1.7.2015 
EN 
Official Journal of the European Union 
L 168/1 
COMMISSION DIRECTIVE (EU) 2015/996
of 19 May 2015
establishing common noise assessment methods according to Directive 2002/49/EC of the European Parliament and of the Council
(Text with EEA relevance)
THE EUROPEAN COMMISSION,
Having regard to the Treaty on the Functioning of the European Union,
Having regard to Directive 2002/49/EC of the European Parliament and of the Council of 25 June 2002 relating to the assessment and management of environmental noise (1), and in particular Article 6, paragraph 2 thereof,
Whereas:
(1) 
According to its Article 1, the aim of Directive 2002/49/EC is to define a common approach intended to avoid, prevent or reduce on a prioritised basis the harmful effects, including annoyance, due to exposure to environmental noise. To that end the Member States shall determine the exposure to environmental noise, through noise mapping, by methods of assessment common to the Member States, shall ensure that information on environmental noise and its effects is made available to the public and shall adopt action plans based upon noisemapping results, with a view to preventing and reducing environmental noise where necessary and particularly where exposure levels can induce harmful effects on human health, and to preserving environmental noise quality where it is good. 
(2) 
According to Article 5 of Directive 2002/49/EC, Member States shall apply the noise indicators (Lden and Lnight) referred to in Annex I to that Directive for the preparation and revision of strategic noise mapping in accordance with Article 7. 
(3) 
According to Article 6 of Directive 2002/49/EC, the values of the noise indicators (Lden and Lnight) shall be determined by means of the assessment methods defined in Annex II to that Directive. 
(4) 
According to Article 6 of Directive 2002/49/EC, the Commission shall establish common assessment methods for the determination of the noise indicators Lden and Lnight through a revision of Annex II. 
(5) 
According to Article 7 of Directive 2002/49/EC, Member States shall ensure that strategic noise maps are made no later than 30 June 2007 and 30 June 2012 and thereafter reviewed, and revised if necessary, at least every 5 years. 
(6) 
Directive 2002/49/EC provides for action plans to be based on strategic noise maps. Strategic noise maps shall be drawn up with the common assessment methods when these methods have been adopted by Member States. However, Member States may use other methods to design measures addressing priorities identified by using the common methods as well as for assessment of other national measures to prevent and reduce environmental noise. 
(7) 
In 2008, the Commission launched the development of the common noise assessment methodological framework through the project ‘Common Noise Assessment Methods in the EU’ (CNOSSOSEU) led by its Joint Research Centre. The project was carried out in close consultation with the Committee established under Article 18 of Directive 2000/14/EC of the European Parliament and of the Council (2), and other experts from Member States. Its results were published in the JRC Reference Report on CNOSSOSEU (3). 
(8) 
The Annex to this Commission Directive sets out the common assessment methods. Member States are required to use these methods from 31 December 2018 onwards. 
(9) 
The assessment methods provided for in the Annex to this Directive are, according to its Article 2, paragraph 1, to be adopted by 31 December 2018 at the latest and until that date Member States may, according to Article 6, paragraph 2 of Directive 2002/49/EC, continue to use the existing assessment methods that they have previously adopted at the national level. 
(10) 
In accordance with Article 12 of Directive 2002/49/EC, the Commission shall adapt Annex II to technical and scientific progress. 
(11) 
Apart from the adaptation to scientific and technical progress in accordance with Article 12 of Directive 2002/49/EC, the Commission shall endeavour to modify the Annex based on the experience from Member States. 
(12) 
The common assessment methods are also to be used for the purpose of other EU legislation where that legislation refers to Annex II to Directive 2002/49/EC. 
(13) 
The measures provided for in this Directive are in accordance with the opinion of the Committee established under Article 13 of Directive 2002/49/EC, 
HAS ADOPTED THIS DIRECTIVE:
Article 1
Annex II to Directive 2002/49/EC is replaced by the text set out in the Annex to this Directive.
Article 2
1. Member States shall bring into force the laws, regulations and administrative provisions necessary to comply with this Directive by 31 December 2018 at the latest. They shall forthwith communicate to the Commission the text of those provisions.
When Member States adopt those provisions, they shall contain a reference to this Directive or be accompanied by such a reference on the occasion of their official publication. Member States shall determine how such reference is to be made.
2. Member States shall communicate to the Commission the text of the main provisions of national law which they adopt in the field covered by this Directive.
Article 3
This Directive shall enter into force on the day following that of its publication in the Official Journal of the European Union.
Article 4
This Directive is addressed to the Member States.
Done at Brussels, 19 May 2015.
For the Commission,
On behalf of the President,
Karmenu VELLA
Member of the Commission
(1) OJ L 189, 18.7.2002, p. 12.
(2) Directive 2000/14/EC of the European Parliament and of the Council of 8 May 2000 on the approximation of the laws of the Member States relating to the noise emission in the environment by equipment for use outdoors (OJ L 162, 3.7.2000, p. 1).
(3) Common Noise Assessment Methods in Europe (CNOSSOSEU) — JRC Reference Report, EUR 25379 EN. Luxembourg: Publications Office of the European Union, 2012, — ISBN 9789279252815
ANNEX
ASSESSMENT METHODS FOR THE NOISE INDICATORS
(Referred to in Article 6 of Directive 2002/49/EC)
1. INTRODUCTION
The values of Lden and Lnight shall be determined at the assessment positions by computation, according to the method set out in Chapter 2 and the data described in Chapter 3. Measurements may be performed according to Chapter 4.
2. COMMON NOISE ASSESSMENT METHODS
2.1. General provisions — Road traffic, railway and industrial noise
2.1.1. Indicators, frequency range and band definitions
Noise calculations shall be defined in the frequency range from 63 Hz to 8 kHz. Frequency band results shall be provided at the corresponding frequency interval.
Calculations are performed in octave bands for road traffic, railway traffic and industrial noise, except for the railway noise source sound power, that uses third octave bands. For road traffic, railway traffic and industrial noise, based on these octave band results, the Aweighted long term average sound pressure level for the day, evening and night period, as defined in Annex I and referred to in Art. 5 of Directive 2002/49/EC, is computed by summation over all frequencies:

(2.1.1) 
where

Ai denotes the Aweighting correction according to IEC 616721 

i = frequency band index 

and T is the time period corresponding to day, evening or night. 
Noise parameters:
Lp 
Instantaneous sound pressure level 
[dB] (re. 2 10–5 Pa) 
LAeq,LT 
Global longterm sound level LAeq due to all sources and image sources at point R 
[dB] (re. 2 10–5 Pa) 
LW 
‘In situ’ sound power level of a point source (moving or steady) 
[dB] (re. 10–12 W) 
LW,i,dir 
Directional ‘in situ’ sound power level for the ith frequency band 
[dB] (re. 10–12 W) 
LW′ 
Average ‘in situ’ sound power level per metre of source line 
[dB/m] (re. 10–12 W) 
Other physical parameters:
p 
r.m.s. of the instantaneous sound pressure 
[Pa] 
p 0 
Reference sound pressure = 2 10–5 Pa 
[Pa] 
W 0 
Reference sound power = 10–12 W 
[watt] 
2.1.2. Quality framework
All input values affecting the emission level of a source shall be determined with at least the accuracy corresponding to an uncertainty of ± 2dB(A) in the emission level of the source (leaving all other parameters unchanged).
In the application of the method, the input data shall reflect the actual usage. In general there shall be no reliance on default input values or assumptions. Default input values and assumptions are accepted if the collection of real data is associated with disproportionately high costs.
Software used to perform the calculations shall prove compliance with the methods herewith described by means of certification of results against test cases.
2.2. Road traffic noise
2.2.1. Source description
The road traffic noise source shall be determined by combining the noise emission of each individual vehicle forming the traffic flow. These vehicles are grouped into five separate categories with regard to their characteristics of noise emission:
Category 1 
: 
Light motor vehicles 
Category 2 
: 
Medium heavy vehicles 
Category 3 
: 
Heavy vehicles 
Category 4 
: 
Powered twowheelers 
Category 5 
: 
Open category 
In the case of powered twowheelers, two separate subclasses are defined for mopeds and more powerful motorcycles, since they operate in very different driving modes and their numbers usually vary widely.
The first four categories shall be used, and the fifth category is optional. It is foreseen for new vehicles that may be developed in the future and may be sufficiently different in their noise emission to require an additional category to be defined. This category could cover, for example, electric or hybrid vehicles or any vehicle developed in the future substantially different from those in categories 1 to 4.
The details of the different vehicle classes are given in Table [2.2.a].
Table [2.2.a]
Vehicle classes
Category 
Name 
Description 
Vehicle category in EC Whole Vehicle Type Approval (1) 

