Xuetao Zhang
Energy Technology SP Report 2012:17SP T
ech
ni
ca
l Re
se
arch
I
nstitu
te of Sweden
Measuring and Modelling Snowmobile
Noise
Abstract
Two typical snowmobiles, one with two-stroke engine (113 kW) and the other with four-stroke engine (88.4 kW), have been measured under different operating conditions:
stationary and at the idling engine rpm, or at a high rpm just below the engagement, or at the one in the between;
stationary and with engine rpm swept from the idling rpm to the high rpm;
cruise-by at a speed of 25 km/h, 50 km/h or 75 km/h;
pass-by at full throttle.
Data were collected by recording the time history of the noise emission, at four receiver positions. This time-history recording of the noise emission is of a great advantage in post-analysis because it saves the full information of the noise emission. Then any desired quantity such as SEL, Leq, or LpASmax, in any time period, becomes available in a post-processing. The four receiver positions are (d/h): 7.5/1.2 m which is the standard position in measuring road traffic noise and railway noise, 15/1.2 m which is the standard position according to the North-America test method for acoustical certification of snowmobiles, 25/1.2 m and 25/3.5 m which are the two optional standard positions in measuring road traffic noise and railway noise and which can also be used for evaluating ground effect on sound propagation.
Also measured is the sound propagation above snow, by using a point sound source. Moreover, the noise emission data of the four snowmobiles collected in 2009 by SP Acoustics, and the snowmobile noise data found in literature, are referred to when making the data analysis.
The main findings, outputs and conclusions are:
Engine/exhaust noise is the primary snowmobile noise. The speed dependence of it can be described by 25*log10(V), where V is the speed;
Non-engine noise is negligible when a snowmobile drives on a soft trail;
On a trail of compacted snow, non-engine noise can become comparable to the engine noise, above 300 Hz, for some snowmobile models. Non-engine noise can become more serious if a snowmobile drives on an ice-hard surface;
The modified one-parameter impedance model (see section 3 for the definition) can be used to handle sound propagation above snow;
A simple source model of snowmobile noise has been worked out. Then, the sound power level of snowmobile noise at a given speed becomes predictable, for the two typical snowmobiles tested within this project. The North-America noise indicator, LpASmax, has also been related to in this source modelling;
Based on a study of available noise emission data of snowmobiles, the tested two snowmobiles are found to be representative. Therefore, this simple source model of snowmobile noise is believed to be generally applicable;
Sound attenuation in forest has been proposed based on a literature study together with a reference to Finnish data;
Sound attenuation of snowmobile noise has been calculated and is provided in tabular values for four typical terrains and up to 10 km. The four typical terrains are: a flat and open field covered by soft snow, a flat and open field of an ice-hard surface such as a frozen lake, a forest of a flat ground covered by soft snow, and an open mountain area with a grade of 30% (i.e. a raise of 3 m for every 10 m of run) covered by soft snow;
Noise impact of snowmobiles has also been calculated for the four typical terrains and up to 10 km;
Modelling the non-engine noise is left to future research, because it requires an extensive measurement study on a variety of snowmobile types combined with several typical trail conditions (i.e. soft, moderate hard and ice-hard trails), which has not been covered by this small project.
It was found that the noise indicators, LpASmax and LeqA, are not simply related. If LpASmax limit value will be lowered by dL, the consequential reduction in
eqA
L will be less and possibly be about 0.6dL ~ 0.7dL (if there is no change in the traffic volume).
Key words: snowmobile, noise emission, measurement, engine noise, non-engine noise, source modelling, flow resistivity, sound attenuation in forest, sound attenuation above snow
SP Sveriges Tekniska Forskningsinstitut
SP Technical Research Institute of Sweden
SP Report :2012:17 ISBN 978-91-87017-37-7 ISSN 0284-5172
Contents
Abstract
3
Contents
5
Preface 7
Sammanfattning
8
1
Introduction
11
2
Measurement of snowmobile noise
13
2.1 Measurement site and measurement conditions 13
2.2 The tested snowmobiles 14
2.3 Measurements 15
2.3.1 Sound propagation of a point source above snow 15
2.3.2 Snowmobile stationary noise 15
2.3.3 Snowmobile pass-by noise 15
3
Sound propagation above snow
17
3.1 The impedance model 17
3.2 Effective flow resistivity of a snow covered field 18 3.3 Sound propagation of the point source at the measurement site 19
4
Analysis of the measurements
23
4.1 The engine noise of the two snowmobiles 23
4.1.1 Measurement arrangement 23
4.1.2 The noise spectrum of snowmobile engine noise 24 4.1.3 The rpm dependence of snowmobile engine noise 25 4.1.4 The source height of snowmobile engine noise 27 4.2 Snowmobile pass-by noise: under a constant speed 30
4.2.1 Measurement arrangement 30
4.2.2 The noise spectrum of snowmobile pass-by noise 31 4.2.3 The speed dependence of snowmobile pass-by noise 36 4.3 Snowmobile pass-by noise: at full throttle 38
4.3.1 The North-America method SAE J192 38
4.3.2 Measurement arrangement 39
4.3.3 The noise spectrum 39
4.4 Referring to some earlier measurements 41
4.4.1 Non-engine noise 42
4.4.2 Directivity of engine noise 46
5
Modelling snowmobile noise
47
5.1 Literature study 47
5.2 Noise source and source height 48
5.3 Sound power and speed dependence 48
5.4 Other factors 51
5.5 Exhaust modification 51
6
Noise impact of snowmobiles
53
6.1 Propagation attenuation of snowmobile noise 53
6.2 Noise impact of snowmobiles 58
6.2.1 Instantaneous SPL caused by a single snowmobile 58 6.2.2 Equivalent SPL of snowmobile traffic noise 62
6.2.3 Consideration of noise spectrum components 70
6.3 About the new noise limit values 71
7
Summarization and discussion
73
Preface
This project is funded by the Swedish Transport Agency (Transportstyrelsen), with the registration number TSG 2011-6321.
