A comparative study of wind-induced dynamic response models of long-span bridges using artificial neural networks, support vector regression and buffeting theory

https://doi.org/10.1016/j.jweia.2020.104484Get rights and content
Under a Creative Commons license
open access

Highlights

  • MLP and SVR algorithms were trained with buffeting response from full-scaled measurements of the Hardanger bridge campaign.

  • Analytical estimation of buffeting response is computed with buffeting theory applied to a FEM model of the bridge.

  • Machine learning showed to be accurate for surrogate modelling of buffeting response enhancing the simulation speed.

  • SVR algorithms showed better accuracy than MLP and buffeting theory for full-scaled buffeting response estimation.

Abstract

Long-span cable-supported bridges are structures susceptible to high dynamic responses due to the buffeting phenomenon. The current state-of-the-art method for buffeting response estimation is the buffeting theory. However, previous research has shown discrepancies between buffeting theory estimates and full-scale measured response, revealing a weakness in the theoretical models. In cases where wind and structural health monitoring data are available, machine learning algorithms may enhance the buffeting response estimation speed with less computational effort by bypassing the analytical model’s assumptions. In this paper, multilayer perceptron and support vector regression models were trained with synthetic and full-scale measured data from the Hardanger Bridge. The analytical response was also computed from buffeting theory applied to a finite element model of the bridge, and the estimates are compared. The prediction accuracy was evaluated with the normalized root mean square error, the mean absolute percent error and the coefficient of determination (R2). The machine learning models trained with synthetic datasets achieved very high accuracy with normalized root mean square errors ranging from 1.46E-04 to 7.21E-03 and are therefore suitable for efficient surrogate modeling. Further, the support vector regression model trained with the full-scale measured dataset achieved the best accuracy compared with the other methods.

Keywords

Long-span bridges
Full-scaled measurements
Buffeting response
Multilayer perceptron
Supporting vector regression

Cited by (0)