1 
Light motor vehicles 
Passenger cars, delivery vans ≤ 3,5 tons, SUVs (2), MPVs (3) including trailers and caravans 
M1 and N1 

2 
Medium heavy vehicles 
Medium heavy vehicles, delivery vans > 3,5 tons, buses, motorhomes, etc. with two axles and twin tyre mounting on rear axle 
M2, M3 and N2, N3 

3 
Heavy vehicles 
Heavy duty vehicles, touring cars, buses, with three or more axles 
M2 and N2 with trailer, M3 and N3 

4 
Powered twowheelers 
4a 
Two, Three and Fourwheel Mopeds 
L1, L2, L6 
4b 
Motorcycles with and without sidecars, Tricycles and Quadricycles 
L3, L4, L5, L7 

5 
Open category 
To be defined according to future needs 
N/A 
In this method, each vehicle (category 1, 2, 3, 4 and 5) is represented by one single point source radiating uniformly into the 2π half space above the ground. The first reflection on the road surface is treated implicitly. As depicted in Figure [2.2.a], this point source is placed 0,05 m above the road surface.
Figure [2.2.a]
Location of equivalent point source on light vehicles (category 1), heavy vehicles (categories 2 and 3) and twowheelers (category 4)
Equivalent source (0,05 m high)
Equivalent source (0,05 m high)
Equivalent source (0,05 m high)
The traffic flow is represented by a source line. In the modelling of a road with multiple lanes, each lane should ideally be represented by a source line placed in the centre of each lane. However, it is also acceptable to model one source line in the middle of a two way road or one source line per carriageway in the outer lane of multilane roads.
The sound power of the source is defined in the ‘semifree field’, thus the sound power includes the effect of the reflection of the ground immediately under the modelled source where there are no disturbing objects in its immediate surroundings except for the reflection on the road surface not immediately under the modelled source.
The noise emission of a traffic flow is represented by a source line characterised by its directional sound power per metre per frequency. This corresponds to the sum of the sound emission of the individual vehicles in the traffic flow, taking into account the time spent by the vehicles in the road section considered. The implementation of the individual vehicle in the flow requires the application of a traffic flow model.
If a steady traffic flow of Qm vehicles of category m per hour is assumed, with an average speed vm (in km/h), the directional sound power per metre in frequency band i of the source line LW′, eq,line,i,m is defined by:

(2.2.1) 
where LW,i,m is the directional sound power of a single vehicle. LW′,m is expressed in dB (re. 10–12 W/m). These sound power levels are calculated for each octave band i from 125 Hz to 4 kHz.
Traffic flow data Qm shall be expressed as yearly average per hour, per time period (dayeveningnight), per vehicle class and per source line. For all categories, input traffic flow data derived from traffic counting or from traffic models shall be used.
The speed vm is a representative speed per vehicle category: in most cases the lower of the maximum legal speed for the section of road and the maximum legal speed for the vehicle category. If local measurement data is unavailable the maximum legal speed for the vehicle category shall be used.
In the traffic flow, all vehicles of category m are assumed to drive at the same speed, i.e. vm , the average speed of the flow of vehicles of the category.
A road vehicle is modelled by a set of mathematical equations representing the two main noise sources:
1. 
Rolling noise due to the tyre/road interaction; 
2. 
Propulsion noise produced by the driveline (engine, exhaust, etc.) of the vehicle. 
Aerodynamic noise is incorporated in the rolling noise source.
For light, medium and heavy motor vehicles (categories 1, 2 and 3), the total sound power corresponds to the energetic sum of the rolling and the propulsion noise. Thus, the total sound power level of the source lines m = 1, 2 or 3 is defined by:

(2.2.2) 
where LWR,i,m is the sound power level for rolling noise and LWP,i,m is the sound power level for propulsion noise. This is valid on all speed ranges. For speeds less than 20 km/h it shall have the same sound power level as defined by the formula for vm = 20 km/h.
For twowheelers (category 4), only propulsion noise is considered for the source:
LW,i,m = 4 (vm = 4 ) = LWP,i,m = 4 (vm = 4 ) 
(2.2.3) 
This is valid on all speed ranges. For speeds less than 20 km/h it shall have the same sound power level as defined by the formula for vm = 20 km/h.
2.2.2. Reference conditions
The source equations and coefficients are valid for the following reference conditions:
— 
a constant vehicle speed 
— 
a flat road 
— 
an air temperature τref = 20 °C 
— 
a virtual reference road surface, consisting of an average of dense asphalt concrete 0/11 and stone mastic asphalt 0/11, between 2 and 7 years old and in a representative maintenance condition 
— 
a dry road surface 
— 
no studded tyres. 
2.2.3. Rolling noise
The rolling noise sound power level in the frequency band i for a vehicle of class m = 1,2 or 3 is defined as:

(2.2.4) 
The coefficients AR,i,m and BR,i,m are given in octave bands for each vehicle category and for a reference speed vref = 70 km/h. ΔLWR,i,m corresponds to the sum of the correction coefficients to be applied to the rolling noise emission for specific road or vehicle conditions deviating from the reference conditions:
ΔLWR,i,m = ΔLWR,road,i,m + ΔLstuddedtyres,i,m + ΔLWR,acc,i,m + ΔLW,temp 
(2.2.5) 
ΔLWR,road,i,m accounts for the effect on rolling noise of a road surface with acoustic properties different from those of the virtual reference surface as defined in Chapter 2.2.2. It includes both the effect on propagation and on generation.
ΔLstudded tyres,i,m is a correction coefficient accounting for the higher rolling noise of light vehicles equipped with studded tyres.
ΔLWR,acc,i,m accounts for the effect on rolling noise of a crossing with traffic lights or a roundabout. It integrates the effect on noise of the speed variation.
ΔLW,temp is a correction term for an average temperature τ different from the reference temperature τref = 20 °C.
In situations where a significant number of light vehicles in the traffic flow use studded tyres during several months every year, the induced effect on rolling noise shall be taken into account. For each vehicle of category m = 1 equipped with studded tyres, a speeddependent increase in rolling noise emission is evaluated by:
Δstud,i (v) = 
ai + bi × lg(50/70) for v < 50 km/h 
(2.2.6) 
ai + bi × lg(v/70) for 50 ≤ v ≤ 90 km/h 

ai + bi × lg(90/70) for v > 90 km/h 
where coefficients ai and bi are given for each octave band.
The increase in rolling noise emission shall only be attributed according to the proportion of light vehicles with studded tyres and during a limited period Ts (in months) over the year. If Qstud,ratio is the average ratio of the total volume of light vehicles per hour equipped with studded tyres during the period Ts (in months), then the yearly average proportion of vehicles equipped with studded tyres ps is expressed by:

(2.2.7) 
The resulting correction to be applied to the rolling sound power emission due to the use of studded tyres for vehicles of category m = 1 in frequency band i shall be:

