Improvement and validation of a novel moving load identification technique

With the growth of freight traffic and the ever-increasing use of structural monitoring on existing bridges, the installation of sensor networks has become a common practice. Based on this consideration, a recent study proposed by the Authors adopts accelerometer recordings to identify heavy vehicles in transit. The approach is based on a multi-parametric identification method that, by comparing the measured dynamic responses with those obtained from a one-dimensional analytical model of the bridge, allows for the estimation of the descriptive parameters of the vehicle: the total weight and the loads distribution on the axles, their respective spacings, and the transversal eccentricity of the moving vehicle with respect to the deck axis. The analytical beam model considers flexural–torsional couplings, so as to consider the effects of a potential skew angle in the deck geometry. Parameter identification is performed using the Differential Evolution (DE) genetic algorithm, already tested for different objective functions, integrating several experimental quantities extracted from both the time and frequency domains. The method has been validated using numerically simulated data containing noise pollution. Here the method is improved and validated considering more generalized conditions. The main novelty of this approach lies in the effective integration between the analytical structural model and the DE algorithm, capable of accurately reconstructing the distribution of vehicular loads on skewed road bridges.