Windblown Sand Modelling and Mitigation
Transcript of Windblown Sand Modelling and Mitigation
![Page 1: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/1.jpg)
111
Windblown Sand Modelling and Mitigation
January 28th, 2021
von Karman Institute for Fluid Dynamics
Environmental and Applied Fluid Dynamics department
Lorenzo RAFFAELE
MAFD - TCS
![Page 2: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/2.jpg)
2
Industrial Motivation: coastal zones
+44%
+96%
Windstorm frequency
Windstorm intensity
US
EU
![Page 3: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/3.jpg)
3
Industrial Motivation: desert regions
Structure scale Urban scale Infrastructure scale
![Page 4: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/4.jpg)
4
0
4
3
0
1
0
0
6
5
0
6
5
0
9
8
0
9
8
0
9
8
0
9
8
Industrial Motivation: railway megaprojects
B$
Market potential
β’
Railway length
Kmβ’
arid regions
northern desert belt
0
100
200
300
400
2010 2012 2014 2016 2018
B$
year
Arab League Network
Gulf Cooperation Council Network
Iron Silk Road
Railway megaprojects
![Page 5: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/5.jpg)
5
Sand π β 0.063,2 mm
Dust π < 0.063 mm β’ long-term suspension
β’ short-term suspension
Engineering interest
β’ creep
β’ saltation
π = ΰΆ±0
+β
π(π§) ππ§Sand transport rate
[ππ πβ1π β1]
WbS saltation condition
π π§ > 0 π > ππ‘
π’β > π’βπ‘π’β = π/ππ
Threshold shear velocity
Saltation
π₯
π§
Phenomenology
![Page 6: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/6.jpg)
6
Deterministic models
Microscopic models
β’ Equilibrium of the
momentsEntraining aerodynamic forces VSStabilizing forces
Macroscopic models
β’ Semi-empirical
(free parameters)
β’ Trend VS d
πΏ
π·
πΌ
πΊ
π
Probabilistic models
β’ Scatter of experimental data
β’ Random turbulent wind flow, bed
grain geometry, interparticle
forces Zimon (1982)
Duan et al. (2013)
4 microscopic r.v.s
Modelling and technical
difficulties
π’βπ‘ = π¨ππ β ππ
ππππ
Bagnold (1941) dust sand
Modelling: π’βπ‘
![Page 7: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/7.jpg)
7
Modelling: πSemi-empirical models
Bagnold type
π~π’β3
Modified Bagnold type
π~π’β,πππ3 π’β, π’βπ‘
Oβ Brien-Rindlaub type
π~π’
Complex
Dong et al. (2003)
Large discrepancies due to:
β’ large randomness of physical
phenomenon
β’ debated scaling
π = 2.78ππππ’β3 1 β
π’βπ‘2
π’β2
1 +π’βπ‘π’β
π = 0.25 +ππ 3π’β
ππππ’β3 1 β
π’βπ‘2
π’β2
π = 6.7π
ππ
ππππ’β3 1 β
π’βπ‘π’β
π = 5ππππ’βπ‘π’β
2 1 βπ’βπ‘2
π’β2
Kawamura (1951)
Lettau & Lettau (1951)
Owen (1964)
Kok et al. (2012)
![Page 8: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/8.jpg)
8
upwind strip downwind strip
source windblown sand path receiver
Source SMM
β’ Layer system
gravel surfacestraw checkerboard array line-like obstacles
asphalt-latex mixture
β’ Hedge system
natural crusting
π β π’β,πππ β π’βπ β π’βπ‘
π
Hedge system Layer system
Sand Mitigation Measures: Source
Path SMMReceiver SMM
![Page 9: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/9.jpg)
9
upwind strip downwind strip
source windblown sand path receiver
Path SMM
Sand Mitigation Measures: Path
Straight Vertical Wall (SVW)
β’ Surface-like
β’ Volume-like
porous fence Shield for Sand (S4S)
dyke ditch
WO 2016/181417 A1
porous solid
![Page 10: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/10.jpg)
10
upwind strip downwind strip