(2.2.8) 
For vehicles of all other categories no correction shall be applied:
ΔLstuddedtyres,i,m ≠ 1 = 0 
(2.2.9) 
The air temperature affects rolling noise emission; the rolling sound power level decreases when the air temperature increases. This effect is introduced in the road surface correction. Road surface corrections are usually evaluated at an air temperature of τref = 20 °C. In the case of a different yearly average air temperature °C, the road surface noise shall be corrected by:
ΔLW,temp,m (τ) = Km × (τref – τ) 
(2.2.10) 
The correction term is positive (i.e. noise increases) for temperatures lower than 20 °C and negative (i.e. noise decreases) for higher temperatures. The coefficient K depends on the road surface and the tyre characteristics and in general exhibits some frequency dependence. A generic coefficient Km = 1 = 0,08 dB/°C for light vehicles (category 1) and Km = 2 = Km = 3 = 0,04 dB/°C for heavy vehicles (categories 2 and 3) shall be applied for all road surfaces. The correction coefficient shall be applied equally on all octave bands from 63 to 8 000 Hz.
2.2.4. Propulsion noise
The propulsion noise emission includes all contributions from engine, exhaust, gears, air intake, etc. The propulsion noise sound power level in the frequency band i for a vehicle of class m is defined as:

(2.2.11) 
The coefficients AP,i,m and BP,i,m are given in octave bands for each vehicle category and for a reference speed vref = 70 km/h.
ΔLWP,i,m corresponds to the sum of the correction coefficients to be applied to the propulsion noise emission for specific driving conditions or regional conditions deviating from the reference conditions:
ΔLWP,i,m = ΔLWP,road,i,m + ΔLWP,grad,i,m + ΔLWP,acc,i,m 
(2.2.12) 
ΔLWP,road,i,m accounts for the effect of the road surface on the propulsion noise via absorption. The calculation shall be performed according to Chapter 2.2.6.
ΔLWP,acc,i,m and ΔLWP,grad,i,m account for the effect of road gradients and of vehicle acceleration and deceleration at intersections. They shall be calculated according to Chapters 2.2.4 and 2.2.5 respectively.
The road gradient has two effects on the noise emission of the vehicle: first, it affects the vehicle speed and thus the rolling and propulsion noise emission of the vehicle; second, it affects both the engine load and the engine speed via the choice of gear and thus the propulsion noise emission of the vehicle. Only the effect on the propulsion noise is considered in this section, where a steady speed is assumed.
The effect of the road gradient on the propulsion noise is taken into account by a correction term ΔLWP,grad,m which is a function of the slope s (in %), the vehicle speed vm (in km/h) and the vehicle class m. In the case of a bidirectional traffic flow, it is necessary to split the flow into two components and correct half for uphill and half for downhill. The correction term is attributed to all octave bands equally:

For m = 1


For m = 2


For m = 3


For m = 4

The correction ΔLWP,grad,m implicitly includes the effect of slope on speed.
2.2.5. Effect of the acceleration and deceleration of vehicles
Before and after crossings with traffic lights and roundabouts a correction shall be applied for the effect of acceleration and deceleration as described below.
The correction terms for rolling noise, ΔLWR,acc,m,k , and for propulsion noise, ΔLWP,acc,m,k , are linear functions of the distance x (in m) of the point source to the nearest intersection of the respective source line with another source line. They are attributed to all octave bands equally:

(2.2.17) 

(2.2.18) 
The coefficients CR,m,k and CP,m,k depend on the kind of junction k (k = 1 for a crossing with traffic lights; k = 2 for a roundabout) and are given for each vehicle category. The correction includes the effect of change in speed when approaching or moving away from a crossing or a roundabout.
Note that at a distance x ≥ 100 m, ΔLWR,acc,m,k = ΔLWP,acc,m,k = 0.
2.2.6. Effect of the type of road surface
For road surfaces with acoustic properties different from those of the reference surface, a spectral correction term for both rolling noise and propulsion noise shall be applied.
The road surface correction term for the rolling noise emission is given by:

(2.2.19) 
where

αi,m is the spectral correction in dB at reference speed vref for category m (1, 2 or 3) and spectral band i. 

βm is the speed effect on the rolling noise reduction for category m (1, 2 or 3) and is identical for all frequency bands. 
The road surface correction term for the propulsion noise emission is given by:
ΔLWP,road,i,m = min{αi,m ;0} 
(2.2.20) 
Absorbing surfaces decrease the propulsion noise, while nonabsorbing surfaces do not increase it.
The noise characteristics of road surfaces vary with age and the level of maintenance, with a tendency to become louder over time. In this method the road surface parameters are derived to be representative for the acoustic performance of the road surface type averaged over its representative lifetime and assuming proper maintenance.
2.3. Railway noise
2.3.1. Source description
For the purposes of this noise calculation method, a vehicle is defined as any single railway subunit of a train (typically a locomotive, a selfpropelled coach, a hauled coach or a freight wagon) that can be moved independently and can be detached from the rest of the train. Some specific circumstances may occur for subunits of a train that are a part of a nondetachable set, e.g. share one bogie between them. For the purpose of this calculation method, all these subunits are grouped into a single vehicle.
For the purpose of this calculation method, a train consists of a series of coupled vehicles.
Table [2.3.a] defines a common language to describe the vehicle types included in the source database. It presents the relevant descriptors to be used to classify the vehicles in full. These descriptors correspond to properties of the vehicle, which affect the acoustic directional sound power per metre length of the equivalent source line modelled.
The number of vehicles for each type shall be determined on each of the track sections for each of the time periods to be used in the noise calculation. It shall be expressed as an average number of vehicles per hour, which is obtained by dividing the total number of vehicles travelling in a given time period by the duration in hours of this time period (e.g. 24 vehicles in 4 hours means 6 vehicles per hour). All vehicle types travelling on each track section shall be used.
Table [2.3.a]
Classification and descriptors for railway vehicles
Digit 
1 
2 
3 
4 
Descriptor 
Vehicle type 
Number of axles per vehicle 
Brake type 
Wheel measure 
Explanation of the descriptor 
A letter that describes the type 
The actual number of axles 
A letter that describes the brake type 
A letter that describes the noise reduction measure type 
Possible descriptors 
h high speed vehicle (> 200 km/h) 
1 
c castiron block 
n no measure 
m selfpropelled passenger coaches 
2 
k composite or sinter metal block 
d dampers 

p hauled passenger coaches 
3 
n nontread braked, like disc, drum, magnetic 
s screens 

c city tram or light metro selfpropelled and nonselfpropelled coach 
4 

o other 

d diesel loco 
etc. 



e electric loco 




a any generic freight vehicle 




o other (i.e. maintenance vehicles etc.) 



