As rail transportation continues to advance, it provides significant convenience to the public. However, the environmental vibration and noise generated during its operation have become major concerns for residents living near rail lines. In response to these concerns, the “Law on the Prevention and Control of Noise Pollution” was promulgated in China, bringing attention to this issue within the rail transportation sector. This review summarizes the regular features observed in environmental vibration and secondary structural noise tests on different sections, including embankment sections, bridge sections, underground railroads and vehicle sections. Furthermore, it introduces several physical models utilized in the study of environmental vibration and secondary structural noise, focusing on three key aspects: excitation sources, propagation paths and the modelling of building structures. This paper also explores the introduction of data-driven models related to big data and artificial intelligence to enhance the accuracy and efficiency of research in this field and provides an overview of commonly used measures to control train-induced environmental vibrations and secondary noise in buildings. These measures are discussed in terms of excitation sources, propagation paths, and receivers, offering insights into effective strategies for mitigating the impact of rail transportation on nearby residents. Finally, this study highlights the primary findings and offers pertinent recommendations. These recommendations include considerations regarding both laboratory and on-site testing procedures, challenges associated with the deployment of data-driven models and key parameters for designing and utilizing low-stiffness fasteners.

Concrete box girder bridges occupy over 80% of the total mileage of the Chinese high-speed railway. The box girder structure has many natural modes of low frequencies, which can be excited by a train passing at high-speed, generating low-frequency bridge noise. This paper is concerned with the prediction of such bridge noise and reports a prediction model. The model, as other existing models of the same nature, also incorporates two parts, one dealing with vehicle–track-viaduct dynamics and the other dealing with sound radiation from the girders, but takes into account more features related to high-speed. In this model, vehicle–track-viaduct dynamics is dealt with in the frequency-domain based on the theory of infinitely long periodic structure and the Fourier-series method, predicting vibration frequency spectra for each and every box girder. The predicted vibration frequency spectra of all the box girders are expressed as a sum of propagating waves at different wavenumbers, and sound radiation from each propagating wave is evaluated using the 2.5D acoustic boundary element method. This approach to sound radiation enables contributions from all the box girders to be included at a reasonable computational cost. This paper continues with a comparison in bridge vibration and noise between prediction and measurement for a typical site. And finally, based on the parameters of that site, characteristics of noise radiation from the concrete box girders are studied using the prediction model.

A three-dimensional wheel–rail rolling contact model (3D WRRCM) developed using the explicit finite element method (FEM) can be used to analyze the dynamic performances of wheelsets and tracks while considering the wheelset rotation and translation, as well as the real wheel–rail contact mechanism. When using the FEM with static initialisation (traditional model), the first step is to calculate the static wheel–rail displacement, strain, and stress when applying axle loading on the wheel and neglecting the wheelset rotation and translation. Then, using these results as the initial conditions, the dynamic wheel–rail displacement, strain, and stress are calculated while considering the wheelset rotation and translation by adopting the explicit dynamic solution method. However, obtaining a solution using this traditional method is time consuming, and the solution can be divergent because of the impulse produced by the suddenly increasing wheelset rotation when proceeding from the initial condition to the dynamic. The optimised method developed in this study can solve this problem. A comparison of the results from the traditional and optimised methods reveals that in the optimised method, the softer initial impulse effectively shortens the track length needed for the calculation and improves the solution efficiency.

It is a common experience that the interior noise of an underground train vehicle increases suddenly when it just runs into a track section with steel spring floating slabs, significantly reducing passengers’ comfort. This paper investigates the reasons, and control, of such noise through field tests and model simulations. The effect of rail surface irregularity on interior noise was investigated first by analysing data collected from interior noise tests, rail surface irregularity measurement and by analysing the modal behaviour of the slab. Noise radiation of the track slab was then analysed using the acoustic boundary element method while the vibration of the steel spring floating slab track is predicted based on an infinitely long periodic structure model; And finally with the prediction models, a parameter study was performed for the steel spring floating slab, on the purpose of improving the parameters to make the track less noisy. The research presented in this paper may provide a reference for the design of steel spring floating slabs in terms of noise control.

For large infrastructures, dynamic displacement measurement in structures is an essential topic. However, limitations imposed by the installation location of the displacement sensor can lead to measurement difficulties. Accelerometers are characterized by easy installation, good stability and high sensitivity. For this regard, this paper proposes a structural dynamic displacement estimation method based on a one-dimensional convolutional neural network and acceleration data. It models the complex relationship between acceleration signals and dynamic displacement information. In order to verify the reliability of the proposed method, a finite element-based frame structure was created. Accelerations and displacements were collected for each node of the frame model under seismic response. Then, a dynamic displacement estimation dataset is constructed using the acceleration time series signal as features and the displacement signal at a certain moment as target. In addition, a typical neural network was used for a comparative study. The results indicated that the error of the neural network model in the dynamic displacement estimation task was 9.52 times higher than that of the one-dimensional convolutional neural network model. Meanwhile, the proposed modelling scheme has stronger noise immunity. In order to validate the utility of the proposed method, data from a real frame structure was collected. The test results showed that the proposed method has a mean square error of only 5.097 in the real dynamic displacement estimation task, which meets the engineering needs. Afterwards, the outputs of each layer in the dynamic displacement estimation model are visualized to emphasize the displacement calculation process of the convolutional neural network.