Evaluation of Unknown Foundations of Bridges Subjected to Scour

Abstract

Missing substructure information has impeded the safety assessment of bridges with unknown foundations, especially for scour-prone bridges. An approach based on artificial neural networks (ANNs) was developed to identify the inherent patterns in the substructure design of bridges with commonly available evidence (e.g., geometric characteristics of superstructures, loading conditions, soil properties, year built, and location) and then to generalize them further to bridges with unknown foundations. The proposed ANN models were trained with information collected for an inventory of bridges with available foundation records located in the Bryan District of the Texas Department of Transportation. Results showed that the proposed ANN models were able to make successful predictions about the foundation type and the embedment depth for deep foundations. In addition, the degree of uncertainty in the models’ predictions was evaluated by performing the random subsampling method. Graphs of the probability of exceedance were generated that allowed for factoring the predicted pile depth on the basis of a reasonable probability of failure caused by scour. As a consequence, departments of transportation can adopt a similar methodology to reclassify the bridges with unknown foundations included in the National Bridge Inventory.

Probabilistic Evaluation of Unknown Foundations for Scour Susceptible Bridges

“…A new probabilistic methodology was developed in this study, incorporating estimates of bearing capacity predicted by ANN ensemble models and Bayesian inference framework for a full probabilistic characterization of unknown bridge foundations. Unlike the deterministic solution proposed by the authors (Yousefpour et al 2014), the probabilistic solution here aims at providing a transparent and systematic assessment of uncertainty of model parameters and predictions for unknown foundations.…”

“…A deterministic evidence-based approach using ANNs has been already introduced by the authors to predict the type and the embedment depth of unknown foundations (Yousefpour et al 2014). The proposed method predicts foundation characteristics based on available evidence of a bridge, including superstructure characteristics, loading, year built, and location.…”

Machine Learning Solutions for Bridge Scour Forecast Based on Monitoring Data

“…The advantage of this method over the first method is that, unlike LSTMs that are bridge-specific, the models can be trained by data from various bridges and subsequently can make predictions for any bridge at any location. Multilayer Perceptron networks (MLPs) are feed-forward neural networks proven successful for prediction and approximation in high-dimensional problems [25][26][27]. Three-layer MLP networks were developed to predict the maximum scour depth based on selected physical and engineering features (input parameters) governing the scour process.…”

“…Multilayer perceptron networks (MLPs) are feed-forward neural networks proven successful for prediction and approximation in high-dimensional problems ( 26 – 28 ). Three-layer MLP networks were developed to predict the maximum scour depth based on selected physical and engineering features (input parameters) governing the scour process.…”

Stiffness and Strength of Stabilized Organic Soils—Part II/II: Parametric Analysis and Modeling with Machine Learning

“…ANNs have been widely used in various geotechnical applications [21], such as in prediction of pile capacity [22][23][24][25][26][27][28][29][30], constitutive modeling of soil [31][32][33][34][35][36][37], site characterization [38,39], earth-retaining structures [40], settlement of foundations [41,42], prediction of unknown foundations [43], slope stability [44,45], design of tunnels and underground openings [46,47], liquefaction [48][49][50][51][52], soil permeability and hydraulic conductivity [53], soil compaction [54,55], and soil classification [56,57].…”