The existing tracks may differ because there are several elements contributing to and characterising their acoustic properties. The track types used in this method are listed in Table [2.3.b] below. Some of the elements have a large influence on acoustic properties, while others have only secondary effects. In general, the most relevant elements influencing the railway noise emission are: railhead roughness, rail pad stiffness, track base, rail joints and radius of curvature of the track. Alternatively, the overall track properties can be defined and, in this case, the railhead roughness and the track decay rate according to ISO 3095 are the two acoustically essential parameters, plus the radius of curvature of the track.
A track section is defined as a part of a single track, on a railway line or station or depot, on which the track's physical properties and basic components do not change.
Table [2.3.b] defines a common language to describe the track types included in the source database.
Table [2.3.b]
Digit 
1 
2 
3 
4 
5 
6 
Descriptor 
Track base 
Railhead Roughness 
Rail pad type 
Additional measures 
Rail joints 
Curvature 
Explanation of the descriptor 
Type of track base 
Indicator for roughness 
Represents an indication of the ‘acoustic’ stiffness 
A letter describing acoustic device 
Presence of joints and spacing 
Indicate the radius of curvature in m 
Codes allowed 
B Ballast 
E Well maintained and very smooth 
S Soft (150250 MN/m) 
N None 
N None 
N Straight track 
S Slab track 
M Normally maintained 
M Medium (250 to 800 MN/m) 
D Rail damper 
S Single joint or switch 
L Low (1 000 500 m) 

L Ballasted bridge 
N Not well maintained 
H Stiff (8001 000 MN/m) 
B Low barrier 
D Two joints or switches per 100 m 
M Medium (Less than 500 m and more than 300 m) 

N Nonballasted bridge 
B Not maintained and bad condition 

A Absorber plate on slab track 
M More than two joints or switches per 100 m 
H High (Less than 300 m) 

T Embedded track 


E Embedded rail 



O Other 


O Other 


Figure [2.3.a]
Equivalent noise sources position
The different equivalent noise line sources are placed at different heights and at the centre of the track. All heights are referred to the plane tangent to the two upper surfaces of the two rails.
The equivalent sources include different physical sources (index p). These physical sources are divided into different categories depending on the generation mechanism, and are: (1) rolling noise (including not only rail and track base vibration and wheel vibration but also, where present, superstructure noise of the freight vehicles); (2) traction noise; (3) aerodynamic noise; (4) impact noise (from crossings, switches and junctions); (5) squeal noise and (6) noise due to additional effects such as bridges and viaducts.
(1) 
The roughness of wheels and railheads, through three transmission paths to the radiating surfaces (rails, wheels and superstructure), constitutes the rolling noise. This is allocated to h = 0,5 m (radiating surfaces A) to represent the track contribution, including the effects of the surface of the tracks, especially slab tracks (in accordance with the propagation part), to represent the wheel contribution and to represent the contribution of the superstructure of the vehicle to noise (in freight trains). 
(2) 
The equivalent source heights for traction noise vary between 0,5 m (source A) and 4,0 m (source B), depending on the physical position of the component concerned. Sources such as gear transmissions and electric motors will often be at an axle height of 0,5 m (source A). Louvres and cooling outlets can be at various heights; engine exhausts for dieselpowered vehicles are often at a roof height of 4,0 m (source B). Other traction sources such as fans or diesel engine blocks may be at a height of 0,5 m (source A) or 4,0 m (source B). If the exact source height is in between the model heights, the sound energy is distributed proportionately over the nearest adjacent source heights. For this reason, two source heights are foreseen by the method at 0,5 m (source A), 4,0 m (source B), and the equivalent sound power associated with each is distributed between the two depending on the specific configuration of the sources on the unit type. 
(3) 
Aerodynamic noise effects are associated with the source at 0,5 m (representing the shrouds and the screens, source A), and the source at 4,0 m (modelling all over roof apparatus and pantograph, source B). The choice of 4,0 m for pantograph effects is known to be a simple model, and has to be considered carefully if the objective is to choose an appropriate noise barrier height. 
(4) 
Impact noise is associated with the source at 0,5 m (source A). 
(5) 
Squeal noise is associated with the sources at 0,5 m (source A). 
(6) 
Bridge noise is associated with the source at 0,5 m (source A). 
2.3.2. Sound power emission
The model for railway traffic noise, analogously to road traffic noise, describes the noise sound power emission of a specific combination of vehicle type and track type which fulfils a series of requirements described in the vehicle and track classification, in terms of a set of sound power per each vehicle (LW,0).
The noise emission of a traffic flow on each track shall be represented by a set of 2 source lines characterised by its directional sound power per metre per frequency band. This corresponds to the sum of the sound emissions due to the individual vehicles passing by in the traffic flow and, in the specific case of stationary vehicles, taking into account the time spent by the vehicles in the railway section under consideration.
The directional sound power per metre per frequency band, due to all the vehicles passing by each track section on the track type (j), is defined:
— 
for each frequency band (i), 
— 
for each given source height (h) (for sources at 0,5 m h = 1, at 4,0 m h = 2), 
and is the energy sum of all contributions from all vehicles running on the specific jth track section. These contributions are:
— 
from all vehicle types (t) 
— 
at their different speeds (s) 
— 
under the particular running conditions (constant speed) (c) 
— 
for each physical source type (rolling, impact, squeal, traction, aerodynamic and additional effects sources such as for example bridge noise) (p). 
To calculate the directional sound power per metre (input to the propagation part) due to the average mix of traffic on the jth track section, the following is used:

(2.3.1) 
where
Tref 
= 
reference time period for which the average traffic is considered 
X 
= 
total number of existing combinations of i, t, s, c, p for each jth track section 
t 
= 
index for vehicle types on the jth track section 
s 
= 
index for train speed: there are as many indexes as the number of different average train speeds on the jth track section 
c 
= 
index for running conditions: 1 (for constant speed), 2 (idling) 
p 
= 
index for physical source types: 1 (for rolling and impact noise), 2 (curve squeal), 3 (traction noise), 4 (aerodynamic noise), 5 (additional effects) 
LW′,eq,line,x 
= 
xth directional sound power per metre for a source line of one combination of t, s, c, p on each jth track section 
If a steady flow of Q vehicles per hour is assumed, with an average speed v, on average at each moment in time there will be an equivalent number of Q/v vehicles per unit length of the railway section. The noise emission of the vehicle flow in terms of directional sound power per metre LW′,eq,line (expressed in dB/m (re. 10–12 W)) is integrated by:
(for c = 1) 
(2.3.2) 
where
— 
Q is the average number of vehicles per hour on the jth track section for vehicle type t , average train speed s and running condition c 
— 
v is their speed on the jth track section for vehicle type t and average train speed s 
— 
LW,0,dir is the directional sound power level of the specific noise (rolling, impact, squeal, braking, traction, aerodynamic, other effects) of a single vehicle in the directions ψ, φ defined with respect to the vehicle's direction of movement (see Figure [2.3.b]). 
In the case of a stationary source, as during idling, it is assumed that the vehicle will remain for an overall time Tidle at a location within a track section with length L. Therefore, with Tref as the reference time period for the noise assessment (e.g. 12 hours, 4 hours, 8 hours), the directional sound power per unit length on that track section is defined by:
(for c = 2) 
(2.3.4) 
In general, directional sound power is obtained from each specific source as:
LW,0,dir,i (ψ,φ) = LW,0,i + ΔLW,dir,vert,i + ΔLW,dir,hor,i 
(2.3.5) 
where
— 
ΔLW,dir,vert,i is the vertical directivity correction (dimensionless) function of ψ (Figure [2.3.b]) 
— 
ΔLW,dir,hor,i is the horizontal directivity correction (dimensionless) function of φ (Figure [2.3.b]). 
And where LW,0,dir,i(ψ,φ) shall, after being derived in 1/3 octave bands, be expressed in octave bands by energetically adding each pertaining 1/3 octave band together into the corresponding octave band.
Figure [2.3.b]
Geometrical definition
Travelling direction
Vehicle (equivalent point source)
Emission direction
Horizontal plane
For the purpose of the calculations, the source strength is then specifically expressed in terms of directional sound power per 1 m length of track LW′,tot,dir,i to account for the directivity of the sources in their vertical and horizontal direction, by means of the additional corrections.
Several LW,0,dir,i (ψ,φ) are considered for each vehicletrackspeedrunning condition combination:
— 
for a 1/3 octave frequency band ( i ) 
— 
for each track section ( j ) 
— 
source height ( h ) (for sources at 0,5 m h = 1, at 4,0 m h = 2) 
— 
directivity ( d ) of the source 
A set of LW,0,dir,i (ψ,φ) are considered for each vehicletrackspeedrunning condition combination, each track section, the heights corresponding to h = 1 and h = 2 and the directivity.
The vehicle contribution and the track contribution to rolling noise are separated into four essential elements: wheel roughness, rail roughness, vehicle transfer function to the wheels and to the superstructure (vessels) and track transfer function. Wheel and rail roughness represent the cause of the excitation of the vibration at the contact point between the rail and the wheel, and the transfer functions are two empirical or modelled functions that represent the entire complex phenomena of the mechanical vibration and sound generation on the surfaces of the wheel, the rail, the sleeper and the track substructure. This separation reflects the physical evidence that roughness present on a rail may excite the vibration of the rail, but it will also excite the vibration of the wheel and vice versa. Not including one of these four parameters would prevent the decoupling of the classification of tracks and trains.
Rolling noise is mainly excited by rail and wheel roughness in the wavelength range from 5500 mm.
The roughness level Lr is defined as 10 times the logarithm to the base 10 of the square of the mean square value r2 of the roughness of the running surface of a rail or a wheel in the direction of motion (longitudinal level) measured in μm over a certain rail length or the entire wheel diameter, divided by the square of the reference value
:
dB 
(2.3.6) 
where
r 0 
= 
1 μm 
r 
= 
r.m.s. of the vertical displacement difference of the contact surface to the mean level 
The roughness level Lr is typically obtained as a spectrum of wavelength λ and it shall be converted to a frequency spectrum f = v/λ, where f is the centre band frequency of a given 1/3 octave band in Hz, λ is the wavelength in m, and v is the train speed in km/h. The roughness spectrum as a function of frequency shifts along the frequency axis for different speeds. In general cases, after conversion to the frequency spectrum by means of the speed, it is necessary to obtain new 1/3 octave band spectra values averaging between two corresponding 1/3 octave bands in the wavelength domain. To estimate the total effective roughness frequency spectrum corresponding to the appropriate train speed, the two corresponding 1/3 octave bands defined in the wavelength domain shall be averaged energetically and proportionally.
The rail roughness level (track side roughness) for the ith wavenumber band is defined as Lr,TR,i
By analogy, the wheel roughness level (vehicle side roughness) for the ith wavenumber band is defined as Lr,VEH,i .
The total and effective roughness level for wavenumber band i (LR,tot,i ) is defined as the energy sum of the roughness levels of the rail and that of the wheel plus the A3(λ) contact filter to take into account the filtering effect of the contact patch between the rail and the wheel, and is in dB:

(2.3.7) 
where expressed as a function of the ith wavenumber band corresponding to the wavelength λ.
The contact filter depends on the rail and wheel type and the load.
The total effective roughness for the jth track section and each tth vehicle type at its corresponding v speed shall be used in the method.
Three speedindependent transfer functions, LH,TR,i LH,VEH,i and LH,VEH,SUP,i , are defined: the first for each jth track section and the second two for each tth vehicle type. They relate the total effective roughness level with the sound power of the track, the wheels and the superstructure respectively.
The superstructure contribution is considered only for freight wagons, therefore only for vehicle type ‘a’.
For rolling noise, therefore, the contributions from the track and from the vehicle are fully described by these transfer functions and by the total effective roughness level. When a train is idling, rolling noise shall be excluded.
For sound power per vehicle the rolling noise is calculated at axle height, and has as an input the total effective roughness level LR,TOT,i as a function of the vehicle speed v, the track, vehicle and superstructure transfer functions LH,TR,i , LH,VEH,i and LH,VEH,SUP,i , and the total number of axles Na :
for h = 1:
LW,0,TR,i = LR,TOT,i + LH,TR,i + 10 × lg(Na ) 
dB 
(2.3.8) 
LW,0,VEH,i = LR,TOT,i + LH,VEH,i + 10 × lg(Na ) 
dB 
(2.3.9) 
LW,0,VEHSUP,i = LR,TOT,i + LH,VEHSUP,i + 10 × lg(Na ) 
dB 
(2.3.10) 
where Na is the number of axles per vehicle for the tth vehicle type.
Figure [2.3.c]
Scheme of the use of the different roughness and transfer function definitions
Track transfer function for rail sleeper and ballast/slab emission
Htr
Vehicle transfer function for superstructure emission
Hsup
Vehicle transfer function for wheel and bogie emission
Hveh
Sound power of wheel and bogie emission
Lw,0,veh
Total effective roughness
Sound power of rail sleeper and ballast/slab emission
Lw,0,tr
Sound power of superstructure emission
Lw,0,sup
Rail roughness
rtr
Contact filter
Cf
Wheel roughness
rveh
Train speed
v
A minimum speed of 50 km/h (30 km/h only for trams and light metro) shall be used to determine the total effective roughness and therefore the sound power of the vehicles (this speed does not affect the vehicle flow calculation) to compensate for the potential error introduced by the simplification of rolling noise definition, braking noise definition and impact noise from crossings and switches definition.
Impact noise can be caused by crossings, switches and rail joints or points. It can vary in magnitude and can dominate rolling noise. Impact noise shall be considered for jointed tracks. For impact noise due to switches, crossings and joints in track sections with a speed of less than 50 km/h (30 km/h only for trams and light metro), since the minimum speed of 50 km/h (30 km/h only for trams and light metro) is used to include more effects according to the description of the rolling noise chapter, modelling shall be avoided. Impact noise modelling shall also be avoided under running condition c = 2 (idling).
Impact noise is included in the rolling noise term by (energy) adding a supplementary fictitious impact roughness level to the total effective roughness level on each specific jth track section where it is present. In this case a new LR,TOT + IMPACT,i shall be used in place of LR,TOT,i and it will then become:

dB 
(2.3.11) 
LR,IMPACT,i is a 1/3 octave band spectrum (as a function of frequency). To obtain this frequency spectrum, a spectrum is given as a function of wavelength λ and shall be converted to the required spectrum as a function of frequency using the relation λ = v/f, where f is the 1/3 octave band centre frequency in Hz and v is the sth vehicle speed of the tth vehicle type in km/h.
Impact noise will depend on the severity and number of impacts per unit length or joint density, so in the case where multiple impacts are given, the impact roughness level to be used in the equation above shall be calculated as follows:

dB 
(2.3.12) 
where LR,IMPACT–SINGLE,i is the impact roughness level as given for a single impact and nl is the joint density.
The default impact roughness level is given for a joint density nl = 0,01 m–1, which is one joint per each 100 m of track. Situations with different numbers of joints shall be approximated by adjusting the joint density nl . It should be noted that when modelling the track layout and segmentation, the rail joint density shall be taken into account, i.e. it may be necessary to take a separate source segment for a stretch of track with more joints. The LW,0 of track, wheel/bogie and superstructure contribution are incremented by means of the LR,IMPACT,i for +/– 50 m before and after the rail joint. In the case of a series of joints, the increase is extended to between – 50 m before the first joint and + 50 m after the last joint.
The applicability of these sound power spectra shall normally be verified onsite.
For jointed tracks, a default nl of 0,01 shall be used.
Curve squeal is a special source that is only relevant for curves and is therefore localised. As it can be significant, an appropriate description is required. Curve squeal is generally dependent on curvature, friction conditions, train speed and trackwheel geometry and dynamics. The emission level to be used is determined for curves with radius below or equal to 500 m and for sharper curves and branchouts of points with radii below 300 m. The noise emission should be specific to each type of rolling stock, as certain wheel and bogie types may be significantly less prone to squeal than others.
The applicability of these sound power spectra shall normally be verified onsite, especially for trams.
Taking a simple approach, squeal noise shall be considered by adding 8 dB for R < 300 m and 5 dB for 300 m < R < 500 m to the rolling noise sound power spectra for all frequencies. Squeal contribution shall be applied on railway track sections where the radius is within the ranges mentioned above for at least a 50 m length of track.
Although traction noise is generally specific to each characteristic operating condition amongst constant speed, deceleration, acceleration and idling, the only two conditions modelled are constant speed (that is valid as well when the train is decelerating or when it is accelerating) and idling. The source strength modelled only corresponds to maximum load conditions and this results in the quantities LW,0,const,i = LW,0,idling,i . Also, the LW,0,idling,i corresponds to the contribution of all physical sources of a given vehicle attributable to a specific height, as described in 2.3.1.
The LW,0,idling,i is expressed as a static noise source in the idling position, for the duration of the idling condition, and to be used modelled as a fixed point source as described in the following chapter for industrial noise. It shall be considered only if trains are idling for more than 0,5 hours.
These quantities can either be obtained from measurements of all sources at each operating condition, or the partial sources can be characterised individually, determining their parameter dependency and relative strength. This may be done by means of measurements on a stationary vehicle, by varying shaft speeds of the traction equipment, following ISO 3095:2005. As far as relevant, several traction noise sources have to be characterised which might not be all directly depending on the train speed:
— 
noise from the power train, such as diesel engines (including inlet, exhaust and engine block), gear transmission, electrical generators, mainly dependent on engine round per minute speed (rpm), and electrical sources such as converters, which may be mostly loaddependent, 
— 
noise from fans and cooling systems, depending on fan rpm; in some cases fans can be directly coupled to the driveline, 
— 
intermittent sources such as compressors, valves and others with a characteristic duration of operation and corresponding duty cycle correction for the noise emission. 
As each of these sources can behave differently at each operating condition, the traction noise shall be specified accordingly. The source strength is obtained from measurements under controlled conditions. In general, locomotives will tend to show more variation in loading as the number of vehicles hauled and thereby the power output can vary significantly, whereas fixed train formations such as electric motored units (EMUs), diesel motored units (DMUs) and highspeed trains have a better defined load.
There is no a priori attribution of the source sound power to the source heights, and this choice shall depend on the specific noise and vehicle assessed. It shall be modelled to be at source A (h = 1) and at source B (h = 2).
Aerodynamic noise is only relevant at high speeds above 200 km/h and therefore it should first be verified whether it is actually necessary for application purposes. If the rolling noise roughness and transfer functions are known, it can be extrapolated to higher speeds and a comparison can be made with existing highspeed data to check whether higher levels are produced by aerodynamic noise. If train speeds on a network are above 200 km/h but limited to 250 km/h, in some cases it may not be necessary to include aerodynamic noise, depending on the vehicle design.
The aerodynamic noise contribution is given as a function of speed:

dB 
For h = 1 
(2.3.13) 

dB 
For h = 2 
(2.3.14) 
where

v 0 is a speed at which aerodynamic noise is dominant and is fixed at 300 km/h 

LW,0,1,i is a reference sound power determined from two or more measurement points, for sources at known source heights, for example the first bogie 

LW,0,2,i is a reference sound power determined from two or more measurement points, for sources at known source heights, for example the pantograph recess heights 

α1,i is a coefficient determined from two or more measurement points, for sources at known source heights, for example the first bogie 

α2,i is a coefficient determined from two or more measurement points, for sources at known source heights, for example the pantograph recess heights. 
W,dir,hor,i in dB is given in the horizontal plane and by default can be assumed to be a dipole for rolling, impact (rail joints etc.), squeal, braking, fans and aerodynamic effects, given for each ΔLith frequency band by:
ΔLW,dir,hor,i = 10 × lg(0,01 + 0,99 · sin2 φ) 
(2.3.15) 
W,dir,ver,i in dB is given in the vertical plane for source A (h = 1), as a function of the centre band frequency ΔLfc,i of each ith frequency band, and for – π/2 < ψ < π/2 by:

(2.3.16) 
For source B (h = 2) for the aerodynamic effect:
ΔLW,dir,ver,i = 10 × lg(cos2 ψ) 
for ψ < 0 
(2.3.17) 
ΔLW,dir,ver,i = 0 elsewhere
Directivity ΔLdir,ver,i is not considered for source B (h = 2) for other effects, as omnidirectionality is assumed for these sources in this position.
2.3.3. Additional effects
In the case where the track section is on a bridge, it is necessary to consider the additional noise generated by the vibration of the bridge as a result of the excitation caused by the presence of the train. Because it is not simple to model the bridge emission as an additional source, given the complex shapes of bridges, an increase in the rolling noise is used to account for the bridge noise. The increase shall be modelled exclusively by adding a fixed increase in the noise sound power per each third octave band. The sound power of only the rolling noise is modified when considering the correction and the new LW,0,rolling–and–bridge,i shall be used instead of LW,0,rollingonly,i :
LW,0,rolling–and–bridge,i = LW,0,rolling–only,i + Cbridge 
dB 
(2.3.18) 
where Cbridge is a constant that depends on the bridge type, and LW,0,rolling–only,i is the rolling noise sound power on the given bridge that depends only on the vehicle and track properties.
Various sources like depots, loading/unloading areas, stations, bells, station loudspeakers, etc. can be present and are associated with the railway noise. These sources are to be treated as industrial noise sources (fixed noise sources) and shall be modelled, if relevant, according to the following chapter for industrial noise.
2.4. Industrial noise
2.4.1. Source description
The industrial sources are of very variable dimensions. They can be large industrial plants as well as small concentrated sources like small tools or operating machines used in factories. Therefore, it is necessary to use an appropriate modelling technique for the specific source under assessment. Depending on the dimensions and the way several single sources extend over an area, with each belonging to the same industrial site, these may be modelled as point sources, source lines or area sources. In practice, the calculations of the noise effect are always based on point sources, but several point sources can be used to represent a real complex source, which mainly extends over a line or an area.
The real sound sources are modelled by means of equivalent sound sources represented by one or more point sources so that the total sound power of the real source corresponds to the sum of the single sound powers attributed to the different point sources.
The general rules to be applied in defining the number of point sources to be used are:
— 
line or surface sources where the largest dimension is less than 1/2 of the distance between the source and the receiver can be modelled as single point sources, 
— 
sources where the largest dimension is more than 1/2 of the distance between the source and the receiver should be modelled as a series of incoherent point sources in a line or as a series of incoherent point sources over an area, such that for each of these sources the condition of 1/2 is fulfilled. The distribution over an area can include vertical distribution of point sources, 
— 
for sources where the largest dimensions in height are over 2 m or near the ground, special care should be administered to the height of the source. Doubling the number of sources, redistributing them only in the zcomponent, may not lead to a significantly better result for this source, 
— 
in the case of any source, doubling the number of sources over the source area (in all dimensions) may not lead to a significantly better result. 
The position of the equivalent sound sources cannot be fixed, given the large number of configurations that an industrial site can have. Best practices will normally apply.
The following information constitutes the complete set of input data for sound propagation calculations with the methods to be used for noise mapping:
— 
Emitted sound power level spectrum in octave bands 
— 
Working hours (day, evening, night, on a yearly averaged basis) 
— 
Location (coordinates x, y) and elevation (z) of the noise source 
— 
Type of source (point, line, area) 
— 
Dimensions and orientation 
— 
Operating conditions of the source 
— 
Directivity of the source. 
The point, line and area source sound power are required to be defined as:
— 
For a point source, sound power LW and directivity as a function of the three orthogonal coordinates (x, y, z); 
— 
Two types of source lines can be defined: 
— 
source lines representing conveyor belts, pipe lines, etc., sound power per metre length LW′ and directivity as a function of the two orthogonal coordinates to the axis of the source line, 
— 
source lines representing moving vehicles, each associated with sound power LW and directivity as a function of the two orthogonal coordinates to the axis of the source line and sound power per metre LW′ derived by means of the speed and number of vehicles travelling along this line during day, evening and night; The correction for the working hours, to be added to the source sound power to define the corrected sound power that is to be used for calculations over each time period, CW in dB is calculated as follows:
Where:

— 
For an area source, sound power per square metre LW/m2 , and no directivity (may be horizontal or vertical). 
The working hours are an essential input for the calculation of noise levels. The working hours shall be given for the day, evening and night period and, if the propagation is using different meteorological classes defined during each of the day, night and evening periods, then a finer distribution of the working hours shall be given in subperiods matching the distribution of meteorological classes. This information shall be based on a yearly average.
The correction for the working hours, to be added to the source sound power to define the corrected sound power that shall be used for calculations over each time period, CW in dB is calculated as follows:

(2.4.2) 
where

T is the active source time per period based on a yearly averaged situation, in hours; 

Tref is the reference period of time in hours (e.g. day is 12 hours, evening is 4 hours, night is 8 hours). 
For the more dominant sources, the yearly average working hours correction shall be estimated at least within 0,5 dB tolerance in order to achieve an acceptable accuracy (this is equivalent to an uncertainty of less than 10 % in the definition of the active period of the source).
The source directivity is strongly related to the position of the equivalent sound source next to nearby surfaces. Because the propagation method considers the reflection of the nearby surface as well its sound absorption, it is necessary to consider carefully the location of the nearby surfaces. In general, these two cases will always be distinguished:
— 
a source sound power and directivity is determined and given relative to a certain real source when this is in free field (excluding the terrain effect). This is in agreement with the definitions concerning the propagation, if it is assumed that there is no nearby surface less than 0,01 m from the source and surfaces at 0,01 m or more are included in the calculation of the propagation, 
— 
a source sound power and directivity is determined and given relative to a certain real source when this is placed in a specific location and therefore the source sound power and directivity is in fact an ‘equivalent’ one, since it includes the modelling of the effect of the nearby surfaces. This is defined in ‘semifree field’ according to the definitions concerning the propagation. In this case, the nearby surfaces modelled shall be excluded from the calculation of propagation. 
The directivity shall be expressed in the calculation as a factor ΔLW,dir,xyz (x, y, z) to be added to the sound power to obtain the right directional sound power of a reference sound source seen by the sound propagation in the direction given. The factor can be given as a function of the direction vector defined by (x,y,z) with
. This directivity can also be expressed by means of other coordinate systems such as angular coordinate systems.2.5. Calculation of noise propagation for road, railway, industrial sources.
2.5.1. Scope and applicability of the method
This document specifies a method for calculating the attenuation of noise during its outdoor propagation. Knowing the characteristics of the source, this method predicts the equivalent continuous sound pressure level at a receiver point corresponding to two particular types of atmospheric conditions:
— 
downwardrefraction propagation conditions (positive vertical gradient of effective sound celerity) from the source to the receiver, 
— 
homogeneous atmospheric conditions (null vertical gradient of effective sound celerity) over the entire area of propagation. 
The method of calculation described in this document applies to industrial infrastructures and land transport infrastructures. It therefore applies in particular to road and railway infrastructures. Aircraft transport is included in the scope of the method only for the noise produced during ground operations and excludes takeoff and landing.
Industrial infrastructures that emit impulsive or strong tonal noises as described in ISO 19962:2007 do not fall within the scope of this method.
The method of calculation does not provide results in upwardrefraction propagation conditions (negative vertical gradient of effective sound speed) but these conditions are approximated by homogeneous conditions when computing Lden.
To calculate the attenuation due to atmospheric absorption in the case of transport infrastructure, the temperature and humidity conditions are calculated according to ISO 96131:1996.
The method provides results per octave band, from 63 Hz to 8 000 Hz. The calculations are made for each of the centre frequencies.
Partial covers and obstacles sloping, when modelled, more than 15° in relation to the vertical are out of the scope of this calculation method.
A single screen is calculated as a single diffraction calculation, two or more screens in a single path are treated as a subsequent set of single diffractions by applying the procedure described further.
2.5.2. Definitions used
All distances, heights, dimensions and altitudes used in this document are expressed in metres (m).
The notation MN stands for the distance in 3 dimensions (3D) between the points M and N, measured according to a straight line joining these points.
The notation M̂N stands for the curved path length between the points M and N, in favourable conditions.
It is customary for real heights to be measured vertically in a direction perpendicular to the horizontal plane. Heights of points above the local ground are denoted h, absolute heights of points and absolute height of the ground are to be noted by the letter H.
To take into account the actual relief of the land along a propagation path, the notion of ‘equivalent height’ is introduced, to be noted by the letter z. This substitutes real heights in the ground effect equations.
The sound levels, noted by the capital letter L, are expressed in decibels (dB) per frequency band when index A is omitted. The sound levels in decibels dB(A) are given the index A.
The sum of the sound levels due to mutually incoherent sources is noted by the sign ⊕ in accordance with the following definition:

(2.5.1) 
2.5.3. Geometrical considerations
Real sources are described by a set of point sources or, in the case of railway traffic or road traffic, by incoherent source lines. The propagation method assumes that line or area sources have previously been split up to be represented by a series of equivalent point sources. This may have occurred as preprocessing of the source data, or may occur within the pathfinder component of the calculation software. The means by which this has occurred is outside the scope of the current methodology.
The method operates on a geometrical model consisting of a set of connected ground and obstacles surfaces. A vertical propagation path is deployed on one or more vertical planes with respect to the horizontal plane. For trajectories including reflections onto vertical surfaces not orthogonal to the incident plane, another vertical plane is subsequently considered including the reflected part of the propagation path. In these cases, where more vertical planes are used to describe the entire trajectory from the source to the receiver, the vertical planes are then flattened, like an unfolding Chinese screen.
The equivalent heights are obtained from the mean ground plane between the source and the receiver. This replaces the actual ground with a fictitious plane representing the mean profile of the land.
Figure 2.5.a
Equivalent heights in relation to the ground
1 
: 
Actual relief 
2 
: 
Mean plane 
The equivalent height of a point is its orthogonal height in relation to the mean ground plane. The equivalent source height zs and the equivalent receiver height zr can therefore be defined. The distance between the source and receiver in projection over the mean ground plane is noted by dp .
If the equivalent height of a point becomes negative, i.e. if the point is located below the mean ground plane, a null height is retained, and the equivalent point is then identical with its possible image.
In the plane of the path, the topography (including terrain, mounds, embankments and other manmade obstacles, buildings, …) may be described by an ordered set of discrete points (xk, Hk ); k є {1,…,n}. This set of points defines a polyline, or equivalently, a sequence of straight segments Hk = akx + bk, x є [xk, xk + 1 ]; k є {1,…n}, where:

ak = (Hk + 1 – Hk )/(xk + 1 – xk ) 
(2.5.2) 
bk = (Hk · xk + 1 – Hk + 1 · xk )/(xk + 1 – xk ) 
The mean plane is represented by the straight line Z = ax + b; x є [x 1, xn ], which is adjusted to the polyline by means of a leastsquare approximation. The equation of the mean line can be worked out analytically.
Using:


(2.5.3) 

The coefficients of the straight line are given by:


(2.5.4) 

Where segments with xk + 1 = xk shall be ignored when evaluating eq. 2.5.3.
Contributions from reflections are taken into account by the introduction of image sources as described further.
2.5.4. Sound propagation model
For a receiver R the calculations are made according to the following steps:
(1) 
on each propagation path:

