DAM OVERALL RISK ANALYSIS IN PARTICULAR TO OVERTOPPING

1
DAM OVERALL RISK ANALYSIS- IN PARTICULAR TO
OVERTOPPING

• ???
• ????????????

2
Outlines

• Background
• Dam overtopping evaluation
• Frequency analysis
• Flood wind
• Uncertainty analysis
• IS-LHS method
• Reservoir routing wind setup and run-up
• Risk analysis
• Case study- Shihmen Dam

3
Background

• Dam is defined
• as a barrier of earth, masonry, etc.
• Dam failures have claimed many lives in many
  countries since the first dam was built 7,500
  years ago in Mesopotamia.

4

• ICOLD ( 1998) estimated

• there were more than 45,000 large dams by the end
  of the last century.

• ICOLD (1973)

• overtopping causes about 35 of all earth dam
  failures

• Cheng (1993) investigated

• the average overall failure probability is around
  10-3 per dam per year for damsgt 15 m in height .

5

• These data are insufficient for a statistical
  analysis of dam failure because

• the sample size is too small to make a
  statistically reliable conclusion for
  low-probability but high-consequence events

• the dam failure sample may not be representative

• the physical conditions of the dams are
  time-variant.

6
agricultural, industrial, and domestic water
supply hydropower generation flood control.

• SHIHMEN DAM

To allow for higher flood recurrences, engineers
constructed additional two tunnel spillways in
1985. The capacity of water release facilities
has increased from 10,000 m3/s to 12,400 m3/s.
7
(No Transcript)
8
Dam overtopping evaluation
9
(No Transcript)
10
Frequency analysis

• Chow et al., 1988
• The primary object of frequency analysis is to
  relate the magnitude of extreme events to their
  frequency of occurrence through the use of
  probability distributions

11
Flood

• In previous studies, the flood frequency analysis
  has emphasized
• the annual maximum (AM) model
• without much regard to the monthly maximum (MM)
  model.

12

• There are two types of error associated with
  flood frequency analysis (quantile estimation).

• The first type arises from the assumption that
  the observations follow a particular distribution

• the second type is the error inherent in
  parameter estimates from small samples

13

• These annual and monthly maximum three-day
  average flow rates are subjected to the
  Anderson-Darling goodness-of-fit test.

• Normal
• Lognormal
• Pearson type III
• LogPearson type III
• Weibull
• Gumbel

• Normal
• Lognormal
• Pearson type III
• LogPearson type III
• Weibull
• Gumbel

14

• Annual maximum (AM) model
• Lognormal
• LogPearson type III
• Gumbel
• Monthly maximum (MM) model
• Lognormal
• LogPearson type III
• Gumbel

15
Return periods of different frequency models
16
Return periods of different frequency models
17
Wind

• The distributional properties of wind speed are
  then adopted from Juang (2001).

18
(No Transcript)
19
Uncertainty analysis

• The main purpose of uncertainty analysis is to
  quantify system outputs or responses as affected
  by the stochastic basic parameters in the system.
  

20

• Analytic methods
• Fourier, Laplace, exponential ,and Mellin
  transforms
• Approximation methods
• FOSM, AFOSM, PEMs
• Sampling methods
• Monte Carlo sampling
• Latin hypercube sampling
• Importance sampling

21
Concept of sampling methods
various times
22
IS-LHS method
p
23
p
p5
p4
p3
p2
p1
q1
q5
0
.
w1
w5
.
24
IS-LHS method
p
q,
w
25
(No Transcript)
26
(No Transcript)
27
Sample from this area
28
(No Transcript)
29
Reservoir routing wind setup and runup
I-O dS/dt
Overtopping induced by flood and wind
Overtopping induced only by flood
Overtopping induced only by wind
30
(No Transcript)
31
Risk analysis

• The failure of an engineering system can be
  defined as the loading on the system (L)
  exceeding the resistance of the system (R).

• The failure of an engineering system can be
  defined as the loading on the system (L)
  exceeding the resistance of the system (R).
• Risk can be represented as

32
Case Study
Overtopping caused by flood and wind without
considering flow rate uncertainty
Overtopping caused by flood and wind considering
flow rate uncertainty
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
Conclusions

• The proposed sampling method

• Wind effect

• Model distribution

• Frequency model

37
Thanks for your attention

• Questions?