35w and 10 years of change - umn ccaps · 1/30/2018 1 35w and 10 years of change ed lutgen| bridge...

1/30/2018

1

35W and 10 Years of ChangeEd Lutgen| Bridge Construction and Maintenance Engineer

Catherine French | Professor, University of MinnesotaKevin Western | State Bridge Engineer

Topics

2

• Bridge Info, Collapse, and Initial Response

• On Site Perspective

• National Transportation Safety Board (NTSB) Findings

• Changes in MnDOT and National Processes

• New Bridge – Selection, Design, and Construction

• U of M Research

• Chapter 152 Funding

• Final Thoughts

1/30/2018

3

I-35W Bridge

Universityof

Minnesota

DowntownMinneapolis

http://maps.google.com/1/30/2018

I‐35W Bridge

4

Bridge Background• Bridge Completed in 1967

• 1907 Feet Long

•3 span continuous truss

• Main span 456’

• 2 trusses

• 11 approach spans

• Continuous steel multi beam

• Continuous concrete slabs

• Average Daily Traffic (ADT) 141,000

• 4 lanes each direction

• Structurally Deficient – SR 50.0

• Annually Inspected – In depth Fracture Critical

• MnDOT 20 year plan called for Replacement 2020‐25July 1967

1/30/2018

1/30/2018

2

51/30/2018

Two Studies of Fatigue Potential

6

• 2001 University of Minnesota –Fatigue Evaluation of the Deck Truss of Bridge 9340

• 2007 – URS – Fatigue Evaluation and Redundancy Analysis

• Multi Girder Approach Spans Had Required Retrofits for Fatigue Cracks

1/30/2018

Previous Bridge Modifications

• 1977 Concrete overlay

• 1998 Railing repair, drainage system, minor deck repair

• 1999 Painting portion of truss spans

• 1999 Anti icing system

• 2001 Curb & slab repair

Average yearly MnDOT Maintenance Hours: 500

71/30/2018

Contract Maintenance Work at Time of Collapse

• Contract to replace concrete overlay, repair joints and lighting, and install guardrail

• Work completed:North bound – two inside lanesSouth bound – two outside lanes

• Scheduled completion date was September 30, 2007

• Cost ‐ $9 million

• Contractor employees and MnDOT inspectors on bridge at collapse

81/30/2018

1/30/2018

3

August 1, 2007: The Collapse

Collapse occurs at 6:05 p.m.Wednesday, August 1, 2007

9

Over 100 vehicles on the bridge at the time

13 fatalities140 injuries1/30/2018 10

North End – I 35W Bridge – South End• 1/30/2018

11North End• 1/30/2018 12

Pier 7 North End Main Span• 1/30/2018

1/30/2018

4

13Main Span

• 1/30/2018 14Pier 6 South End Main Span• 1/30/2018

15South Spans• 1/30/2018 161/30/2018

1/30/2018

5

Emergency Response

• 6:05 p.m.

• Numerous concurrent calls to State Patrol Dispatch (911) and from field employees to RTMC and Maintenance Dispatch

• Motorists on bridge, construction workers, citizens in area assist injured

• Emergency personnel from Twin Cities and Western Wisconsin respond

• 6:10 p.m.

• District Emergency Operations Center activated

• Immediate traffic control for ramp and freeway closures provided by FIRST units, maintenance units, and contractors in the vicinity

17

MnDOT’s Regional Transportation

Management Center (RTMC)

1/30/2018

Response the First 12 Hours

• 6:20 p.m. – State Emergency Operations Center

• Unified command center set up on collapse site. Authority in command changed as incident progressed:

• Minneapolis Fire Department in charge of rescue

• Hennepin County Sheriff in charge of recovery (MnDOT assisted with some demolition)

• MnDOT assumed command after recovery was completed

181/30/2018

First 12 Hours – (continued)

6:20 p.m.

• Started converting I‐35W temporary traffic control measures to longer term traffic control standards with barriers and signs

• 20 changeable message signs activated

• MnDOT Metro District provides maintenance staff and equipment for security efforts

• 800 megahertz communication system was critical for responders

191/30/2018

First 12 Hours – (continued)

• 7:00 p.m.• Over 150 employees activated, most just

returned without a call• MnDOT structural engineers called to site

• 10:00 p.m.• Governor and Mayor provide an update to

public. Number of victims unknown.• Rescue Operations Ended; Recovery

Begins• 11:00 p.m.

