Bridge foundation scour depth identification method based on vehicle braking excitation and generalized regression neural network

Currently, scour detection for short and medium-span bridges often relies on specialized underwater inspection equipment or health monitoring systems, which can be both costly and time-consuming. This study proposes an innovative method that integrates with the existing periodic dynamic load testing framework. By utilizing vehicle braking tests and surrogate models, this method seeks to accurately identify the scour depth of pile foundations. First, vehicle braking tests are conducted to induce significant longitudinal vibrations in the bridge, thereby enabling efficient modal testing of the pier. Second, a surrogate model is developed based on the generalized regression neural network (GRNN) to quantify the relationship between scour depth and the modal parameters of piers. To enhance the reliability of the GRNN model for practical applications, the updated parameters are expanded to include the bearing stiffness and soil properties. The proposed method is validated through numerical simulations as well as field tests on typical bridges. The results demonstrate that braking actions can excite multiple scour-sensitive modes of the pier. Moreover, all updated parameters can be accurately identified using merely two to three measurement points on the pier, and the field-identified results show a high level of consistency with the actual conditions. This method enables a preliminary evaluation of foundation scour conditions and screening for defective substructures, thereby reducing excessive and unnecessary underwater inspections or scour monitoring.