The project has been carried out by Dr. Xuetao Zhang, Acoustics Section, SP Technical Research Institute of Sweden. Dr. Krister Larsson (SP) is the project leader.
Dr. Krister Larsson, Dr. Hans Jonasson (SP) and Stefan Frisk (SMP Swedish Machinery Testing Institute) engaged in the measurement planning. Stefan Frisk and Joachim Grönlund (SMP) took part in carrying out the measurement. Stefan Frisk provided valuable knowledge on snowmobiles.
All these supports are greatly appreciated!
Xuetao Zhang
Sammanfattning
Två typiska snöskotrar, en med tvåtaktsmotor (113 kW) och en med fyrtaktsmotor (88,4 kW), har testats under olika driftsförhållanden:
Stillastående på tomgång, eller vid ett högt motorvarvtal under ingrepp, eller vid ett motorvarvtal mellan dessa;
Stillastående med motorvarvtalet svept mellan tomgång och det högre varvtalet;
förbipassage med en hastighet av 25 km/t, 50 km/t och 75 km/t;
förbipassage vid full acceleration.
Data samlades in genom att registrera tidshistoriken av ljudet vid fyra mottagarpositioner. En inspelning av ljudet är en stor fördel i analysen av data, eftersom all information har sparats och önskade mått som SEL, Leq eller LpASmaxblir tillgängliga i varje tidsperiod. De fyra mottagarpositioner är (d / h): 7.5/1.2 m, som är standardpositionen vid mätning av vägtrafikbuller och järnvägstrafikbuller, 15/1.2 m som är standardposition enligt Nord-Amerika testmetoden för certifiering av en snöskoter, 25/1.2 m och 25/3.5 m som är de två valfria standard positioner vid mätning av vägtrafikbuller och järnvägstrafikbuller och som också kan användas för att utvärdera markeffekter på ljudutbredning.
Dessutom mättes ljudutbredningen över snö med hjälp av en punktljudkälla. Dessutom utnyttjades data för fyra snöskotrar som samlades in 2009 av SP Akustik, samt bullerdata som finns i litteraturen när dataanalys gjordes.
De viktigaste resultaten och slutsatserna är:
Motor/avgasljudet är den största bullerkällan hos snöskotrar. Hastighetsberoendet kan beskrivas med, baserat på tillgängliga data, 25 * log10 (V), där V är
hastigheten i km/t;
Icke-motorljud är försumbart när en snöskoter kör på ett mjukt underlag;
På ett underlag med packad snö, kan icke-motorljudet bli jämförbart med motorbullret över 300 Hz, för vissa snöskoter modeller. Om en snöskoter driver på en is-hård yta kan icke-motorljudet bli högre än motorljudet;
Den modifierade enparameter impedans modellen (se avsnitt 3 för definitionen) kan användas för att hantera ljudutbredning över snö;
En enkel källmodell av snöskoterbuller har utarbetats. Med hjälp av den kan ljudeffektnivå av snöskoterbuller vid en given hastighet förutsägas för de två typiska skotrarna som testats inom detta projekt. Sambandet med det
Nordamerikanska bullermåttet har också utarbetats i denna källmodell;
De två testade snöskotrar har befunnits vara representativa för snöskotrar baserat på en studie av tillgängliga data. Därför tros denna enkla källmodell av
snöskoterbuller vara allmänt tillämplig;
Ljuddämpning i skogen har föreslagits utifrån en litteraturstudie tillsammans med en hänvisning till uppgifter från finska studier;
Ljuddämpning av snöskoterbuller har beräknats och anges i tabellform för fyra typiska terränger upp till 10 km avstånd. De fyra typiska terrängerna är: ett platt och öppet område som täcks av mjuk snö, ett platt och öppet fält av en is-hård yta, t.ex. en frusen sjö, en skog med plan mark täckt av mjuk snö, och ett öppet fjällområde med en lutning av 30% (dvs. en höjning på 3 meter för varje 10 m körning), som täcks av mjuk snö;
Bullernivåer från snöskotrar har också beräknats för de fyra typiska terrängerna upp till 10 km avstånd;
Modellering av icke-motorljudet lämnas till framtida forskning, eftersom det kräver en omfattande mätstudie på en variationnav olika snöskotrar i kombination
med flera typiska markförhållanden (dvs. mjuka, måttliga hårda och is-hårda spår), som inte omfattas av detta projektet.