• Detour maps for morning rush posted on MnDOT web site

• Overnight• Expanded signing and barricades of closed

I‐35W • Converted T.H. 280 to a freeway

201/30/2018

1/30/2018

6

I‐35W Detour Map

211/30/2018

August 2, 2008

– Recovery operations continue– Engineers assess stability of wreckage for recovery personnel

–NTSB leads investigation–MnDOT contacts Wiss Janney Elstner (WJE) and TranSystems/Lichtenstein for investigation

– Engineering team begins to organize for rapid replacement

221/30/2018

Media Coverage August 2nd

– Governor conducts interviews throughout morning

– 2:00 pm Press Conference:Governor announces

• Emergency statewide bridge inspections beginning with underdeck trusses

• Forensic Investigation Team• Review of MnDOT Inspection

Program

23

NTSB Chairman Mark Rosenker

1/30/2018

Managing Media Requests

24

• Held Daily 2:00 pm Press Conference• Only means to manage volume of requests• Format was statement, update on specific issue,

open to questions• Length was held to a reasonable time – 30 to 45

minutes• When it ended, held questions till next day

• Governor and Commissioner directed MnDOT to be transparent

• Document requests were overwhelming• Website posting of plans, inspection reports,

bridge studies• Dedicated I‐35W Website includes all documents• Document Management System established to

gather and store all bridge information• Later removed inspectors names after harassing

phone calls to their homes1/30/2018

1/30/2018

7

Managing Media Requests (continued)

• Our Priority was local media requests

• Within 24 hours media began their own investigations and speculation.

• Our time was consumed correcting misinformation

• After 3 weeks ended face to face interviews and responded to written questions.

251/30/2018 26

U.S. Navy Dive Team

Recovery of Victims and Bridge Removal• Access to site controlled by police and fencing – 24/7

• Site protocol followed during recovery of victims

• Navy Divers from Norfolk, VA assist in recovery

1/30/2018

• Aug. 20 – Navy Divers recover 13th victim, site turned over to MnDOT

• Sept. 6 ‐ Navigation channel opened to commercial traffic

• Sept. 27 – Final steel removed from river

271/30/2018

Topics

1/30/2018 28

• My Experiences

• Inspection Changes

• Construction Changes

• Load Rating Changes

• Maintenance Changes

• Programming Changes

1/30/2018

8

My Experiences – August 1

1/30/2018 291/30/2018


1/30/2018 30


1/30/2018 31

My Experiences ‐ Visuals

1/30/2018 32

1/30/2018

9

My Experiences ‐ Documentation

1/30/2018 33

My Experiences – Bohemian Flats

1/30/2018 34

My Experiences ‐ Afton

1/30/2018 35

My Experiences – Final Report

1/30/2018 36

1/30/2018

10

My Experiences ‐ Pieces

1/30/2018 37

• 35W Steel Pieces

• Historical Society

• Victims

• First Responders

• Educational

• Recycling

My Experiences ‐ Universities

1/30/2018 38

The Cause ‐ U10W

1/30/2018 39

The Cause – U10W

1/30/2018 40

Compression diagonal

Tension diagonal

Stress

Yieldstress

0

Orange and red shading: exceeds yield stress

Allowable

1/30/2018

11

The Cause – U10W

Compressiondiagonal

Tensiondiagonal

1/30/2018 41

The Cause – U10W

1/30/2018 43

Vertical U10‐L10

Tension Diagonal U10‐L11

Top Chord U9‐U11

Compression Diagonal U10‐L9

The Cause – U10W

1/30/2018

Center of spanU10

44L9

Pier

1/30/2018

12

The Cause

U2 U4 U6 U8 U10 U12 U14

L1L3 L5

L7 L9L11 L13

U0

1/2” thick gusset plate (50 ksi) 10 of 29 gusset plates5/8” thick gusset plate (50 ksi) 4 of 29 gusset plates1” thick gusset plate (50 ksi) 13 of 29 gusset plates1 3/8” thick gusset plate (100 ksi) 2 of 29 gusset plates

1/30/2018 45

NTSB/WJE Theories Considered

• Corrosion damage• Fracture of a floor truss• Pre‐existing cracking or fatigue• Locked bearings and piers• Terrorism• Sinkhole south side of river• Railroad vibrations• Seismic movement• Approach creep• Material deficiencies• Fire truck loading (overload)• Scour undermining the footing• Thermal loads• Overloaded bridge (contractor and traffic)• Drilled shaft hinging• Settling or moving piers• U of MN coal tunnel 1/30/2018 46

Why Change

“Failure is a great teacher, and I think when you make mistakes and you recover from them and you treat them as valuable learning experiences, then you’ve got something to share.”

Steve Harvey, famous actor and author

1/30/2018 47

Failures Lead to Change

• 1967 – Silver Creek bridge in WV/OH collapse

• 1987 – Schoharie Creek in NY collapse

• 2007 – 35W bridge

1/30/2018 48

1/30/2018

13

Inspections – FHWA

Bridge Inspection Reference Manual (BIRM) 2006

• Gusset plate only briefly mentioned couple times

• “By contrast, a slightly deteriorated gusset plate at a panel point of a truss may not be critical.” page 4.4.4

1/30/2018 49

2006 BIRM


Bridge Inspection Reference Manual (BIRM) 2012

• Added Chapter 10.8.1 specifically for gusset plates.

• Over 30 pages of details, failure mechanisms, repairs, inspection techniques, conditions to look for and document.