source windblown sand path receiver
Receiver SMM
Sand Mitigation Measures: Receiver
β’ Aerodynamic based
β’ Sand-resistantCN/102002916
Humped sleepersUS4958806A
T-Track system Continuous slab Lubricant free turnout
Jet roofs Venturi effect-based
![Page 11: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/11.jpg)
11
an uncharted territory for modern engineeringβ¦
![Page 12: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/12.jpg)
12
Windblown Sand Action: categorization
Environmental
β’ site-dependent
β’ inborn
randomness
probabilistic
modelling
~ wind
an
alo
go
us
to
Free
β’ wind-
dependent
accumulation
~ windblown snow
Wind and sand
modelling
Variable
β’ long-term varying
accumulation
process
~ snow
β’ non monotonic
(periodic sand
removal)
~/β snow
Time-variant
reliability analysis
Evaluation of sand
removal period
2 months laterβ¦
mo
de
llin
g
fallo
ut
Win
db
low
n
Sa
nd
![Page 13: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/13.jpg)
13
Windblown Sand Action: modelling chainSite
an
aly
sis
Incoming Wind
Aerodynamics
Wind action
Ass
ess
me
nt
Incoming Windblown Sand
Aerodynamics / Morphodynamics
Windblown sand action
β’ π10
β’ πΉβ’ π πΉ β¦
β’ πππ(π10, π)
πππ
π
ππππ
Windblown sand action VS Wind action
π10
ππΉ
π10 πΉ
β’ πΆπ π, Ξ0, π
aerodyn. morphodyn.
β’ πΆπ π, Ξ0
1
2
3
![Page 14: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/14.jpg)
14
Incoming Windblown Sand
Uncertainty
Aleatory
Epistemic
β’ Sand uncertainties: grain size,
shape, relative position, surface
cleanliness, grain size
distribution.
β’ Wind uncertainties: turbulent flow inborn variability, uncontrolled
environmental conditions, e.g. temperature, humidity.
β’ Model uncertainty: simplified representation of the real physical
behaviour, identification of relevant variables, hypothesis, interactions
left out. Lack of a shared definition of π’βπ‘ Shao (2008)
β’ Measurement uncertainty: errors and/or different procedures.
β’ Parameter uncertainty: values of model parameters.
βLack of exact knowledge, regardless of what is
the cause of this deficiencyβ Refsgaard et al. (2007)
Statistical approach β’ Nonlinear regression
β’ Copula-based regression
πππ π’βπ‘ , ππ , π’β
![Page 15: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/15.jpg)
15
Incoming Windblown Sand: probabilistic π’βπ‘
β’ Collected studies range from 1937 to 2004.
β’ #=133 #=109 from in L. Raffaele, L. Bruno, F. Pellerey, L. Preziosi (2016), Aeolian Res.
![Page 16: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/16.jpg)
16
π 2 β 0.75
π’βπ‘ = π¨πππ β ππππ
ππ
π’βπ‘ = π¨πππ β ππππ
ππ +πΈ
πππ
Bagnold (1941)
Shao & Lu (2000)
fine medium coarse
Incoming Windblown Sand: probabilistic π’βπ‘
![Page 17: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/17.jpg)
17
Incoming Windblown Sand: probabilistic π’βπ‘
πΉ(π), πΉ(π’βπ‘) , π, π’βπ‘ β β
Fitting of marginal distributions
πΉ π, π’βπ‘ = πΆ πΉ(π), πΉ(π’βπ‘) πΆ: 0,1 2 β 0,1
πΆ π’, π£ = π’ + π£ β 1 + 1 β π’ β1/πΌ + 1 β π£ β1/πΌ β 1βπΌ
π’, π£ β 0,1 , πΌ > 0
Fitting of Inverted Clayton Copula
π, π’βπ‘ = πΉβ1 π’ , πΉβ1 π£
From copula to original scale
From original to copula scale1 2
34
![Page 18: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/18.jpg)
18
Incoming Windblown Sand: probabilistic π’βπ‘
from in L. Raffaele, L. Bruno, F. Pellerey, L. Preziosi (2016), Aeolian Res.
