Monitoring of a movable bridge mechanical components for damage identification using artificial neural networks

Abstract

This paper presents a review of the results of a structural health monitoring (SHM) study to track the performance of a gearbox and rack-pinion of an operating movable bridge. These mechanical components are critical parts of bascule type bridges and damage of these components need to be identified and diagnosed, since an early detection of faults may help to avoid major damage to the structure and also avoid unexpected bridge closures. The prediction of the gearbox and rack-pinion fault detection is carried out with artificial neural networks (ANN) using the time domain vibration signals. Several statistical parameters are selected as characteristic features of the time-domain vibration signals. Monitoring data is collected during regular opening and closing of the bridge, as well as during artificially induced damage conditions. The results indicate that the vibration monitoring data, with selected statistical parameters and particular network architecture, give good results to predict the undamaged and damaged condition of the bridge.