1/30/2018 50

2012 BIRM

Inspections – Gusset Plate

• New Inspection Elements for Trusses

• Element 162: gusset plates

• Element 120:trusses

• Element 113: stringer

• Element 152: floorbeam

1/30/2018 51


23 FHWA Inspection Metrics

• Assesses compliance with NBIS at 23 CFR Part 650, subpart C

• Compliance levels

• Compliance

• Substantial compliance

• Conditional compliance

• Non‐compliance

1/30/2018 52

1/30/2018

14

• MN has 205 agencies with at least 1 bridge

• Agencies broken into classes based on number of eligible structures

Inspections – Bridge Office Audit

1/30/2018 53


1/30/2018 54

• Constant improvement since 2012 implementation

• All metrics have shown similar improvements


70%

75%

80%

85%

90%

95%

100%

2011 2012 2013 2014 2015 2016

Metric 6 Compliance

1/30/2018 55

Inspections – Section Loss

Pre 2007 inspection notes for gusset plate

• Up to 3/16” section loss on a ½” plate

• 3/16” / ½” = 38%

• Take entire critical cross section not localized damage

1/30/2018 56

1/30/2018

15

Inspections – Section Loss

Better Inspection Documentation

• BSIPM pages section loss guidance B.4.1.2

• UT and Phased array

1/30/2018 57

Inspections – Fracture Critical

• All FC inspections for state and local agencies performed by Bridge Office

• Best access equipment

• More NDT trained staff

• Consistent documentation

• Meet FHWA Metric 10 and 16

• Fracture critical bridges have an independent structural review every 2 years.

1/30/2018 58

Inspections – SIMS

SIMS – Structure Information Management System

• All inspection reports

• Pictures

• Migrated old Pontis data into SIMS

• Reports for inspections due

• Queries of data

• Program Administrator reviews

• Bridge maintenance module

1/30/2018 59

Construction – Specs

1/30/2018 60

Figure 16 NTSB Final ReportNTSB Factual Report

1/30/2018

16

Construction – Specs

1/30/2018 61

Specification 1513 Load Restrictions

• Material 65,000 lb per 1000 ft2

• Materials 25,000 lb per 100 ft2

• Truck 80,000 lb

• All truck, material 200,000 lb

1/30/2018 62

• Prior 2007

• No gussets analyzed

• Simple span trusses only

• Post 2007

• All trusses

• Structural gussets

Load Rating – Analysis

Truss and Gusset Plate Rating

• Blatnik

• Winona

• Osceola

• DeSoto

1/30/2018 63

Truss ProjectsTruss Projects

1/30/2018 64

Truss Projects

1/30/2018

17

Bolt Replacement

65

Maintenance ‐ Resources

Tipped Bearing

Debris Removal

In response to an identified condition.

Rail Repair

Approach panel foam jacking

Spall Repair

1/30/2018 66

Identify Maintenance Tasks from: • Inspection Reports • Assessments• Flushing• Date Last Performed• Condition of Elements

Maintenance – Work

1/30/2018

Maintenance ‐ Data Tracking and Reporting

67SIMS Maintenance Module

Oracle Business Intelligence

1/30/2018

Bridge Maintenance Training

68


• Bridge Maintenance Academy (BMA I – III)

• Welding

• Shotcrete

• High Angle Rescue

• Statewide annual worker conferences

1/30/2018

1/30/2018

18


69

Bridge Maintenance Training• Preventive Maintenance eLearning Modules

• Crack Sealing

• Gland Repair

• Flushing (in progress)

• Poured Joint Sealing (in progress)

• BMA I (in progress)

• http://www.dot.state.mn.us/bridge/training.html

• Bridge Maintenance Manual – soon to be on the MnDOTBridge Office website

1/30/2018

Bridge Planning ‐ BRIM

• BRIM Process:

• Ranking each bridge based on the probability and consequence of a service interruption (Bridge Planning Index).

• Identifying preservation and improvement needs.

• Conducting an expert review.

1/30/2018 70

Risk Assessment

• Work Type• Costs• Timeframe

District Review

35W Bridge Replacement

• Process for Contractor Selection

• Design, Construction and Innovation

• What Went Well and Challenges

• Health Monitoring (Cathy)

• Final Thoughts

711/30/2018

New Bridge is Needed Quickly

• 141,000 cars a day used the bridge

• One of the busiest bridges in Minnesota

• Serves major traffic areas of:

• U of M

• Downtown Minneapolis

• $400K a day estimated in road users costs

• MnDOT needs to demonstrate abilities and begin rebuilding public confidence

721/30/2018

1/30/2018

19

Planning for Replacement Begins Early Morning Hours of August 2

• Replacement team begins to form

• Discussion begins on delivery method

• Need quality, buildable replacement bridge plan

731/30/2018

New Bridge

• August 2nd• Design build best value procurement method

selected• August 3rd

• Core team identified• Three experienced design build project managers

on team• Bridge Office personnel

• Three bridge designers• Two construction engineers with bridge

experience• Empowered to be decision makers

741/30/2018

Design Build – Why?

