Thermal response separation for bridge long-term monitoring systems using multi-resolution wavelet-based methodologies

This paper presents an application of the multi-resolution wavelet-based methodology to uncover thermal effects from bridge responses based on the distinguished frequency bandwidths. First, periods of daily and seasonal temperature effects are verified using field-measured mid-span deflections of a cable-stayed bridge. Then, the formula to determine the decomposition level is derived on the basis of its capability in dividing the total bandwidth of signals. Moreover, simulated signals on two scales (i.e., minutely and hourly) are applied to validate the effectiveness of the proposed formula. Meanwhile, wavelet basis function ‘coif5′ is selected for thermal response separation due to its more stable performance compared to other functions. In-situ measured mid-span deflections of Xihoumen bridge are taken as an example to validate the effectiveness of the method through the similarity of daily vehicle gross mass and daily mean reconstructed deflection signals (without effects of temperature actions). As a result, the trends of daily mean deflections and vehicle gross mass are generally similar, except for the beginning and end of the time window. In conclusion, it is of practicality to determine the parameters by following the suggestion in this paper when using multi-resolution wavelet-based method to separate temperature responses.