Fluid Structure Interaction Effects
Fluid Structure Interaction Effects
TABLE OF CONTENTS
1
Simplified methods
Numerical methods
Applications
It covers:
It covers:
State-of-the-art review
PHYSICAL PHENOMENON
Sloshing means any motion of the free liquid surface inside its container
PHYSICAL PHENOMENON
Motion of the fluid can be caused by different dynamic excitations:
Earthquake
The base of the container is subjected to ground or floor seismic motion
Transportation
Tanker ships, aircraft & spacecraft fuel tanks, road tankers
PHYSICAL PHENOMENON
Motion of the fluid can be caused by different dynamic excitations:
Earthquake
PHYSICAL PHENOMENON
Why to worry about sloshing in isolated NPPs?
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Boundary conditions
Vertical velocity of liquid along tank base is equal to ground velocity
Radial velocities of liquid and tank wall are the same
Free surface at top
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Modal masses
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Expressions by Graham & Housner are widely used in seismic design of tanks:
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Sloshing frequencies (Hz):
fi =
1
2
g
h
tanh i
R
R
tank radius
height of fluid
1 = 1.841
2 = 5.331
3 = 8.536
4 = 11.706
(vary with tank dimensions and are generally independent of liquid height, except for
very shallow depths)
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Sloshing pressures along height (first 2 modes) for various = h/R ratios
Maximum convective pressure:
at z = h
r =R
SIMPLIFIED METHODS
Assumption of a rigid container (Housner, 1957)
Maximum vertical displacement of fluid surface:
d max = 0.837 R Sa
where Sa is the spectral acceleration at the fundamental sloshing
frequency and damping ratio (about 0.5%) expressed in g
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
Co = (R/h) Co
(see figure in next viewgraph)
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
Top of fluid
Vertical distribution of
impulsive hydrodynamic
maximum pressures.
Squat tank
Tank bottom
Slender tank
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
Cl 1
fi =
2 h
Cl = CW
W
l
where:
h
R
CW
s
W
l
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
CW
(Haroun & Housner, 1981)
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
1
fi =
2 R g ( )
E s1 / 3
h
where:
E
s1/3
h
R
g()
(EC-8)
SIMPLIFIED METHODS
Flexible containers (ASCE, 1987)
The main difference between the response of a rigid and flexible tank
has to do with the nature of the acceleration (impulsive) component:
In a rigid tank, the response is proportional to the maximum ground
acceleration.
In a flexible tank, the response is governed by the spectral acceleration
corresponding to the fundamental frequency of the tank-liquid vibration
and associated damping ratio (about 2%).
NUMERICAL METHODS
Only way to deal with complicated geometries and excitations
Highly non-linear
Numerical techniques should be very robust.
Requires large computer resources
NUMERICAL METHODS
Solution of Navier-Stokes equations in the time domain combined with
methods for free surface tracking
KASHIWAZAKI-KARIWA NPP
SFP overflow during 2007 Niigataken Chuetsu-oki Earthquake
CRIEPRI (Japan)
Own VOF-3D code (non-linear
wave behaviour).
3 axis motion from seismic data.
Similar overflows and wave
shape as recorded during EQ.
ELSY
Study of sloshing of liquid lead at ELSY reactor vessel
NRG (Netherlands)
OpenFOAM code (VOF)
Simple harmonic horizontal
velocity motion from seismic data
FRS at 3 first natural frequencies
Shape and load assessment
IDOM (Spain)
DualSPHysics (SPH)
Motion derived from Floor Response
Spectra
2D and 3D simulations
With and without vertical motion
Target: assess loss of water in the pool
TRILLO NPP - 1
Stress tests. Assessment of Essential Services Pond overflows
IDOM (Spain)
DualSPHysics (SPH)
Motion derived from Ground Response
Spectra
2D simulations
Without vertical motion (low influence)
Target: assess loss of water in the pond
TRILLO NPP - 1
Stress tests. Assessment of Essential Services Pond overflows
No overflow expected
Induced vortex near lateral embankments
Attractive features
Lagrangian character / Interaction with structures is made easier
Exact treatment of convection (mass conservation)
Precise definition of interphases
Easy introduction of complex Physics
Present applications
Multiphase flows / Viscous flows
Fluid-structure interaction
High-speed impact problems
Landslide generated waves
kernel function
particle property
As (r ) = mi
i
smoothed
property at
position r
Ai
W (r ri , h )
particle
mass and
density
smoothing
length
Hydrodynamic pressures
Directly proportional to density
Molten
lead
Density (kg/m3)
1000
10470
1500
1740
0,466
0,434
1,002
1,882
4.7 10-4
0.4 10-4
Sloshing.avi
Experimental facilities
Experimental facilities
Non-Coupled SPH
Sloshing3.avi
Sloshing4.avi