75

• Considers quality and cost

• 7th design build best value project

• Past design build projects successful

• Geometric improvements desired

• Public input/communication and

visual quality important

• High level of national interest

• Utilize expertise of DB teams

• Allows construction to begin quickly

• First complex bridge with design build process

1/30/2018

Challenges of I‐35W Site

1/30/2018 76

• Demolition – when will it be completed

• 24/7 work will be necessary

• Hydraulic scour at site

• Work adjacent to lock and dam

• Substantial winter work required

1/30/2018

20

Procurement Process

• Issued Statement of Qualification on August 4

• Held daily confidential meeting with each proposing team

• Held public meetings, DBE and utility coordination meetings

• Addressed media and public questions

• Developed Request for Proposals in 3 weeks

• Responded to clarifications and issued addenda

1/30/2018 77

Established Procurement Timeline

• August 1 – Collapse occurs• August 4 – Issue Request for Qualifications• August 8 – Short listed teams• August 23 – Request For Proposals released• September 14 ‐ Technical proposals received• September 18

• Evaluation and Scoring Completed• Financial Proposals Received

• September 19 – Project Letting• September 20 – City of Minneapolis Grants

Municipal Consent• Project Award – Early October

1/30/2018 78

WHAT DID WE HEAR?“Slow Down; You are Moving Too Fast”

• A landmark bridge should be considered

• You don’t know why the previous bridge collapsed

• No ugly freeway‐style bridge

• Design bridge for a future light rail line

• Minnesota politics

1/30/2018 79

RFP Evaluation/Scoring Criteria

• Quality (50 percent)

• Quality control/quality assurance

• Safety

• Aesthetics/Visual Quality (20 percent)

• Visual enhancements to the structure

• Enhancements (15 percent)

• Public Outreach/Involvement (15 percent)

• Impacts to the public

• Approach to communication

1/30/2018 80

1/30/2018

21

Co‐Housing

• Important to establish a partnership relation

• All team members represented

• 24/7 work until design was way ahead of construction

• Scheduled and just‐in‐time meetings every day

• Prioritizing review tasks a continual challenge

1/30/2018 81

Oversight – Peer Review Team

• Parsons Transportation Group brought concrete box expertise to the team

• First test of our Peer Review process• Bridge Modeling – In parallel with EOR

• Established protocol for results comparisons

• Reviewed temporary works as well

• Same team (and members) provided field support• Great benefit in understanding what is important

• Unique staff with both experiences and background

1/30/2018 82

Bridge Description

•Four‐span bridge approximately 1225’ in length

•Cast‐in‐Place approach spans and Precast Segmental river span (120 segments)

•Variable depth superstructure 25’ to 11’

1/30/2018 83

Bridge Description• Two parallel bridges, each with two box girders

• Striped for 5 lanes each direction (10 total) with 13’ and 14’ shoulders (actual design loading considers for 7 lanes each direction)

• Future configuration of 4 lanes each direction plus light rail line or bus transit lane (lane drop for ramps)

1/30/2018 84

90’4”

11’

1/30/2018

22

100 Year Design Life

1/30/2018 85

•Include corrosion resistant design details with post‐tensioning

•Utilize high performance materials

•Provide multiple layers of protection of key structural elements

•Provide high quality construction

High Performance Concrete

• Performance specifications • Impacts on schedule and quality• Strength, permeability, chloride resistance

• Slag, fly ash, micro silica• Self consolidating concrete

• Primarily used for drilled shafts• Viscosity Modifying Admixtures (VMA)

• Helped prevent segregation in mix• Contractor and supplier innovation

• Developed mix designs• 45 mix designs• Silica Fume in Superstructure (approx. 4% by vol)• Used high range water reducing agents and

retarders• Multiple test pours and mock ups

1/30/2018 86

Public Input in Design

1/30/2018 87

• First time charrette process used in MN

• Pier shape

• Rail type

• Native stone abutments

More than 400 peoattended on July 5

Sidewalk Talk: More than 400 people attended on July 5, 20081/30/2018 88

1/30/2018

23

Safety Management

• Partnership• MnDOT• Flatiron‐Manson• Mn/OSHA

• Training of all workers assigned to project• Required escorts for visitors

• Large safety team

• Audits performed weekly

• Consistency from top to bottom

• MnDOT incentive

1/30/2018 89

Falsework System

1/30/2018 90

1/30/2018 91

Segment Casting and Delivery

• Contractor used 35W roadway

• Long line casting method

• First time in Minnesota

• 8 separate beds

• Travelling housing

• Side by side segments with closure pour

1/30/2018 92

1/30/2018

24

Casting Yard

1/30/2018 931/30/2018 94

1/30/2018 95 Lifting Segments961/30/2018

1/30/2018

25

Transporting Segments1/30/2018 97 1/30/2018 98

Segment Erection1/30/2018 99 1/30/2018 100

1/30/2018

26

Aesthetic Lighting

1/30/2018 101

Open to Traffic September 19, 2008

1/30/2018 102

Completed in 11 Months – 3 Months Early

1/30/2018 103

Schedule and Budget

• Schedule• Contract completion date: December 24, 2008• Open to traffic: September 18, 2008• Substantial completion reached 90 days ahead of schedule

• Budget• Little cost growth (1% ±)• Incentives

• $18 million on time• $7 million No‐Excuse Bonus

1/30/2018 104

1/30/2018

27

35W ‐ What went well

• High performance concrete

• Mass concrete

• Self‐consolidating concrete

• Cold weather protection

• Involvement of Engineer of Record

• LED lighting

• Smart Bridge Technology

1/30/2018 105

35W – What Went Well

• Safety• Quality

• Engineer of Record

• Peer Review

• MnDOT staff

• Team Partnership

• Project Communication

1/30/2018 106

35W ‐ Challenges

• Anti‐icing system

• Definition of falsework requirements in project documents

• Long‐term creep for strength and fatigue design of modular joints

• Litigation ‐ but did validate selection process

1/30/2018 107

Health Monitoring

1/30/2018 108

1/30/2018

28

Department of Civil, Environmental, and Geo- Engineering13 February 2018 - Structural Seminar Series 109

