The purpose of this study is to facilitate damage detection and health monitoring in concrete bridge girders without the need for visual inspection while minimizing field measurements. Simple span beams with different geometry, material and cracking parameters were modeled using Abaqus finite element analysis software to obtain stiffness values at specified nodes. The resulting databases were used to train two Artificial Neural Networks (ANNs). The first network (ANN1) solves the forward problem of providing a health index parameter based on predicted stiffness values. The second network (ANN2) solves the inverse problem of predicting the most probable cracking pattern. For the forward problem, ANN1 had the geometric, material and cracking parameters as inputs and stiffness values as outputs. This network provided excellent prediction accuracy measures (R² > 99%). ANN2 had the geometric and material parameters as well as stiffness values as inputs and cracking parameters as outputs. This network provided less accurate predictions compared to ANN1, however, ANN2 results were reasonable considering the non-uniqueness of this problem's solution. An experimental verification program will be conducted to qualify the effectiveness of the method proposed. This test program is described in details in the present paper.
[1]
Foad Nazari,et al.
Detection of multiple cracks in beams using particle swarm optimization and artificial neural network
,
2011,
2011 Fourth International Conference on Modeling, Simulation and Applied Optimization.
[2]
Yacoub M. Najjar,et al.
Simulating the stress–strain behavior of Georgia kaolin via recurrent neuronet approach
,
2007
.
[3]
Yacoub M. Najjar,et al.
Intelligent Damage Detection in Bridge Girders: Hybrid Approach
,
2013
.
[4]
Hongpo Xu,et al.
Damage Detection in a Girder Bridge by Artificial Neural Network Technique
,
2006,
Comput. Aided Civ. Infrastructure Eng..
[5]
S. Masri,et al.
Application of Neural Networks for Detection of Changes in Nonlinear Systems
,
2000
.