by brian walsh & arturo gonzález
DESCRIPTION
An Overview of the Application of Neural Networks to the Monitoring of Civil Engineering Structures. By Brian Walsh & Arturo González. With thanks thanks to the 6 th European Framework Project ARCHES for their generous support. Contents. Introduction to neural networks (NNs) - PowerPoint PPT PresentationTRANSCRIPT
An Overview of the Application of Neural Networks to the Monitoring of Civil Engineering Structures
By Brian Walsh & Arturo González
With thanks thanks to the 6th European Framework Project ARCHES for their generous support
Contents
1. Introduction to neural networks (NNs)
2. Damaged beam simulation
3. Network training
4. Results
• Number of hidden nodes
• Number of input nodes
• Size of training set
1. Introduction to NNs
Synapses
Cell Body
Activation Function
Weighted Connections
1. Introduction to NNs
2. Damaged Beam Simulation
2. Damaged Beam Simulation
Reduced Stiffness
2. Damaged Beam Simulation
3. Network Training
Error BP
4. Results
Net Output Category
Net indicates lowest EI value in correct element
Net indicates lowest EI value in correct element, and healthy elements elsewhere
EIpredicted / EItarget < 1.03
Best performance Category
Location Identified
EI Profile Identified
Severity Estimated
Beam Identified
4. Results
4.1 Number of Nodes in Hidden Layer
4. Results
4.1 Number of Nodes in Hidden Layer
4. Results
4.2 Number of Input Nodes
4. Results
4.3 Size of Training Set
5. Conclusions
• NNs can be an effective tool for damage detection
• NNs sensitive to number of nodes & training patterns
• Further work
Thank you for listening!