![Page 19: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/19.jpg)
19
β’ Discrepancy among semi-empirical laws
β’ Sedimentation velocity affects the mode
of transport, distribution of particles above
the ground, and transport rate
β’ Sedimentation velocity bound to drag
coefficient
from L. Raffaele, L. Bruno, D. Sherman (2020), Aeolian Research
Incoming Windblown Sand: probabilistic ππ
πππ π’βπ‘ , ππ , π’β
![Page 20: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/20.jpg)
20
πΉπ =1
2ππππ
2πΆπππ2
4
πΉπ = πππππ3
6
πΉπ = πππππ3
6
π π =ππ π
ππ
πΆπ =4
3
ππ(ππ β ππ)
π π2ππ2 ππ3
π =3
4
πΆππ π2ππ
2
ππ ππ β ππ π
1/3
ππ =π π πππ
Incoming Windblown Sand: probabilistic ππ
![Page 21: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/21.jpg)
21
Incoming Windblown Sand
Threshold shear velocity π’βπ‘ β π’βπ‘(π)
Wind shear velocity π’β = π’β π10, π§0
πππ = ΰΆ±0
+β
π(π§) ππ§
β πππ(π’β, π’βπ‘)
Incoming
sand transport rate
π(π’βπ‘ π from Raffaele et al (2016)
if
if
π’β > π’βπ‘
π’β β€ π’βπ‘
π(πππ) =π΄
π
ππ
ππππ π’β
3 1 βπ(π’βπ‘ π
π(π’β)
0
mean wind velocity profile
shear stress @ wind-sand interfacesand bed
sand flux profile
1
2
3
π(π’β) =ππ π10,πππ
ln π§πππ/π§0
π’βπ‘
π
from Raffaele et al (2017a)
![Page 22: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/22.jpg)
22
Aerodynamics/Morphodynamics1
2
3source windblown sand path structure
Plan view
Side view
πππSedimentation coefficient
β’ πΆπ = ππ /πππ β 0,1
β’ πΆπ π, Ξ0, π
β’ monotonic decreasing vs π
β’ No closed forms
πΆπ
π
WT or CFD testing
π ππ = πΆπ π πππ
π πππ’π‘ = 1 β πΆπ π πππ
Sedimentation rate
Outgoing transport rate
![Page 23: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/23.jpg)
23
Windblown sand action1
2
3
Resistant sand volume ππ
β’ Structure/Infrastructure
β’ SMM
SLS
efficiency (πͺπ)
Time-variant WbS action π(π‘)
ππ(π‘) =
π=1
β
π1 β ππ β β―β ππ ππ P[ππ = π]
e.g. ππ,π = 5%
π π‘ < ππ Time variant reliability analysis
ππ
t
ππ = ππβ1(ππ,π)
Characteristic time of failure
ππ π‘ = π π π‘ β₯ ππ
= ΰΆ±0
+β
πΉππ π₯ ππ π₯, π‘ ππ₯ = 1 β πΉπ(ππ , π‘)
Probability of failure
sand removal period β€ ππ
![Page 24: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/24.jpg)
24
Site characteristicsπ§0 = 4π β 3 mWind statistics
from Raffaele et al (2017b)
πβπ statistics (π = 0.35 mm)
.5
1
1.5
2
.2 .4 .6 .8π’βπ‘
ππ’βπ‘π
![Page 25: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/25.jpg)
25
Sedimentation coefficient: SVW
3 mπ = 90Β°
Wind flow: RANS k-
Sand phase: sand mass conservation equation
πππ ππ‘
+ π» β π = 0
π = ππ‘πππ +ππππ β ππππππ πβ1π»ππ
Eulerian 1st order multiphase model for windblown sand
Straight Vertical Wall (SVW) from multiphase CFD simulation
from A. Lo Giudice, L. Preziosi (2020), App. Math. Modelling
![Page 26: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/26.jpg)
26
Sedimentation coefficient: SVW
![Page 27: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/27.jpg)
27
Sedimentation coefficient: SVW
π
πΆπ
π/ ΰ΄€π0 .2 .4 .6 .8 1
.2
.4
.6
.8
1data
fitting
β’ Multiphase simulation
β’ Standard CFD simulation
![Page 28: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/28.jpg)
28
Sedimentation coefficient: S4S
1. Trapping vortex
2. Reversed flow close to the ground
3. Sand subtraction from the
