Time-Dependent Reliability Modeling for Use in Major ... · Foundations Karst Used Expert-Opinion...
Transcript of Time-Dependent Reliability Modeling for Use in Major ... · Foundations Karst Used Expert-Opinion...
Time-Dependent Reliability Modelingfor Use in Major Rehabilitation of
Embankment Dams and Foundation
Robert C. PatevNAD Regional Technical Specialist
Navigation/Risk and Reliability
Outline
� Time-Dependent Reliability� Hazard Functions� Wolf Creek Major Rehabilitation Report
(MRR)� Time-Dependent Reliability Modeling� Expert-Opinion Elicitation
Reliability� Probability of unsatisfactory performance
(PUP)� Limit state defined before failure occurs� Problem – snapshot in time� Not cumulative (does not account for previous
loadings)� Must account for degradation of structures
� Mechanisms� Corrosion� Fatigue� Freeze-Thaw� Wear� Abrasion/Erosion
Time-Dependent Reliability� Geotechnical Aspects
� Difficult task-at-hand� Foundations
� Karst� Used Expert-Opinion Elicitation to define PUPs� Snapshot
� Alluvial� Used Taylor Series� Snapshot
� Degradation� Data/Rates� Models
Time-Dependent Reliability
f(R)
f(D)
Time, t
R , D L(t) = f(R)*g(t) f(D)
h(t) = - ln d/dt (L(t))
Time-Dependent Reliability
� Hazard Function� Developed by actuaries in 1880’s� Used by aerospace industry in early 1950’s� Conditional probability
� h(t) = PUP [t + dt, t]� PUP in time, t+ dt, given you have survived up to
time, t� Based on efforts on Ohio River Mainstem Study
ttime,toupsurvivedofNumber1ttime,inesperformanctoryunsatisfacofNumberh(t) +=
Hazard Function
Time, t
Burn-in
h(t) dec.
h(t)
Burn-out
h(t) inc.
Life phase
h(t) constant to inc.
Wolf Creek Dam Project
� Concrete 240’ ht 1836’ length� Earth Emb 200’ ht 3900’ length
Wolf Creek Dam - History� Designed in late 1930’s� Construction began 1941� Completion delayed until 1951 due to WWII� 1967-68 Sinkholes near Switchyard + right
d/s abutment wet areas D/S of embankment,muddy flows
� 1968-72 Emergency exploration / grouting� 1975-79 Diaphragm walls construction
(ICOS)
Wolf Creek Dam – Typical Section
0 1+00A 2+00A 3+00A 4+00A 5+00A1+00B2+00B3+00B4+00B5+00B5+00B
EL. 660
EL. 690
EL. 730
EL. 763EL. 773
EL. 730
EL. 690
EL. 665
EL. 640
Flow
Axis of Dam
C.L. Roadway & Embankment Sta. 0+16B
Foundation of Toe Drain
End of Drainage Blanket
Core trench
Drainage
Blanket
Bedrock
Emb
Qal
Wolf Creek Dam – 1967/1968Events
Wet Area
No. 1
Wet Area
No. 2
Wet Area
No. 5
Wet Area
No. 3
Wet Area
No. 6
Sinkhole 3’ Dia.
15 August 1967
Sta. 56+75L,4+36B
Wet Area No. 4
Sinkhole No. 1
13 March 1968
Sta. 32+16L, 4+13B
13’ Dia. & 10’ Deep
Sinkhole No. 2
22 April 1968
Sta. 32+58L, 4+13B
6’ Dia. & 6’ Deep
Muddy
Water
Downward Slope of
Cut Off Trench
Cut Off Trench
Toe of Random Fill
Wolf Creek Dam – Piezometers
0 1+00A 2+00A 3+00A 4+00A 5+00A1+00B2+00B3+00B4+00B5+00B5+00B
EL. 660
EL. 690
EL. 730
EL. 763EL. 773
EL. 730
EL. 690
EL. 665
EL. 640
Flow
Axis of Dam
C.L. Roadway & Embankment Sta. 0+16B
Foundation of Toe Drain
End of Drainage Blanket
Diaphragm Wall
PZ No. DC-248R
Sta. 35+49L 0+06A
PZ No. D-323
Sta. 35+55L 0+35B
March 03March 03
Wolf Creek Dam – CutoffTrench
Wolf Creek Dam – CutoffTrench
Wolf Creek Dam – DiaphragmWall, 1975-1979
Diaphragm Wall in Emb Concrete
Limestone
Wolf Creek Dam – DistressIndicators
C o m p aris o n of D is tres s Ind ic ato rs 1 96 8-20 04 at W o lf C re ek P ro jec t
3
8
0 .2 0 0 0 0
6
1
37
4 .2
2
64
3
7
0
5
10
15
20
25
30
35
40
N o. o f S ink s N o . o f W etA reas
% of W et A reasD /S
S eepageIns tab ility o fR ive r B ank
Z ones , N o . o fLoca tions
N o . o f A rtes ianP Z 's
C racks per 100F ee t a t R oad
S urface (100 '=1 )S ta . 35+11-
39+00
E m bankm entS ettlem en t
(0 .1 '=1 ) M axS e ttlem en t atS ta tion 37+00
P Z H ead(10% =1)
H ighest H eadsBe tw een 35+11-
40+00Q u alita tiv e D is tress In d icato r
* T em pera ture m easurem ents from P Z s in 1968 and 2004 show anom a lis tic co ld a reas dow ns tream of dam axis .