I35W St. Anthony Falls BridgeSmart Bridge System

Catherine French, Carol Shield, Brock Hedegaard, Lauren Linderman, Ben Jilk Department of Civil, Environmental, and Geo‐ Engineering

University of Minnesota – Twin Cities


Project Background – Planning/Construction Phase

Design‐build project awarded to Flatiron/Manson & FiggBridge Engineers featured a “Smart Bridge” system

• strain gages• thermistors• fiber optic sensors• linear potentiometers• accelerometers• corrosion sensors

• strain gages• thermistors• fiber optic sensors• linear potentiometers• accelerometers• corrosion sensors

UMN invited to participate in instrumentation plan discussions Proposed modifications/additions to address MnDOT issues of interest Gathered material samples

“Smart Bridge” consisted of

over 500 sensors


Originally Proposed Instrumentation – Strain Gages

• Originally proposed Strain gages at midspans and supports

Behaviors of interest:• Curvatures


195 vibrating wire (VW) strain gages (38 at midspan of Span 2)

Behaviors of interest:• Load distribution (curvatures)• Shear• Torsion

Installed Instrumentation: VW, thermistors

26 VWSG 38 VWSG

243 thermistors

• Thermal gradient

Transverse VW strain gageLongitudinal VW strain gage

6 in.

Instrumented Sections

1/30/2018

29


Installed Instrumentation: Fiber optic sensors

12 Fiber optic sensors (12 ft. lengths)(6 pairs along SB Span 2)

Behaviors of interest:• Load distribution (average curvatures)

SB Span 2



• Originally proposed Accelerometers at midspans

Behaviors of interest:• Deflections

Originally Proposed Instrumentation – Accelerometers


Installed Instrumentation: Accelerometers


26 accelerometers(bottom of deck in each box)10 moveable along the length of SB Span 2

Behaviors of interest:• Frequencies• Mode shapes• Damping


Installed Instrumentation: Linear Potentiometers


12 Linear potentiometersin each box at each expansion joint(4 @ abutment 1, 8 @ pier 4)

Behaviors of interest:• Overall movement of bridge

1/30/2018

30


Project originally intended to be review of previously performed and available technologies…

…became demonstration project of how available technologies play a role in foundation health monitoring

State of the Practice and Art forStructural Health Monitoring ofBridge SubstructuresPUBLICATION NO. FHWA-HRT-09-040 MAY 2014

led by Prof. Gray Mullins, U of S Florida, Tampa

Project Background – Planning/Construction Phase

FHWA Project on Drilled Shafts


FHWA Project on Drilled Shafts

Photos from FHWA-HRT-09-040 May 2014Photos from FHWA-HRT-09-040 May 2014


Instrumentation of Pier 2

Photos from FHWA-HRT-09-040 May 2014


Summary of Bridge Instrumentation

LP Accelerometer VWSG

Fiber Optic

ThermistorsSB

NB

1/30/2018

31


Sensor wiring completion after closure pour


UMN Assisted with “Punch List”


Lightning Strikes

https://i.pinimg.com/736x/17/2c/a4/172ca4cf5b7615b8d57a519b9d2ba2b4--photo-and-video-lightning.jpghttps://www.johnweeks.com/i35w/ms32.html

• Isolation transformers installed for surge protection • Metal‐oxide varistors (MOVs) installed to help against lightning strikes.


Project Background – UMN’s role in monitoring

Objectives:• Investigate structural characteristics and changes over time• Evaluate design assumptions

(load distribution, deformations, response to environmental effects)• Evaluate effectiveness of the selected instrumentation• Develop long‐term monitoring system

1/30/2018

32


Truck Tests• 8 trucks for a total load of 400 kips (1.8MN)

• 5 static load configurations at multiple locations

• Additional dynamic load configurations

Establish “baseline behavior” – Truck Tests

Establish a baseline


Measured Strain Data for First Five Years

Daily changes: ~100 με(approximately 500 psi)

Total changes: ~400 με

*Plot represents mechanical strain plus creep and shrinkage strains

Challenges of Developing Monitoring System

8 fully loaded trucks (400 k) induce at most 50 µε


• Data from multiple sources in multiple formats collected with multiple data acquisition systems at multiple rates (static and dynamic data)

• Most individual sensor data not meaningful Groups of sensors through section required to determine curvatures,

thermal gradients,… Groups of curvature readings required to determine load distribution “Baseline” data varies over course of day/season due to thermal and

time‐dependent effects (need to combine multiple types of data)

• VOLUME OF DATA…

Model Development• Historical Data• Finite Element Model

Data Processing

• What is “normal” behavior?• How might damage manifest itself in data?

