Title:
SHM for Damage Prediction Leveraging Kinematic Contact Enforcement-based Moving Load and Deep Learning Techniques
Author(s):
Ham
Publication:
Web Session
Volume:
Issue:
Appears on pages(s):
Keywords:
DOI:
Date:
10/23/2022
Abstract:
This paper presents a study on the development of hybrid models featuring structural health monitoring (SHM) inspection using bridge weigh-in-motion (BWIM) physics-based models. Artificial intelligence (AI) techniques helped improve the structural damage prediction during SHM inspection of bridges. This study introduces a comprehensive assessment of 1) a unique finite element (FE) simulation approach, which leverages the kinematic contact enforcement (KCE) method, verified with the vehicle-bridge interaction theory, and 2) machine learning (ML) techniques to identify and automatically predict structural damages from the structural response. The KCE method is a new approach to simulating vehicle motion in a BWIM model, which is used to carry out actual structural responses to motion providing contact conditions between elements (i.e., contact type, material properties, and element moving speed), which enables the realistic vehicle motion and structural response to be simulated. The FE model is designed with four different classes of damages with three different damage locations applied under different load conditions. These responses obtained from the FE simulation are further examined by using the feature selection method, which provides the importance rank for ML models. A prediction model includes: 1) a decision tree, 2) a support vector machine, 3) backpropagation, and 4) XGBoost. Among these prediction models, the deep learning XGBoost with its assembly decision tree provided the most reliable results. The results in this paper verify that structural damage prediction can be achieved by using ML as well as the BWIM structural response, which provides high accuracy in bridge damage prediction.