Article Preview
Top1. Introduction
As an important part of the transportation system, bridges play an important role in national transportation and economic development (Lertpaitoonpan et al. 2000, Chen et al. 2016). However, with the growth of bridge service time, the continuous erosion under the natural environment, and the rapid growth of traffic volume and load, the local structure of the bridge appears natural accumulation or accidental damage. Once the bridge structure is damaged (Kameshwar et al. 2021, Obrien et al. 2021) and cannot be maintained in time, it will not only shorten the service life of the bridge, but will also affect the safety of driving, and even may cause bridge safety accidents when the situation is serious. It is significantly important to maintain the health of the bridges in transportation system.
In recent years, by integrating science and technology (Celik et al. 2020), sensor technology (Ristic et al. 1994, Ali et al. 2014), signal processing & analysis technology (O'Neal et al. 2018, Nagarajaiah et al. 2010), and structural analysis (Domede et al. 2013, Wen et al. 2021) into bridge health detection, the bridge health detection can include more contents and the detection accuracy is improved. In the bridge health detection, the vibration detection (Chang et al. 2016, Wickramasinghe et al. 2016) can effectively reveal the overall dynamic characteristics of the bridge which belongs to the overall bridge structure detection method. Through vibration detection to find out the reasons of vibration changes inside the bridge, it can grasp the safety status of the bridge structure in real time.
In this paper, the actual vibration signal of the bridge is sampled by the acceleration sensor, and then the obtained signal is processed and analyzed, finally, the state characteristics of the measured object are distinguished by combining with the results of time-frequency analysis. When vehicles pass through the bridge, the vibration of bridge structure will be produced, which often presents non-stationary characteristics. The traditional stationary signal analysis method is either time domain analysis or frequency domain analysis. They cannot obtain the local information of time domain and frequency domain at the same time. Thus, these methods cannot deal with the response of non-stationary and nonlinear system. The time-frequency combined analysis can deal with non-stationary signals by observing the law of signal changing in time and frequency domain at the same time. Wavelet analysis (Ding et al. 2008) and Wigner Ville distribution (Liao et al. 2018) have been proved to have obvious advantages in processing non-stationary signals, and are widely used in vibration signal, image processing and other engineering fields.