incoming flux
Shield for Sand (S4S) from WT tests @
3 m
![Page 29: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/29.jpg)
29
Sedimentation coefficient: S4S
β’ Wind tunnel setup in L-1B
β’ PIV-PTV measurement setup
![Page 30: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/30.jpg)
30
Sedimentation coefficient: S4S
β’ Sand concentration and morphodynamics
from Raffaele et al (under review)
![Page 31: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/31.jpg)
31
Sedimentation coefficient: S4S
data
fitting
π/ ΰ΄€π0 .2 .4 .6 .8 1
.2
.4
.6
.8
1
πΆπ
π
β’ Wind Tunnel test
β’ Standard CFD simulation
![Page 32: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/32.jpg)
32
data
fitting
ππ / ΰ΄€π
0 .2 .4 .6 .8 1
.2
.4
.6
.8
1
πΆπ
40% porosity fence from WT tests Hotta and Horikawa (1990)
ππ ππ ππ
3 m
Sedimentation coefficient: porous fence
![Page 33: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/33.jpg)
33
Sedimentation coefficient: embankment & track
SMM embankmenttrack
4 mπ = 90Β°
From WT tests Hotta & Horikawa (1990)
conjectured
πΆπ
0.25 m
2.5 m
Ballast void filling
ππ = 0,9 ΰ΄€π
SULS full coveringdata
fitting
πΆπ
ππ
ππ
source
![Page 34: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/34.jpg)
34
Results: SVW configuration, North side
![Page 35: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/35.jpg)
35
Results: SVW configuration, North side
ππ ππ
ππ,π=5%
ππ
ππ,π=5%
ππ,π=5%
![Page 36: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/36.jpg)
36
Lesson learnt: design perspective
β’ The higher the wind occurrence (i.e. North side):
β’ the lower ππ, for a given SMM capacity
β’ the higher the required capacity, for a given ππ
ππ as a design requirement for SMM
β’ S4S scores overall good performance, while SVW shows the poorest performance.
ππ,π‘ππππ S4S
ππ,π‘ππππ SVW~6
ππ,πππ S4S
ππ,πππ SVW~6
ππ,πππ S4S
ππ,πππ SVW~3
from Raffaele & Bruno (2020)
![Page 37: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/37.jpg)
37
Life Cycle Cost Analysis
paritySMM completion
S4S savings ~6M$/km
SVW savings ~3M$/km
S4S savings ~12M$/km
SVW savings ~6M$/km
cumulated savings are impressive w.r.t. railway avg worth (in ME ~4M$/km)from Raffaele & Bruno (2020)
![Page 38: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/38.jpg)
38
Conclusion & Perspectives
To conclude..
β’ Assess the performances of SMMs
β’ Plan sand removal maintenance operations
The proposed modelling framework allows to:
β’ Move from trial-and-error to rationale design
β’ Assess the economic impact of SMMs
Some perspectivesβ¦
β’ Development of innovative Wind-Sand tunnel tests to assess the sedimentation
coefficient of different SMM
β’ Extrapolation from scale to full-scale conditions under different environmental
setups
![Page 39: Windblown Sand Modelling and Mitigation](https://reader034.fdocuments.us/reader034/viewer/2022042412/625ee600f269b83db3030afe/html5/thumbnails/39.jpg)
39
HyPer SMM received funding from the European Unionβs Horizon 2020 research and innovation programme under the Marie SkΕodowskaβCurie grant agreement No 885985
EU Horizon 2020 Marie Curie Individual Fellowship
Hybrid Performance assessment of
Sand Mitigation Measures
Conclusion & Perspectives