dirk van maerckeimagine final conference, budapest, 25.10.2006 1 development of the harmonoise...

Dirk Van Maercke IMAGINE Final Conference, Budapest, 25.10.2006 1

Development of the HARMONOISE Point-To-Point MODEL

Prediction of Excess Attenuation in Outdoor Noise Propagation

Dirk van Maercke

CSTB24, rue Joseph Fourier

F-38400 Saint-Martin d’Hères FRANCE

[email protected]

mailto:[email protected]


The Point-To-Point moduleCalculation of excess attenuation:

over any kind of terrain profile, including multiple diffractions & reflections by natural and man-made obstacles

under different representative meteorological conditions

using physical models, based on “recent” progress in outdoor propagation

more accurate than existing (national, ISO 9613-2 based) models

limited computation time “just good enough”

continuous results, no excessive sensitivity to accuracy of input data

compatible or comparable with basic principles of existing models

ready to be integrated in existing noise mapping software

2

log10free

excess ppL


INPUT TO THE P2P MODEL

cross-section profile

Pi (xi ,yi ), i = 0,1,…N where xi+1 > xi

each segment (Pi, Pi+1) = impedance value / class

source and receiver height

hS ± hS and hR ± hR

any number of diffractions (thin screens, wedges, thick barriers,…)

no distinction between ground (terrain), road surfaces, embankments, barriers, buildings, roofs,…


INPUT TO THE P2P MODEL (continued)

Impedance values- User defined- Delany-Bazley (flow resistivity and layer thickness)- Impedance classes (Nord 2000) + predefined impedance values

Frequency range- Default: 1/3 octave bands, 25 – 10000Hz- User defined

Atmospheric conditions- Sound speed (default = 340 m/s)

- Temperature and humidity (ISO 9613-1)- Sound speed gradient- Turbulence strength


Point-To-Point Model development

Step-by-step, increasing complexity: Ground effect Diffraction Ground + diffraction Combined model : ground + multiple diffraction Meteorological effects Long time averaging

Approach:

Analytical model (base solution, simple situation)

Heuristics (extension & adaptation to more realistic situations)

Validation against numerical “reference” calculations

Validation against experimental results


The P2P model : GROUND REFLECTION

• Chien-Soroka

• Mixed ground & Fresnel weighting

• Deep valley solution

• Roughness, almost flat ground

),(),'(

.1log20, RSpRSp

QwLfree

ifree

iiivalleyG

),,(.),(),'(

1log20 ZkQRSpRSp

Lfree

freeG

i

iGiflatG LwL ,, .

R

S

S'


The P2P model : WEGDE DIFFRACTION

Hadden & Pierce + heuristics

Approximation (from NMPB):

),,()',,(

1.),,(),,'(

1.),,(

)',,'(..

)',,(.

),,'(.

),,(

RDSpRDSp

QRDSpRDSp

QRDSpp

RDSpQQ

RDSpQ

RDSpQ

RDSpp

dif

difR

dif

difSdif

difRS

difR

difS

dif


The P2P model : simplified diffraction formula

RGSGDwedge LLLL ,,

),,()',,(

1log20

),,(),,'(

1log20

),(),,(

log20

,

,

RDSpRDSp

QL

RDSpRDSp

QL

RSpRDSp

L

dif

difRRG

dif

difSSG

free

difD


DIFFRACTION : INSERTION LOSS

2

),,(),,(

log20)(

N

ddpddp

NLRSfree

RSdifD

1)log(1016125.088

25.00126

025.0126

25.00)(

NifNNifN

NifN

NifN

NifNLdif

)(

...)32

().(42

RS

RS

RS

dddd

dd

special case : >> Deygout’s approximation


DIFFRACTION + GROUND

RGSGDgrounddiff LLLL ,,


MULTIPLE DIFFRACTION

two screens, no ground

extension to multiple screens : recursively !

),,(),,( 21121, RDDLRDSLL DDdif

most diffracting edge secondary diffracting edge

),...,(),...,(

),,(),...,(

110

010

Nkkdifkdif

NkDNdif

PPPLPPPL

PPPLPPPL


MULTIPLE DIFFRACTION + GROUND

),(

),'(

)',(

)'(

)(

RDL

DDL

DSL

DL

DLL

ground

ground

ground

dif

dif

)(DLdif

)'(DLdif ),( RDLground

),'( DDLground)',( DSLgroundbinary tree structure !