Challenges of structural monitoring


Finite Element Model (Overview)Constructed using ABAQUS to model Spans 1‐3 of SB Bridge

Pinned Pinned

ExpansionExpansion

BCs pinned @ Piers 2 and 3, rollers @ Abutments 1 and Pier 4

1/30/2018

33


FEM Development and Calibration


Calibration with measured results

Used measured data to validate models

Considered load effects (truck tests) and environmental effects (temperature/gradient)

Incorporated information from material tests Creep Shrinkage Coefficient of Thermal Expansion


Bridge Instrumentation

Linear Potentiometers

12 linear potentiometers located at expansion joints

Behaviors of interest:• Overall movement of bridge due

to time dependent effects and temperature


Challenges of Long-Term Monitoring

• Bridge behavior depends on many complex natural phenomena.

• Damage is not necessarily sudden and can be masked by normal, safe variations in behavior.

Ideal Case Typical Case

1/30/2018

34


Monitoring and Modeling the I35W Bridge

GoalIntegrate monitoring data into maintenance and inspection strategies to extend bridge life.

ChallengesDamage is not necessarily sudden, whilechanging operating conditions cause large variations in behavior.

ApproachDevelop data normalization and anomaly detection techniques for extracting unexpected changes in the structural behavior.


Thermal Gradients

Five largest measured thermal gradients compared to design thermal gradients from AASHTO LRFD (2010) and Priestley (1978).

Longitudinal stress at midspan estimated from measured mechanical strains compared to FEM stresses computed using AASHTO LRFD (2010) and Priestley (1978) gradients.

Parameters that affect behavior


Uniform Temperature

Common design assumption that coefficient of thermal expansion (CTE) is constant.

Estimates of short blocks of measured data show that CTE varied by 20% with concrete temperature.

Believed to be caused by changes in internal relative humidity and surface tension of adsorbed water.


Coefficient of thermal expansion measured from VWSGs in south span of the southbound bridge

Internal relative humidity in concrete dependent on temperature (Grasley and Lange, 2007)


Hydration Concrete strength and modulus increase with time after casting.

Creep Continued deformation under constantly applied load.

Shrinkage Loss of water in concrete as cement cures and hardens.

Time‐dependent effects


1/30/2018

35


Post‐tensioning losses

Post‐tensioning stresses reduce with time, possibly leading to tension and cracking.

Increaseddeflections

Structure continues to deform, either leading to serviceability problems or (in extreme cases) safety issues.

Monitoring and predictions

If monitoring a structure, we will need to differentiate between normal time‐dependent behavior and slow degradation of structure.


Time‐dependent effects – Why of concern?


• Time‐dependent behavior of concrete widely recognized, but many different approaches exist:

– AASHTO LRFD (2010)– ACI‐209 (1982)– CEB‐FIP Model Code (1978, 1990)

– B3 (Bažant and Baweja, 1995)– GL2000 (Gardner and Lockman, 2001)

• Creep models are very different among the listed methods.

“Asymptotic” creep limits to some maximum value

“Logarithmic” creep approaches line in log‐time space



Creep and Shrinkage

Creep and Elastic StrainH = 64%V/S = 8 in.f’c = 7450 psiσ = 1900 psi @ 10 days

Logarithmic Models:– B3 (Bažant and Baweja, 1995)

– GL2000 (Gardner and Lockman, 2001)

Asymptotic Models:– AASHTO LRFD (2010)– ACI‐209 (1982)– CEB/FIP Model Code (1978, 1990)

When compared to NU‐ITI creep and shrinkage database, all models have large coefficients of variation, ranging from 25% to 35%.



Data Normalization

Need to predict the expected behavior due to temperature and time‐dependent phenomena in order to identify anomalous readings.

Challenges• Time‐dependent behavior uncertain• Temperature dependence dominates data

Short‐term (e.g., bearing lock‐up) Long‐term (e.g., slow degradation)

1/30/2018

36


Data Normalization

• Use linear regression on measured data:

j

jjx

ttHtDI

zTdA

A

dAT

A

TdAy ,653

2

21

Uniform Temperature

Thermal Gradient

Time‐Dependent

Error and Data Jumps

• Given: Measured data y, temperature T• Unknown: Coefficients αi, time‐dependent D(t)• Technique: Assume time‐dependent behavior,

compute coefficients, and remove temperature‐dependent behavior.


Data Normalization

LP Deflection

Extract time‐dependent

Bridge TemperatureLP Deflection

Bridge Temperature

The rate of creep and shrinkage of concrete depend on the bridge temperature.

Longitudinal deflections are measured by LPs at the expansion joints.


Node 3

Northbound Span 1

Total LP

Time‐dependent deflection Temperature‐dependent deflection

Data Normalization


Data Normalization

Transform the time‐scale using the Arrhenius equation:

Converts readings with temperature history T(t′) to an equivalent adjusted age assuming constant temperature T0.

0

0

1 1'

0 'QtR T T t

adjt

t t e dt

Bridge TemperatureLP Deflection

Normalized time‐dependent behavior follows line in log‐time.

Transform time‐scale

1/30/2018

37


To compare measured deflections with time‐dependent predictions, the time can be adjusted using the Arrhenius equation:

Warmer temperatures “stretch” time. Colder temperatures “contract” time.