Dis
tres
s
19682004
Wolf Creek Dam – CrestSettlement
W olf C re ek S ettlem en t R ate s
0 .00
0 .05
0 .10
0 .15
0 .20
0 .25
0 .30
0 .35
0 .40
0 .45
0 .50
J-81 J-82 J -83 J -84 J-85 J -86 J -87 J-88 J-89 J -90 J -91 J-92 J -93 J -94 J-95 J-96 J -97 J -98 J-99 J-00 J -01 J -02 J-03 J -04 J -05
D ate
Settl
emen
t (Fe
et)
E M -1E M -2E M -3E M -4
E M -5E M -6E M -7E M -8E M -9E M -1 0E M -1 1E M -1 2E M -1 3E M -1 4E M -1 5E M -1 6E M -1 7E M -1 8E M -1 9E M -2 0E M -2 1E M -2 2E M -2 3E M -2 4E M -2 5E M -2 6
Wolf Creek Dam – PiezometricData - Section 1
WOLF CREEK DAMSelected PZ's (20 Year trend)
540
550
560
570
580
590
600
610
620
630
640
650
660
670
680
690
700
710
720
730
740
750
760
J-84 J-85 J-86 J-87 J-88 J-89 J-90 J-91 J-92 J-93 J-94 J-95 J-96 J-97 J-98 J-99 J-00 J-01 J-02 J-03 J-04
DATE
PZ
Ele
vatio
ns in
Fee
t (M
SL)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Rai
nfal
l in
Inch
es
D-321D-322D-323HeadwaterTailwaterRainfallLinear (D-322)
Wolf Creek Dam – PiezometricData -Section 2
WOLF CREEK DAMSelected PZ's (20 Year trend)
540
550
560
570
580
590
600
610
620
630
640
650
660
670
680
690
700
710
720
730
740
750
760
J-84 J-85 J-86 J-87 J-88 J-89 J-90 J-91 J-92 J-93 J-94 J-95 J-96 J-97 J-98 J-99 J-00 J-01 J-02 J-03 J-04
DATE
PZ
Ele
vatio
ns in
Fee
t (M
SL)
DC-258DC-254RWA-25HeadwaterTailwaterWA-49Linear (WA-49)
Wolf Creek Dam – PiezometricData - Section 3
WOLF CREEK DAMSelected PZ's (20 Year trend)
540
550
560
570
580
590
600
610
620
630
640
650
660
670
680
690
700
710
720
730
740
750
760
J-84 J-85 J-86 J-87 J-88 J-89 J-90 J-91 J-92 J-93 J-94 J-95 J-96 J-97 J-98 J-99 J-00 J-01 J-02 J-03 J-04
DATE
PZ E
leva
tions
in F
eet (
MSL
)
D-275AWA-99FWO-35HeadwaterTailwaterLinear (WA-99F)
Wolf Creek Dam – Time-Dependent Reliability Concepts
PiezometricHead,Flow, Surface
Water
DefinedLimitState
Time1951 1968 1979 2004
Groutwall
Diaphragmwall
1984 2060
0
?
PiezometricHead,Flow, Surface
Water
DefinedLimitState
Time1951 1968 1979 2004
Groutwall
Diaphragmwall
1984 2060
0
?
Time1951 1968 1979 2004
Groutwall
Diaphragmwall
1984 2060
0
?
