Application of Long-Short Term Memory and Convolutional Neural Networks for Real-Time Bridge Scour Prediction
Tahrima Hashem, Negin Yousefpour
TL;DR
This study addresses real-time prediction of bridge scour depth using deep learning on historical sensor data from Alaska and Oregon. It compares LSTM variants and CNN FCN/DCN/VCN architectures, introduces random-search hyperparameter tuning, and investigates feature impacts, including Sonar, Stage, Discharge, and time features, with transfer-learning-inspired sequential training for Oregon. Key contributions include demonstration that FCN can match or exceed LSTM performance at lower computational cost, evidence that random-search can efficiently identify strong hyperparameters, and a transfer-learning strategy that improves cross-location forecast accuracy. The findings highlight the potential and limitations of DL-based scour forecasting for diverse geological and hydraulic conditions, offering a pathway toward real-time early warning in bridge networks. The work also emphasizes the importance of bed-elevation history (Sonar) and suggests future integration with physics-informed approaches to improve generalization.
Abstract
Scour around bridge piers is a critical challenge for infrastructures around the world. In the absence of analytical models and due to the complexity of the scour process, it is difficult for current empirical methods to achieve accurate predictions. In this paper, we exploit the power of deep learning algorithms to forecast the scour depth variations around bridge piers based on historical sensor monitoring data, including riverbed elevation, flow elevation, and flow velocity. We investigated the performance of Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) models for real-time scour forecasting using data collected from bridges in Alaska and Oregon from 2006 to 2021. The LSTM models achieved mean absolute error (MAE) ranging from 0.1m to 0.5m for predicting bed level variations a week in advance, showing a reasonable performance. The Fully Convolutional Network (FCN) variant of CNN outperformed other CNN configurations, showing a comparable performance to LSTMs with significantly lower computational costs. We explored various innovative random-search heuristics for hyperparameter tuning and model optimisation which resulted in reduced computational cost compared to grid-search method. The impact of different combinations of sensor features on scour prediction showed the significance of the historical time series of scour for predicting upcoming events. Overall, this study provides a greater understanding of the potential of Deep Learning algorithms for real-time scour prediction and early warning for bridges with distinct geology, geomorphology and flow characteristics.