0

0

1 1'

0 'cUt

R T T t

adjt

t t e dt

LP DeflectionBest fit of time‐dependent predictions

Data Normalization


Data Normalization

Benefits of Data Normalization

AnomalyDetection

Deviations easier to detect from expected time‐dependent behavior than from total strains and deflections.

FEM ComparisonMeasured data using adjusted age can be compared to constant temperatureFEM results.

VersatilityEffective for temperature‐ and time‐dependent longitudinal strains and deflections.


Time-Dependent FEM• Includes aging, creep,

shrinkage, and relaxation for 150 adjusted age years

• Uses Kelvin chain model to compute viscoelastic behavior

• Includes full erection procedure to accurately capture initial stress state.


Time-Dependent FEM

LP Data and FEM Results

Southb

ound

Sp

an 1

Southb

ound

Sp

an 3

1/30/2018

38


Prototype Monitoring System

Features of Prototype Monitoring System

• Monitors expansion joint deflections, but can be expanded to other system.

• Computes short‐term errors based on predictions using Bayesian regression.

• Computes long‐term errors based on unexpected changes in slope.

• Sends automatic email updates.


Linear Potentiometer Monitoring

Benefits of Using Time‐Dependent Readings

Time‐dependent deflection from Node 4 (SB Span 1)

Narrow bounds over the course of several months

Always increasing, but at a slower rate each year

Thermal behavior is assumed not to change with time, and is therefore easier to remove from LP data than the time‐dependent behaviors.


Anomaly Detection

Approach to Anomaly Detection

Linear Potentiometer

Data

Temperature Data

Data NormalizationTime‐

Dependent Data

Finite Element Model Results

Bayesian Regression(FEM = prior dist.)

Posterior Distribution

Training Set

Test Set

Diagnose Anomalies(Short and Long‐term)


Anomaly Detection

Short‐Term Anomaly Detection• Use Bayesian regression to predict the time‐

dependent behavior over a specified test set.• Prior distribution taken from FEM results and

uncertainty of creep and shrinkage models.

95%‐credible bounds for Southbound south span LP using 1990 CEB Model Code

Prior Distribution

Posterior Distribution

1/30/2018

39


Linear Potentiometer Monitoring

Monitoring System for LP Readings

Short‐term check Every month, ensure that the time‐dependent readings lie within of the defined bounds.

Long‐term checkEvery month, ensure that the annual time‐dependent deflections are (1) positive and (2) less than the deflections from the previous year.

Bearing location Given the LP readings, output the location of the bearing and how much is left for expansion.

Thermal ResponseEnsure that the linear regression coefficients for the thermal response have not changed since the previous year.


Anomaly Detection

Short‐Term Anomaly Detection• Posterior distribution provides rational bounds

for detecting anomalous readings.

Southbound south span LP measured data with bearing lock‐up introduced 10 days from end of test set

Normalized time‐dependent data compared to posterior distribution clearly shows anomaly


Anomaly Detection

Long‐Term Anomaly Detection• Compute slope of time‐dependent data with

respect to adjusted age.• Slope is always decreasing and never negative.

Southbound south span LP

0.25 in. ramp over 18 months starting here

Southbound south span LP

Measured slope with respect to adjusted age

Posterior distribution compared to slope of measured time‐dependent data

Posterior distribution compared to slope with 0.25 in. drift over 18 months


Summary of Monitoring System

Data Normalization

Account for expected safe changes in measured structural behavior.

ModelingCreate computational models to estimate future performance of structure.

Anomaly Detection

Use Bayesian regression to derive rational bounding values for detecting deterioration and damage.

1/30/2018

40


• Difficult to identify problems with systems that may have defects prior to instrumentation.

• Truck load tests and environmental effects can be used to establish “baseline” behavior—periodically repeat tests to identify changes.

• Temperature and time‐dependent effects have prominent influence on concrete bridge behavior, possibly masking deterioration that may occur over time (e.g., corrosion).

• Data normalization on temperature can allow the measured time‐dependent behavior to be compared to FEM results and used in monitoring system.

• FEM in combination with historical data can be used to establish expected metrics and error limits.

• Anomaly detection technique using Bayesian regression has successfully been implemented to detect short‐term and long‐term problems

• Long‐term monitoring can assist in maintenance plans/strategies.

Summary and Conclusions


AcknowledgementsMinnesota Department of Transportation

Minnesota Supercomputing Institute

Truck Test and Lab Helpers: Paul Bergson, Rachel Gaulke, Andrew Gastineau, Andrew Morgan, Anna Flintrop, Ben Jilk, Damien Teichner, Dan Slegh, Dave Klaseus, Joel Petersen‐Gauthier, Max Halverson, Sarah Noe


Questions?


