OBSERVATION METHOD: A NEW TOOL FOR THE ......Scour Rate (mm/hr) Sand D 50 =0.3 mm Scour Rate vs...

Preview:

Citation preview

Jean-Louis BRIAUD

Distinguished Professor Texas A&M University

OBSERVATION METHOD: A NEW TOOL FOR THE BRIDGE

SCOUR ENGINEER

2Jean-Louis BRIAUD

Acknowledgements

• Texas DOT (John Delphia)

• Massachusetts DOT (Rich Murphy, HananFouad)

• Texas A&M University (Hamn-Chin Chen, Francisco Oliveira, Anand Govindasamy, Dekay Kim, Inwoo Jung, Hans Kim)

SCOUR DEPTHS

Normal Water Level

LC

z(abut)

z(abut) Applies z(cont) Applies

Probable Flood Level

z(pier) z(cont)

PROBLEM

1. Comparison between measured and calculated scour depths by current method (HEC-18) exhibits a lot of scatter

2. Comparison between measured and calculated scour depths by current method (HEC-18) shows excessive conservatism on the average

HEC-18 RESULTS

0 2 4 6 8 100

2

4

6

8

10

Measured Landers-Mueller Pier Scour, zmax (m)

Det

erm

inis

tic H

EC-1

8 Sa

nd, z

max

(m)

Zpredicted/Zmeasured = 3.26

LANDERS-MUELLER DATABASEPIER SCOUR

6Jean-Louis BRIAUD

HEC 18 Results–Bridge Case Histories

7Jean-Louis BRIAUD

OMS Results-Bridge Case Histories

8Courtesy of the University of Kentucky at Louisville

9

10

11

12

13

14

Jean-Louis BRIAUD – Texas A&M University

15

EFA - EROSION FUNCTION APPARATUS

Jean-Louis BRIAUD – Texas A&M University

16

Scour Rate vs Shear Stress

0.01

0.10

1.00

10.00

100.00

1000.00

10000.00

0.1 1.0 10.0 100.0

Shear Stress (N/m2)

Scou

r R

ate

(mm

/hr)

Sand D50=0.3 mm

Scour Rate vs Velocity

0.01

0.10

1.00

10.00

100.00

1000.00

10000.00

0.1 1.0 10.0 100.0

Velocity (m/s)

Scou

r R

ate

(mm

/hr)

Sand D50=0.3 mm

EROSION FUNCTION FOR A FINE SAND

Jean-Louis BRIAUD – Texas A&M University

17

Scour Rate vs Shear Stress

0.01

0.10

1.00

10.00

100.00

0.1 1.0 10.0 100.0

Shear Stress (N/m2)

Scou

r R

ate

(mm

/hr)

Porcelain Clay PI=16%Su=23.3 Kpa

Scour Rate vs Velocity

0.01

0.10

1.00

10.00

100.00

0.1 1.0 10.0 100.0

Velocity (m/s)

Scou

r R

ate

(mm

/hr)

Porcelain Clay PI=16%Su=23.3 Kpa

EROSION FUNCTION FOR A LOW PI CLAY

Jean-Louis BRIAUD – Texas A&M University

POCKET ERODOMETERPET test result = Depth of hole in mm after 20 squirts at 8 m/s

18

$0.49 atWalMart

Jean-Louis BRIAUD – Texas A&M University

BOREHOLE EROSION TEST - BET19

Fluid circulation

Borehole caliper

Jean-Louis BRIAUD – Texas A&M University

BOREHOLE EROSION TEST - BET20

21

EROSION CLASSIFICATION

22

EROSION CLASSIFICATION

Jean-Louis BRIAUD – Texas A&M University

25

26Jean-Louis BRIAUD

OBSERVATION METHOD FOR BRIDGE SCOUR

• Step 1: Observe maximum scour depth = Zmo• Step 2: Find out the maximum flood the bridge

has been subjected to : Collect gage data, RI from TAMU-OMS, Qmo/Q100, Vmo/V100

• Step 3: Extrapolate field measurements to predict future scour depth

Zfut / Zmo = F (Vfut / Vmo) • Step 4: Compare future scour depth to

foundation depth (pier) Zfut < Zfound / 2

27Jean-Louis BRIAUD

Step 1: Observe maximum scour depth = Zmo

28Jean-Louis BRIAUD

Step 2: Find out the maximum flood the bridge has been subjected to = Vmo

930 Flow Gages in Texas and neighbor States

29Jean-Louis BRIAUD

Step 2: Find out the maximum flood the bridge has been subjected to = Vmo

Maximum RI map between 1970 and 2005

Automated with TAMU-OMS software

30Jean-Louis BRIAUD

Step 2: Find out the maximum flood the bridge has been subjected to = Vmo

Maximum RI map between 1920 and 2005

Automated with TAMU-OMSsoftware

31Jean-Louis BRIAUD

Step 3: Extrapolates field measurements to predict future depth Zfut

Zfut / Zmo = F (Vfut / Vmo)

•Known = Zmo and Vmo•Choose Vfut•Obtain Zfut from charts•The Z-Future Charts were developed by performing a large number (~350,000) of HEC-18 Clay simulations using varying pier, contraction & abutment scour geometries, varying soil conditions, varying velocities, and varying age of the bridge.

32Jean-Louis BRIAUD

Category III MaterialsUpstream Water Depth (H1): 5 m to 20 mContraction Ratio (Rc) : 0.5 to 0.9Critical Velocity (Vc) : 0.5 m/sPier Diameter (D) : 0.1m to 1.0 mthyd = 25 years

Vfut/Vmo2.00

Z fut

/Zm

o2.8

1.0

Step 3: Extrapolates field measurements to predict future scour depth Zfut/Zmo = Vfut/Vmo

33Jean-Louis BRIAUD

Step 4: Compare future scour depth to foundation depth Zfut < Zfound / 2

34Jean-Louis BRIAUD

• 15 bridges selected (12 scour critical, 3 stable)• 6 scour critical bridges out of the 12 found

stable by the observation method• 3 stable bridges found stable by the

observation method• 6 of 12 bridges originally classified scour

critical were found stable by the observation method

APPLICATION TO SCOUR CRITICAL BRIDGESTexas

35Jean-Louis BRIAUD

36Jean-Louis BRIAUD

37Jean-Louis BRIAUD

38Jean-Louis BRIAUD

39Jean-Louis BRIAUD

40Jean-Louis BRIAUD

Drawbacks

• Requires a good network of flow gages• Cannot be used directly for new bridges• Estimate in filling (USGS and a TxDOT survey

have found that it was rare (10% of the time) and ranged from 2 to 4 ft)

• Not yet developed for layered soil (be careful with thin hard layer over soft layer)

41Jean-Louis BRIAUD

Advantages

• Valuable tool to prioritize bridge repairs, countermeasure decisions, asset management

• Can serve as an input to FHWA risk approach• Part of the practical design concept• No need for erosion testing• Actual soil• Actual flow history• Actual geometry • Based on observed measurements

42

DIFFERENCES BETWEEN HEC 18AND OBSERVATION METHOD

HEC 18

• Flume tests (scale pb?)

• Wrong worst soil (Fine sand)

• Simplified geometry

• Simplified single velocity

OMS

• Full scale

• Right soil

• Exact geometry

• Exact velocity hydrograph

43

COMPARISON

HEC 18 OMS

44

The Million Dollar Stickwith TAMU-OMS

Freecourtesy of TxDOT, MassDOT and TAMU

Recommended