MULTIPLE DIFFRACTION & WIDE BARRIERS

dBRDDLanddBDDLDDas

RDL

RDDLDDL

RDSLDSLL

difG

G

dif

G

dif

G

6),,(6),(:

),(

),,(),(

),,(),(

212121

2

21

21

1

1

6),(3.06),(':2/

2121

DDLDDLandZcasein

GG


The P2P model : MULTIPLE DIFFRACTION + GROUND

),...,( 0 NPPL

transition

),...,( 0 NG PPL

),...,( 0 Nflat PPL

),...,( 0 Nvalley PPL ),...,(),...,(),,(

),...,(

0

0

0

Nk

k

NkD

ND

PPLPPLPPPL

PPL

diffractionno diffraction

recursive call


N

jijG

M

iiDexcess LLL ,

1,

COMBINED MODEL : COMPUTANIONAL EFFORT ?

the computation time varies as linearly as O(M+N)

instead of O(MN) !


METEOROLOGICAL REFRACTION Linear sound speed gradient analytical construction of rays

Curved rays… curved ground analogy ? Conformal mapping !

Ryczc 1)( 0

Rd

Rdcdc

ifconstantalmost...

;2

1)(2

0

jRwwRjyjxw

22'''

yjxw


The Point-To-Point model= DIFFRACTION + GROUND + METEOROLOGICAL REFRACTION

After conformal transformation…

use “straight ray” model over modified terrain model !


Fine-tuning the model & secondary effects… Loss of coherency

Turbulence and scattering (simplified, one parameter)

22222

R

R

S

S

hh

hh

cc

ff

100log10

1000log3log1025 dfL Tscat


VALIDATION AGAINST BEM CALCULATIONS (no meteo)

hS = 0.30 m ; hR = 5.00 m ; d = 100.0 m

thin barrier / wedge , h = 2.00 m, d = 10.0 m

thin barrier wedge


NUMERICAL VALIDATION (BEM)

hS = 0.30 m ; dS = 10.0 m

hR = 5.00 m ; dR = 50.0 m

hB = 3.00 m; dB = 2.00 m

B = 50 or 2000 kNsm-4

hard soft



embankments, earth walls, round hill tops, ….

receiver grids : d = 50 / 100 / 250 / 500 m , h = 1.25 / 2.5 / 5.0 / 10 m

cylindrical hill top embankment



with screen (180 cases)

0

10

20

30

40

50

60

70

80

90

100

-5 -4 -3 -2 -1 0 1 2 3 4 5

with screen

without screen (880 cases)

0

50

100

150

200

250

300

350

400

450

-5 -4 -3 -2 -1 0 1 2 3 4 5

without screen


VALIDATION AGAINST GFPE CALCULATIONS (with meteo)

hS = 0.30 m ; d = 500.0 m

hR = 2.50 / 10.0 m ;

= 200 kNsm-4

Alin= 0.04 s-1 (R ~ 18 D)

hR = 2.5 m hR = 10 m

weak linear gradient


NUMERICAL VALIDATION (GFPE)

hS = 0.30 m

d = 500.0 m

= 2000 kNsm-4

Alin= 0.04 s-1 (R ~ 18 D)

Low source / low receiver ?hR = 5 m

hR = 1.25 mhR = 2.5 m



hS = 0.30 m

hR = 2.50m

= 200 kNsm-4

Alin= 0.12 s-1 (R ~ 6 D)

Stronger gradients ?

d = 500 m

d = 100 m

d = 250 m



A = 0.04 (1/s)

0

20

40

60

80

100

120

-5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0

A = 0.12 (1/s)

0

20

40

60

80

100

120

-5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0


VALIDATION AGAINT EXPERIMENTAL RESULTS

Measurements: 2 to 10 weeks, records of sound levels and meteorological parameters (W,W, 1/L) in 30’ steps

Reference models : lin/log sound speed profiles for each point source and for each time period

Engineering model : 8 wind directions, 3 wind speeds, 5 stabilityclasses = 120 cases + frequency of occurrence


VALIDATION AGAINT EXPERIMENTAL RESULTS

IMAGINE.WP7

EDF / DEBAKOM


BUT :

A MODEL IS JUST A MODEL…

dirk van maerckeimagine final conference, budapest, 25.10.2006 1 development of the harmonoise...

Documents