Wolf Creek Dam MRR
� Time-Dependent Reliability Model� Utilize permanent rise in piezometric
pressures in the foundations� Reliable and most consistent data� Model three different embankment sections� Assume that selected limit state occurs before
any decrease or fall in piezometric pressurewhich would indicate a more critical situation
Wolf Creek Dam MRR
� Time-Dependent Reliability Model� Damage Accumulation Model (DAM)
� Used in fatigue and wear rate analyses� Model cyclic variations of headwater above EL
710
El 710
AnnualDuration
Annual Intensity
Cyclic Headwater
El 710
AnnualDuration
Annual Intensity
Cyclic Headwater
Wolf Creek Dam MRR
� Time-Dependent Reliability Model� DAM Calibration
Annual and Accumulated Damages Versus Time
0.000
2.000
4.000
6.000
8.000
10.000
12.000
1950 1960 1970 1980 1990 2000 2010
Time
Piez
omet
ric R
ise
(in fe
et)
Annual DamagesAccumulated DamagesActual Piezometric Rise
Wolf Creek Dam MRR� Time-Dependent Reliability Model
� Developed Monte Carlo Simulation Model� Accounted for entire life cycle including past remedial
repairs� Performed 50,000 iterations for reliability calculations
� Random Variables� Annual Intensity –
� Based on historical records from 1950 to 2004� Truncated lognormal distribution
� Annual Duration –� Based on historical records from 1950 to 2004� Truncated lognormal distribution
Wolf Creek Dam MRR� Time-Dependent Reliability
� Random Variables (cont’)� Spatial Variability Factor-
� Accounts for variation in piezometric pressures in thefoundation
� Based on a wide range of piezometer data for each section� Incorporated as quadratic modifier to annual damage
accumulation� Uniform distribution
� More uncertainty in Section 3 than Sections 1 or 2 dueto continued wet spots downstream and lack ofdiaphragm wall
Wolf Creek Dam MRR
� Time-Dependent Reliability Model� Limit state
� Defined limit state for unsatisfactoryperformance on piezometric rise in 3 sections
� Used Expert-Opinion Elicitation to quantifythose values
Section Average PZValue in 2004
(in feet)
Rise of PZfrom EOE(in feet)
UnsatisfactoryPerformance Limit State
(in feet)
1 9.3 5 14.32 4.1 5 9.13 1.2 3 4.2
Wolf Creek Dam MRRWolf Creek Expert Elicitation
Event Full Description Expert-opinion elicitation SummaryName of Issue Table
First SecondResponse Response
Unsatisfactoryperformance occurs
in Section 1
At what future change inpiezometric pressure in thefoundation would you expectunsatisifactory performance inSection 1?
Median = 5 Median = 5
ConfidenceMinimum = 1
Expert #1 0.5 1 med 25 Percentile = 3Expert #2 5 5 highExpert #3 10 7 med Median = 5Expert #4 3 2 medExpert #5 5 5 med 75 Percentile = 5Expert #6 5 4 high
90 Percentile = 6
High = 7
Minimum = 0.5 1Median = 5 5
Maximum = 10 7
Wolf Creek Dam MRRTime Dependent Reliability for Sections 1, 2 and 3 - Wolf Creek
Dam
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1940 1960 1980 2000 2020 2040 2060 2080
Time (years)
Rel
iabi
lity Section 1
Section 2Section 3
Wolf Creek Dam MRR
Summary of Hazard Rates for Wolf Creek Dam
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1940 1960 1980 2000 2020 2040 2060 2080
Time (years)
Haz
ard
Rat
e Section 1Section 2Section 3
Wolf Creek Dam MRRWolf Creek Dam Event Tree
Baseline Condition - Section 1
Limit State - PUP Annual Hazard Rate Event Event Probability Repair Scenarios/Costs Effect on Hazard Rate(from Reliabilty Model) (from EOE)
Full UP 0.15 Rebuild New Dam AHR changed to 0 for remainder of lif(Dam breach) 5 Years Design
5 Years Construction
AHR Drain Reservior to El 680Time, T 5 ft unacceptable rise in X Partial UP 0.6 Grout of Dam AHR adjusted back to 0.1
piezometric pressure (Settlement, sinkholes, piping, wet spots) 1 Year Construction Degrades using AHR
Increased Surveillance and Monitoring 0.25 O&M Costs, Instrumentation Time T + 1(Reach limit state but no observable damages) and increased monitoring (AHR changes to next year)
1-X O&M Costs Time T + 11-AHR (AHR changes to next year)
Time-Dependent Reliability� Conclusions
� Time-dependent geotechnical models are difficult� Need to think well outside the box (deductive versus
inductive thinking)� Incorporate and define distress indicators
� My require some data processing
� Understand - models are for major rehabilitationpurpose to gain funding to rehab structure
� Don’t get lost in the fine details� Not the true probability of failure
� Within an order of magnitude� Not for dam safety