Non-contact municipal bridge structure identification based on Bayesian theory

Accurate identification of bridge influence lines is crucial for ensuring safe operation. This study proposes a non-contact identification method for municipal bridges using Bayesian theory. It utilises vehicle big data and bridge response data to extract internal bridge information. Bayesian inversion is employed to calculate the longitudinal coordinates of the influence line. The method was validated by experimental results, which showed a 6.88% error between the estimated and static influence lines obtained from a finite element model under 600 random bus operating conditions. Analysis shows that the number of working conditions narrows the 95% confidence interval width of the influence line’s posterior distribution. Recognition errors for buses, buses, and two-axle six-wheel trucks are 6.68%, 19.47%, and 18.62%, respectively. The experiment illustrates the effectiveness and feasibility of the identification method for bridge influence lines in view of Bayesian theory. It indicates that the number of operating conditions and vehicle type have an impact on the uncertainty of the posterior distribution of the influence line and the accuracy of the identification results of the bridge influence line.