If Bridges Could Talk

Transcription

1 If Bridges Could Talk Maria Feng, Reinwick Professor Director, Sensing, Monitoring and Robotics Technology (SMaRT) Lab, Associate Director, NSF IUCRC Center for Energy Harvesting Materials & Systems Columbia University NAS Workshop on the Role of Advanced Technologies in Structural Engineering for More Resilient Communities, Irvine, CA, 9/26 /17

2 188 million daily crossings on structurally deficient bridges. $123 billion needed for bridge rehabilitation 21 years at the current funding level One of 10 bridges with weight/speed restrictions BUNÆN/FLICKR 2

3 20 years of exploration to make bridges smart 3

4 Smart Bridges I-35W Bridge Tsing Ma Bridge: $22,875 per sensor 4 ssguysutterstock Source: Jeffery Funk

5 Cost Frequency [Hz] f 3= (# of years) f 2= (# of years) f 1= (# of years) Jan 2002 Jan 2003 Jan 2004 Jan 2005 Jan 2006 Jan 2007 Jan 2008 Jan 2009 Winter 02 Spring 02 Summer 02 Fall 02 Winter 03 Spring 03 Summer 03 Fall 03 Winter 04 Spring 04 Summer 04 Fall 04 Winter 05 Spring 05 Summer 05 Fall 05 Timeline 1st Mode Identified w/o Vehicle Interference 1st Mode Identified with Vehicle Interference 2nd Mode Identified w/o Vehicle Interference 2nd Mode Identified with Vehicle Interference 3rd Mode Identified w/o Vehicle Interference 3rd Mode Identified with Vehicle Interference Winter 06 Spring 06 Summer 06 Fall 06 Winter 07 Spring 07 Summer 07 Fall 07 Winter 08 Spring 08 Summer 08 Fall 08 Winter 09 Spring 09 Benefit? 5

6 Evolution toward Internet of Smart Bridges Traditional sensors Emerging sensors Ubiquitous sensors Smart sensors Fiber optic and MEMs sensors Subsurface imaging sensor Vision sensors Smartphone sensors Multifunctional Computing Communication Control Smart materials Energy harvesting Data analytics for damage detection & condition assessment Machine learning Diagnostics & prognostics Intelligent maintenance Post-disaster response Resiliency planning Lab tests Field instrumentation Long-term database Networked systems Decision making 6

7 Computer Vision 7

8 Computer Vision for Structural Health Monitoring Distance: 300m Manhattan Bridge Displacement Measurement Distance: 340m Differential Displacement: Before: 2,438mm After: 210mm Disp(mm)

9 Internet of Things 9

10 10

11 Smartphone Sensors 11

12 Crowedsourcing Cloud Computing Machine Learning Citizen Sensors An ios App 12

13 Application of Artificial Neural Networks in SHM Frequencies & Mode Shapes Artificial Neural Network (ANN) Input layer Hidden layer Assessment of structural Health Change of Stiffness from Baseline Hidden layer Output layer 13

14 Artificial Intelligence Neural Network: Machine Learning Neural Network Shallow layers Supervised learning Deep Learning Deep Neural Network: Applications: Image recognition Speech recognition Deep layers Supervised or unsupervised learning 14

15 Robot Inspector Engineering.com Spider Man 15

16 Self-Powered Smart Structure 16 Yong-Bin Park

17 Bio-Inspired Smart Materials & Structures Piezo Fiber Receptors Floating- Gate Sensing Nodes Alavi Interconnecting Machine Learning Model Harvest energy from extreme event Harvest energy from ambient vibration Activate sensing network Continuous Monitoring Self damage assessment Self Health Diagnostics 17

18 From Diagnostics to Prognostics to Resiliency West St. On-Ramp, Anaheim, CA Frequency [Hz] f 3= (# of years) f 2= (# of years) f 1= (# of years) Jan 2002 Jan 2003 Jan 2004 Jan 2005 Jan 2006 Jan 2007 Jan 2008 Jan 2009 Timeline Winter 02 Spring 02 Summer 02 Fall 02 Winter 03 Spring 03 Summer 03 Fall 03 Winter 04 Spring 04 Summer 04 Fall 04 Winter 05 Spring 05 Summer 05 Fall 05 1st Mode Identified w/o Vehicle Interference 1st Mode Identified with Vehicle Interference 2nd Mode Identified w/o Vehicle Interference 2nd Mode Identified with Vehicle Interference 3rd Mode Identified w/o Vehicle Interference 3rd Mode Identified with Vehicle Interference Winter 06 Spring 06 Summer 06 Fall 06 Winter 07 Spring 07 Summer 07 Fall 07 Winter 08 Spring 08 Summer 08 Fall 08 Winter 09 Spring 09 Update Seismic Fragility Using 8-Years of Bridge Monitoring Data

19 From Diagnostics to Prognostics to Resiliency LA-OC Highway Network Repair Prioritization Based on Cost Based on Importance 19

20 It is Time to Make Bridges Smart