(2) 
accumulation of the longterm sound levels for all paths affecting a specific receiver, therefore allowing the total sound level to be calculated at the receiver point. 
It should be noted that only the attenuations due to the ground effect (Aground ) and diffraction (Adif ) are affected by meteorological conditions.
2.5.5. Calculation process
For a point source S of directional sound power Lw,0,dir and for a given frequency band, the equivalent continuous sound pressure level at a receiver point R in given atmospheric conditions is obtained according to the equations following below.
LF = LW,0,dir – AF 
(2.5.5) 
The term AF represents the total attenuation along the propagation path in favourable conditions, and is broken down as follows:
LF = Adiv + Aatm + Aboundary,F 
(2.5.6) 
where

Adiv is the attenuation due to geometrical divergence; 

Aatm is the attenuation due to atmospheric absorption; 

Aboundary,F is the attenuation due to the boundary of the propagation medium in favourable conditions. It may contain the following terms:

For a given path and frequency band, the following two scenarios are possible:
— 
either Aground,F is calculated with no diffraction (Adif,F = 0 dB) and Aboundary,F = Aground,F ; 
— 
or Adif,F is calculated. The ground effect is taken into account in the Adif,F equation itself (Aground,F = 0 dB). This therefore gives Aboundary,F = Adif,F . 
The procedure is strictly identical to the case of favourable conditions presented in the previous section.
LH = LW,0,dir – AH 
(2.5.7) 
The term AH represents the total attenuation along the propagation path in homogeneous conditions and is broken down as follows:
AH = Adiv + Aatm + Aboundary,H 
(2.5.8) 
where

Adiv is the attenuation due to geometrical divergence; 

Αatm is the attenuation due to atmospheric absorption; 

Aboundary,H is the attenuation due to the boundary of the propagation medium in homogeneous conditions. It may contain the following terms:

For a given path and frequency band, the following two scenarios are possible:
— 
either Αground,H (Adif,H = 0 dB) is calculated with no diffraction and Aboundary,H = Αground,H ; 
— 
or Adif,H (Αground,H = 0 dB) is calculated. The ground effect is taken into account in the Adif,H equation itself. This therefore gives Aboundary,H = Adif,H 
Inside urban areas, a statistical approach to the calculation of the sound propagation behind the first line of buildings is also allowed, provided that such a method is duly documented, including relevant information on the quality of the method. This method may replace the calculation of the Aboundary,H and Aboundary,F by an approximation of the total attenuation for the direct path and all reflections. The calculation will be based on the average building density and the average height of all buildings in the area.
The ‘longterm’ sound level along a path starting from a given point source is obtained from the logarithmic sum of the weighted sound energy in homogeneous conditions and the sound energy in favourable conditions.
These sound levels are weighted by the mean occurrence p of favourable conditions in the direction of the path (S,R):

(2.5.9) 
NB: The occurrence values for p are expressed in percentages. So for example, if the occurrence value is 82 %, equation (2.5.9) would have p = 0,82.
The total longterm sound level at the receiver for a frequency band is obtained by energy summing contributions from all N paths, all types included:

(2.5.10) 
where
n is the index of the paths between S and R.
Taking reflections into account by means of image sources is described further. The percentage of occurrences of favourable conditions in the case of a path reflected on a vertical obstacle is taken to be identical to the occurrence of the direct path.
If S′ is the image source of S, then the occurrence p′ of the path (S′,R) is taken to be equal to the occurrence p of the path (Si,R).
The total sound level in decibels A (dBA) is obtained by summing levels in each frequency band:

(2.5.11) 
where i is the index of the frequency band. AWC is the Aweighting correction according to the international standard IEC 616721:2003.
This level LAeq,LT constitutes the final result, i.e. the longterm Aweighted sound pressure level at the receiver point on a specific reference time interval (e.g. day or evening, or night or a shorter time during day, evening or night).
2.5.6. Calculation of noise propagation for road, railway, industrial sources.
The attenuation due to geometrical divergence, Adiv, corresponds to a reduction in the sound level due to the propagation distance. For a point sound source in free field, the attenuation in dB is given by:
Adiv = 20 × lg(d) + 11 
(2.5.12) 
where d is the direct 3D slant distance between the source and the receiver.
The attenuation due to atmospheric absorption Aatm during propagation over a distance d is given in dB by the equation:
Aatm = αatm · d/1 000 
(2.5.13) 
where

d is the direct 3D slant distance between the source and the receiver in m; 

αatm is the atmospheric attenuation coefficient in dB/km at the nominal centre frequency for each frequency band, in accordance with ISO 96131. 
The values of the αatm coefficient are given for a temperature of 15 °C, a relative humidity of 70 % and an atmospheric pressure of 101 325 Pa. They are calculated with the exact centre frequencies of the frequency band. These values comply with ISO 96131. Meteorological average over the long term shall be used if meteorological data is available.
The attenuation due to the ground effect is mainly the result of the interference between the reflected sound and the sound that is propagated directly from the source to the receiver. It is physically linked to the acoustic absorption of the ground above which the sound wave is propagated. However, it is also significantly dependent on atmospheric conditions during propagation, as ray bending modifies the height of the path above the ground and makes the ground effects and land located near the source more or less significant.
In case the propagation between the source and the receiver is affected by any obstacle in the propagation plane, the ground effect is calculated separately on the source and receiver side. In this case, zs and zr refer to the equivalent source and/or receiver position as indicated further where the calculation of the diffraction Adif is presented.
The acoustic absorption properties of the ground are mainly linked to its porosity. Compact ground is generally reflective and porous ground is absorbent.
For operational calculation requirements, the acoustic absorption of a ground is represented by a dimensionless coefficient G, between 0 and 1. G is independent of the frequency. Table 2.5.a gives the G values for the ground outdoors. In general, the average of the coefficient G over a path takes values between 0 and 1.
Table 2.5.a
G values for different types of ground
Description 
Type 
(kPa · s/m2) 
G value 
Very soft (snow or mosslike) 
A 
12,5 
1 
Soft forest floor (short, dense heatherlike or thick moss) 
B 
31,5 
1 
Uncompacted, loose ground (turf, grass, loose soil) 
C 
80 
1 
Normal uncompacted ground (forest floors, pasture field) 
D 
200 
1 
Compacted field and gravel (compacted lawns, park area) 
E 
500 
0,7 
Compacted dense ground (gravel road, car park) 
F 
2 000 
0,3 
Hard surfaces (most normal asphalt, concrete) 
G 
20 000 
0 
Very hard and dense surfaces (dense asphalt, concrete, water) 
H 
200 000 
0 
Gpath is defined as the fraction of absorbent ground present over the entire path covered.
When the source and receiver are closeby so that dp ≤ 30(zs + zr ), the distinction between the type of ground located near the source and the type of ground located near the receiver is negligible. To take this comment into account, the ground factor Gpath is therefore ultimately corrected as follows:
G′path = 

if dp ≤ 30(zs + zr ) 
(2.5.14) 
Gpath 
otherwise 
where Gs is the ground factor of the source area. Gs = 0 for road platforms (4), slab tracks. Gs = 1 for rail tracks on ballast. There is no general answer in the case of industrial sources and plants.
G may be linked to the flow resistivity.
Figure 2.5.b
Determination of the ground coefficient Gpath over a propagation path
Road G=0
Parking G=0
Cultivated fields G=1
Gardens G=1
The following two subsections on calculations in homogeneous and favourable conditions introduce the generic Ḡw and Ḡm notations for the absorption of the ground. Table 2.5.b gives the correspondence between these notations and the Gpath and G′path variables.
Table 2.5.b
Correspondence between Ḡw and Ḡm and (Gpath , G′path )

Homogeneous conditions 
Favourable conditions 

Aground 
Δground(S,O) 
Δground(O,R) 
Ag round 
Δground(S,O) 
Δground(O,R) 

Ḡw 
G′path 
Gpath 

Ḡm 
G′path 
Gpath 
G′path 
Gpath 
The attenuation due to the ground effect in homogeneous conditions is calculated according to the following equations:
if Gpath ≠ 0
