmental challenges to identify the track irregularities using on-board measurement signals. Nevertheless, it was worth noting that the parameter uncertainty of the dynamic model generated negative effects that could affect the efficiency of track irregularity identification.
In summary, the study proposed a Kalman filter algorithm for identifying track irregularities of railway bridges using a vehicle dynamic response approach. Results showed that the algorithm that considered the vehicle-bridge interactions outperformed the conventional approaches, thus offering a promising solution for real-time rail-track monitoring.
Xiao, X., Sun, Z., & Shen, W. (2020). A Kalman filter algorithm for identifying track irregularities of railway bridges using vehicle dynamic responses. Mechanical Systems and Signal Processing, 138, 106582.