Det visade sig att bullermåtten, LpASmax och LeqA, inte är relaterade på ett enkelt sätt. Om gränsvärdet LpASmax sänks med dL, komner minskningen av ljudnivån
eqA
L att vara mindre, omkring 0.6dL ~ 0.7dL, om det inte finns någon förändring i trafikvolymen.
1
Introduction
Nowadays environmental noise has become a social concern. One of such environmental noise problems is acoustical pollution in noise-free mountain areas. Terrain vehicles including snowmobiles are one of a few kinds of noise sources in these mountain areas. In Sweden, about 8000 snowmobiles are sold every year; and, there are in total nearly 150 000 snowmobiles in use. The Swedish speed limit for snowmobiles is 70 km/h [1, 18].
During 2009 the European Commission was expected to elaborate an addendum to the Directive 2000/14/EC on noise emission from equipment used outdoors. As at present noise emission from snowmobiles has not been regulated, in order to have the noise emission limit included in the ongoing revision of the directive, Sweden is a driving force in terms of noise requirements on snowmobiles. According to the North-America
regulation, the noise limit for snowmobiles is 78 dB(A) (under the prescribed test conditions). Sweden’s ambition is to set new limit values in two steps, with levels 75 dB(A) to 2017 and 73 dB(A) to 2021. For Sweden to work out a proposal of an
international level a solid knowledge base is required. This knowledge is also needed as an assistant to achieve the national target, a magnificent mountain landscape.
The solid knowledge base mentioned above consists of two parts: the sound attenuation above snow for typical terrains as well as a proper description of the sound power of snowmobile noise. Within this project, which was ordered by the Swedish Transport Agency, these two tasks have been worked on.
For sound propagation above snow, a literature study has been made in order to find a proper method to handle such situations. Moreover, for validating the chosen method, a well designed propagation measurement using a point sound source has been carried out. As described in section 3, the two-parameter impedance model proposed in [2-3], or, the modified one-parameter impedance model described in [4] which is in fact the simplified version of the two-parameter model, has been found most suitable in describing sound propagation above snow.
Aiming at determining the sound power of snowmobile noise together with its speed dependence, a systematic measurement investigation has been made. Two typical
snowmobiles, one with two-stroke engine (113 kW) and the other with four-stroke engine (88.4 kW), have been measured under different operating conditions: stationary, cruise-by and pass-by at full throttle. The measurement arrangement and measurement details are described in section 2 while the technical details of the measurement can also be found in section 4. In section 4, data analysis was made, which also refers to the noise emission data of the four snowmobiles measured in 2009 by SP. The engine noise and non-engine noise, as well as the speed dependence, have been studied. Thus, based on this analysis and other noise data found in literature, a source model of snowmobile noise has been worked out. The process to reach this source model is described in section 5.
In section 6 sound attenuation above snow for four typical terrains is provided in tabular values, wherein a neutral weather condition as well as a downwind condition has been assumed. These sound attenuation values, together with the source model of snowmobile noise, provide people a great convenience in evaluating the noise impact of snowmobiles. Moreover, example noise impacts of snowmobiles, the instantaneous SPL of a single snowmobile and the equivalent SPL of snowmobile passing by, have also been calculated and provided in tabular values, for the four typical terrains.
2
Measurement of snowmobile noise
The purpose of the measurements is to collect noise emission data of snowmobiles, to work out a source model for describing snowmobile noise.
The measurement was carried out in Umeå, February 23, 2012.
2.1
Measurement site and measurement conditions
The measurement site is Röbäcksdalen in Umeå, which has a free radius of about 500 m. Two photos given in Figure 2-1 provide a view of the measurement site.
Figure 2-1. The microphone positions and the travel path at the measurement site
Röbäcksdalen, Umeå.