References• American Association of State Highway and Transportation Officials (2010). AASHTO LRFD Bridge Design Specifications, Fifth

Edition, Washington, DC.• American Concrete Institute (1982). “Prediction of creep, shrinkage and temperature effects in concrete structures.” ACI‐209R‐82,

ACI Committee 209, Detroit. • Bažant, Z.P. and Baweja, S. (1995). “Creep and shrinkage prediction model for analysis and design of concrete structures ‐ model

B3,” Materials and Structures, RILEM TC 107‐GCS, Vol. 28, No. 180, pp. 357‐365.• Bažant, Z.P., and Prasannan, S. (1989a). “Solidification theory for concrete creep. I: Formulation,” Journal of Engineering Mechanics,

Vol. 115, No. 8, pp. 1691‐1703.• Bažant, Z.P., and Prasannan, S. (1989b). “Solidification theory for concrete creep. II: Verification and application,” Journal of

Engineering Mechanics, Vol. 115, No. 8, pp. 1704‐1725.• Bažant, Z.P., and Xi, Y. (1995). “Continuous Retardation Spectrum for Solidification Theory of Concrete Creep,” Journal of

Engineering Mechanics, Vol. 121, No. 2, pp. 281‐288.• Comité Euro‐International du Béton (CEB) and the Fédération International de la Précontrainte (FIP) (1978). CEB‐FIP model code

1978.• Comité Euro‐International du Béton (CEB) and the Fédération International de la Précontrainte (FIP) (1990). CEB‐FIP model code

1990.• Gardner, N.J., and Lockman, M.J. (2001). “Design provisions for drying shrinkage and creep of normal‐strength concrete,” ACI

Materials Journal, Vol. 98, No. 2, pp. 159‐167. • Grasley, Z.C., and Lange, D.A. (2007). “Thermal dilation and internal relative humidity of hardened cement paste,” Materials and

Structures, Vol. 40, No. 3, pp. 311‐317.• Hedegaard, B.D., Shield, C.K., and French, C.E.W. (2013). “Finite element modeling of the viscoelastic behavior of reinforced and

prestressed concrete,” Journal of Structural Engineering.• Hedegaard, B.D., French, C.E.W., and Shield, C.K. (2013). “Investigation of thermal gradient effects in the I‐35W St. Anthony Falls

Bridge,” Journal of Bridge Engineering, Vol. 18, No. 9, pp. 890‐900.• Priestley, M.J.N. (1978). “Design of concrete bridges for temperature gradients,” ACI Journal, Vol. 75, No. 5, pp. 209‐217.• Sohn, H. (2007). “Effects of environmental and operational variability on structural health monitoring,” Philosophical Transactions

of the Royal Society A, Vol. 365, pp. 539‐560.

1/30/2018

41

Final Thoughts

1/30/2018 166

Design Quality

• Incorporated peer review of major bridge designs

• Required quality management plans for all consultant contracts

• Modified quality processes within Bridge Office

1/30/2018 167

Transportation Funding

• Previous Minnesota legislative proposals in 2006 & 2007

• 2008 transportation bill & veto override• License plate fee increase

• 5 cent gas tax Increase

• $1.8 billion in bonding of which $600M for bridges

• Up to 3.5 cent added gas tax for bond payments

• By June 2018, replace or rehab all bridges identified in the bill.

1/30/2018 168

Chapter 152 Funds

• 172 bridges were identified• Fracture critical

• Structurally deficient

• Approximately 137 bridges will be replaced or rehabbed

• 123 completed

• Several notable large, important bridges were replaced or rehabbed

1/30/2018 169

1/30/2018

42

Dresbach1701/30/2018

Lafayette1711/30/2018

Winona1721/30/2018

Winona1731/30/2018

1/30/2018

43

Stillwater Lift Bridge (St. Croix)1741/30/2018

St. Croix Crossing1751/30/2018

Red Wing1761/30/2018

Red Wing1771/30/2018

1/30/2018

44

DeSoto Bridge (St. Cloud)1781/30/2018

Granite City Bridge (St. Cloud)1791/30/2018

Kennedy1801/30/2018

Kennedy1811/30/2018

1/30/2018

45

Baudette1821/30/2018

Baudette1831/30/2018

A few final thoughts …

1/30/2018 184

Media

• Daily news conferences are manageable; individual interviews are overwhelming.

• When conference is over, address further questions the next day.

• Be factual and calm; avoid speculation.

• Use non‐engineering terms when possible.

• Keep accusatory questions in perspective – lives have been lost.

1/30/2018 185

1/30/2018

46

Employee Wellness

• Our employees responded on August 1‐2, returning without being called. They have tremendous strength.

• Stress debriefings were valuable tools to share emotions.

• Watch for signs of employee struggles with mental and physical health. Reassign to other duties as needed.

• Protect employees’ privacy; others won’t hesitate to call them at home.

• Communicate with employees as often as possible, rather than having the news be their source.

1/30/2018 186

Other Observations

• Existing partnerships/relationships with the FHWA, City of Minneapolis, and other agencies were key in responding to the tragedy.

• Dedicating a team solely to rapid replacement was needed. Others dealt with collapse.

• Understand politics will be part of it.

1/30/2018 187

Within tragedy is also the impetus to review processes

and improve.

Be open to the opportunity.

1/30/2018 188

Thank you!

Catherine [email protected]

1/30/2018 189

Ed [email protected]

Kevin [email protected]

35w and 10 years of change - umn ccaps · 1/30/2018 1 35w and 10 years of change ed lutgen| bridge...

Documents