Reliability Analysis of River Embankments --using analytical methods and finite elements-Dipl.-Ing. Axel Moellmann, Prof. Dr.-Ing.

Pieter A. Vermeer1 Prof. Dr.-Ing. habil. Bernhard Westrich 2

1

University of Stuttgart, Germany, Institute of Geotechnical Engineering 2 University of Stuttgart, Germany, Institute of Hydraulic Engineering

4th International Symposium on Flood Defence, 6-8 May, 2008, Toronto, Canada
Presentation within the framework of the RIMAX-Project PCRiver Reliability and risk analysis in river flood protection under consideration of geotechnical, hydrological and hydraulic factors
Funding Project Management Coordination

1/16

Benefits of reliability analysis

Aim of a reliability analysis in river flood protection Systematic determination of flood risk as cost-benefit-analysis Risk = Failure probability x Consequence Not: This is a potential weak spot! But: Those are the sections to start improving the flood protection.
2/16

And: Those are the most cost-efficient measures.

Outline
Introduction into reliability analysis Case study Elbe river Probabilistic Finite-Element Analysis of embankment stability Conclusions and Outlook

3/16

Introduction into reliability analysis

Limit state equation:

Z = R S

R: Resistance, S: Stress

Probability density

f (Stress: S)

f (Resistance: R)

S SS

RR R Stress x, Resistance x

Failure probability p(Z<0) Reliability index for Gaussian (normally) distributed variables and Z = R - S:

4/16

R S 2 + 2 R S

Z Z

Case study Elbe river

Torgau / Saxony

5/16

Case study Elbe river

Comparison to a 100-year flood only considering overflow (Dike stretch B)
120

99
100

100-year flood
94 87 88

Return period T [a]

80 60 40 20 0

6/16

Section 1

Validierung HQ100 Section 2 Section 3

Section 4

Case study Elbe river

Computed failure probabilities for dike stretch B
>>5000 1709 3759 >>5000 >>5000

1000 800 600 400 200 0 Overflow / Overtopping

7/16
463 448 331 302 476 461 448 242 238 725

Return period T [a]

Uplift / Piping

Slope instability

Damage of the revetment

Section 1

Section 2

Section 3

Combined failure probability Section 4

Case study Elbe river

Comparison to Elbe flood 2002 - dike failure statistics (Horlacher, 2005) Slope instability
47 % Overflow / Overtopping

Overflow / Overtopping

Uplift / Piping

24 % Uplift / Piping

Failure mode Overflow / Overtopping Uplift / Piping

Dike failure Relative failure statistics probability (PC-Ring) 47 % 24 % 29 % 66 % 28 % 6%

8/16

29 % Slope instability

Slope instability

Case study Elbe river

Reliability water level and reliability freeboard
Reliability water level Reliability freeboard Design flood water level
100 97

Dike body

Freeboard for HQ100 / Reliability freeboard [cm]

100 80 60 40 20 0

94 91

83 46

77

53

Section 1

Section 2

Section 3

Section 4

9/16

Freeboard for a HQ100

Reliability freeboard

Probabilistic Finite-Element Analysis of embankment stability

1.70 m h N 1 5.99 4.06 1 2.83 m

1.00 m

Total incremental displacements

1.44 m

HA = w zA q = 0? Benefits:
10/16

HB = w zB

Stability reserves due to transient seepage effects can be quantified. Zoned dike structure can be taken into account.

Probabilistic Finite-Element Analysis of embankment stability

Stochastic input parameters
p.d.f. 0.2 0.15 0.1 0.05 15

Friction angle 0.2 normally 0.15 distributed

= 17.1 = 1.76 phi @D

p.d.f.

Effective cohesion c lognormally distributed

c = 4.65 kN/m c = 3.72 kN/m

0.1

0.05

18

21

2 4c6 8 10 12 14

p.d.f. 2 107 1.5 107 1 107 5 106

11/16

Permeability k Correlated river water level h lognormally distributed and duration of the flood wave N
k = 5 10-5 m/s k = 2.5 10-5 m/s h 3 Scenarios: hmax,1 = 2.83 m hmax,2 = 2.40 m hmax,3 = 1.20 m

c @kNmD

k = 5. 10-8

k @msD

Probabilistic Finite-Element Analysis of embankment stability

Phase shift between maximum water level and minimum factor of safety

Factor of safety ? needs to be checked for various time steps

for various flow patterns
3.50 Water level above landward dike toe [m] 3.00 2.50 2.00 1.50 1.00 0.50 0.00 0 2 4 6 Time [Days] Water level h Factor of safety

hmax

1.25 1.15 1.10 1.05 1.00 0.95 0.90

8 10 12

?crit

12/16

Factor of safety ?

1.20

Probabilistic Finite-Element Analysis of embankment stability

First Order Reliability Method with Adaptive Response Surface (FORM-ARS)
Perform numerical simulations around mean value Output: Factor of safety ? from a numerical stability analysis Find best-fit Response Surface for ? = b0 + b 1 f + b2 c+ b3 k Find design point for Response Surface Check design point with numerical results IF ( ? 1 ) AND IF ( New design point = Old design point ) then
13/16

Perform numerical simulations around the design point else

Determination of failure probability

Probabilistic Finite-Element Analysis of embankment stability

Response Surfaces for three different maximum water levels Transformation into standard-normalized variables: hmax,3 = 1.20 m hmax,1 = 2.83 m
Permeability u3

X x ui = x

hmax,2 = 2.40 m 3 4.01 2 3.41 1 3.18

Frictio n an gle u1
C u ion s o he 2

Numerical results
14/16

240 Finite-Element calculations Return period of failure: 40,000 years

Conclusions and Outlook

Reliability analysis as basis for a reliable flood risk management Comparable tendency with dike failure statistics during the flood in 2002 Integration of a probabilistic FE-analysis for slope instabilities which regards zoned dikes and transient seepage effects Provision of a tool for risk-based river flood protection Accompanying paper at ISFD4 2008 by Merkel and Westrich
15/16

Thank you for your attention! Questions?

Dipl.-Ing. A. Moellmann Institute of Geotechnical Engineering, University of Stuttgart Pfaffenwaldring 35, 70569 Stuttgart Tel. ++49/ (0)711 / 685 63779, Email: axel.moellmann@igs.uni-stuttgart.de Financial support: Forschungszentrum Karlsruhe in der Helmholtz-Gemeinschaft
Projekttrger Forschungszentrum Karlsruhe

Cooperation partners:
Dam Authority of Saxony, Pirna Regional administrative authority Tbingen Rijkswaterstaat, Dienst Weg-en Waterbouwkunde, Delft, The Netherlands

16/16

4th International Symposium on Flood Defence 6-8 May, 2008, Toronto, Canada