Version HTML de base
60
Spécial “ 5es Assises sonore ”
Acoustique
&
Techniques n° 51
he Directive 2002/49/CE recommends the use
of the RMR Dutch method for railway noise mapping in
countries that don’t have their own official method. As in
France, the official method for railway noise mapping is
NMPB, that was used for a long time ago to predict the
transportation noise, a comparison of the national French
method (NMPB) and the interim Dutch method (RMR) for
railway noise is needed.
On the other hand, the prediction softwares used to
compute the noise maps include several methods. Then, a
cross-validation is needed between softwares and methods
to guarantee the accuracy of the predictions.
In the first paragraph, the implementation of NMPB method
is compared in Mithra, CadnaA and IMMI softwares. And the
implementation of RMR is compared in CadnaA and IMMI.
Then, a comparison of the NMPB and RMR methods
is presented in paragraph 2. Afterwards, noise levels
predicted in simplified situations have been compared. In
the end, real situations have been modelled in order to
compare the results of the simulations (RMR and NMPB)
with measured data.
Comparison of NMPB method in Mithra and
CadnaA softwares
The NMPB calculation method has been compared in Mithra
and CadnaA softwares. The comparison has been carried
out by means of results in punctual receivers and horizontal
and vertical maps.
The obtained results are quite similar with both softwares,
excepted for the night time period, where in some emission
configurations, CadnaA overestimates the noise levels.
It seems to be an error in the software development.
In general, the results depend on the meteorological
conditions and the ground absorption. The average
difference in absolute value calculated in many receivers
positioned at different distances from the source is
presented figure 1 according the ground parameter G.
The higher the ground absorption, the higher the difference
between models. This difference can reach 2,5 dB when
ground is fully absorbent (G=1) which is not representative
in practice.
Noise mapping for railway noise: assessment
of NMPB method as implemented in different
software, comparison with RMR method
Nagore Tellado, Stéphanie Pinaquy
LABEIN-Tecnalia
c/Geldo
Parque Tecnológico de Bizkaia
Edificio 700
48060 Derio Bizkaia
Espagne
E-mail : ntellado@labein.es
E-mail : spinaquy@labein.es
Pierre-Etienne Gautier, Franck Poisson
SNCF
45, rue de Londres
75379 Paris
E-mail : pierre-etienne.gautier@sncf.fr
E-mail : franck.poisson@sncf.fr
Abstract
The Directive 2002/49/CE on the Assessment and Management of Environmental
Noise recommends the use of the interim Dutch method (RMR) [1] for railway noise
mapping in countries that don’t have their own official method. As in France the
official method for railway noise mapping is the “Nouvelle Méthode de Prévision
du Bruit” (NMPB) [2], a comparison of the NMPB and the RMR methods for railway
noise has been developed.
Firstly, the implementation of NMPB and RMR methods in different computation
models is analysed. The implementation of NMPB method is compared in Mithra,
CadnaA and Immi models ; the implementation of RMR is compared in CadnaA and
Immi.
Then, a comparison of the methods has been carried out, afterwards noise levels
predicted in simplified situations have been compared and, at the end, more
complicated situations have been modelled in order to compare the results of the
simulations (RMR and NMPB) with measured data.
Résumé
La Directive 2002/49/CE sur l’évaluation et la gestion du Bruit dans l’environnement
recommande l’utilisation provisoire de la méthode hollandaise (RMR) [1] pour les
cartes de bruit ferroviaire dans les pays qui n’ont pas leur propre méthode officielle.
En France, la méthode officielle pour les cartes de bruit ferroviaire étant la “Nouvelle
méthode de prévision du bruit” (NMPB), une comparaison entre cette méthode et
des méthodes RMR sur le bruit des chemins de fer a été développée.
Tout d’abord, les mises en œuvre de la NMPB et des méthodes RMR dans des
modèles de calcul différents ont été analysées. La mise en œuvre de la méthode
NMPB a été évaluée dans les modèles acoustiques Mithra, CadnaA et Immi ; la mise
en œuvre de RMR dans CadnaA et Immi.
Puis une comparaison entre ces méthodes a été effectuée, à la suite de quoi des
niveaux sonores calculés dans des situations simples ont été comparés, et, enfin,
des situations plus compliquées ont été modélisées pour comparer les résultats
des simulations (RMR et NMPB) avec des données mesurées.
T