The snow depth is about 53 cm at the microphone positions and about 35 cm along the travel path. The temperature is 1 oC and wind speed is about 3~5 m/s, roughly in 45o from the travel path down to the microphone stands.
2.2
The tested snowmobiles
Two snowmobiles of popular brands are tested: Arctic Cat 2012 XF 800 HIGH
COUNTRY Sno Pro OS BLK (hereafter, the Arctic Cat), of two stroke engine and 2012 Yamaha RS Venture GT (hereafter, the Yamaha), of 4 stroke engine. Descriptions of these two snowmobiles are given below.
1. Snowmobile Arctic Cat 2012 XF 800 HIGH COUNTRY SP OS BLK
Chassis no Engine Track width
(mm)
Total weight (kg) 4UF12SNWXCT120214 Two-stroke, two cylinder,
liquid-cooled, 113 kW (~152 hp)
381 254
2. Snowmobile 2012 Yamaha RS Venture GT
Chassis no Engine Track width
(mm)
Total weight (kg) Not recorded Four-stroke, three-cylinder,
twin-valve, liquid-cooled, 88.4 kW (~120 hp)
2.3
Measurements
Three types of measurements have been carried out, as described in the following.
The measurement system is 01dB with four channels.
2.3.1
Sound propagation of a point source above snow
This measurement is to determine the effective flow resistivity of the snow covered ground, which is the necessary parameter in handling sound propagation above snow.A sound source, Brüel & Kjær HP 1001 type (shown in Figure 3-2), was used. The four microphones positions are (d/h): 7.5/1.2, 15/1.2, 25/1.2 and 25/3.5 (m). These
microphone positions have also been used for the pass-by measurements.
The measurement details and the data analysis are described in section 3.
2.3.2
Snowmobile stationary noise
There are no doubts that engine noise is the most important component of snowmobile noise. However, in general, other noise types such as transmission noise, suspension noise and track noise may also contribute to the total snowmobile noise. Therefore, this type of measurements is for collecting engine noise data of the two snowmobiles.
The four microphone positions are (d/h): 7.5/1.2, 7.5/2 and 7.5/3.5 m for the three microphones along the normal line passing through the centre of the snowmobile (here 7.5 m is the distance between the microphone position and the snowmobile axis line); the fourth microphone has a position similar to the first one (7.5/1.2 m) but one snowmobile length (3.3 m) shifted to the rear side from the normal line. This fourth position is for checking out if there is an important effect of horizontal directivity.
To find out the rpm dependence, engine noise has been measured under three rpm values: the idling rpm, a high rpm which is below but close to the engagement rpm1 and a rpm in between. Engine noise under an acceleration operating condition was also measured. However, as only a gentle acceleration was possible to avoid engagement, it in fact becomes a sweep operating condition (i.e. during the measurement the rpm value varies from the idling rpm to the high rpm).
The data analysis for this measurement type is given in sub-section 4.1.
2.3.3
Snowmobile pass-by noise
Four microphone positions are (d/h): 7.5/1.2, 15/1.2, 25/1.2 and 25/3.5 m.
The first microphone position, 7.5/1.2 m, is the standard position used for measuring road vehicle noise and railway noise. The second microphone position, 15/1.2 m, is close to the standard position, 50/4 feet, which is used according to the North-America standard SAE J192 [5]. The use of this position is aiming at understanding the North-America noise limits. The last two microphone positions, 25/1.2 and 25/3.5 m, are used for validating the impedance model in determining sound attenuation above snow.
Two types of pass-by measurements are carried out: pass-by under a constant speed and pass-by at full throttle.
For pass-by under a constant speed, three speeds were taken: 25 km/h, 50 km/h and 75 km/h. This measurement type is for collecting noise emission data for determining the total sound power of snowmobile noise and also for determining the speed dependence of the noise emission. The measurement arrangement and the analysis of the noise emission data are given in sub-section 4.2.
The pass-by at full throttle is for collecting the noise data under the conditions the same as those described in SAE J192, in order to understand the North-America noise limits. The measurement arrangement and the data analysis are given in sub-section 4.3.
3
Sound propagation above snow
The Nord2000 propagation model was used to calculate sound propagation outdoors. As the impedance model for snow has not been implemented in the latest calculation
software, compro18e, this function has been added according to Eqs. (3-1)-(3-3) given in sub-section 3.1.
In the following sub-sections the impedance model, the method to determine the snow flow resistivity and the snow flow resistivity of the measurement site in Umeå will be described.
3.1
The impedance model
To handle sound propagation above a ground of a finite impedance, a proper impedance model is needed. As proposed in the relevant Nordtest method [4], the modified one-parameter model shall be used for a snow covered ground
Zn Zn1*icot
Lk , (3-1)
0.70 0.591000
3
.
10
1000
8
.
10
1
2
f
i
f
c
f
k
, (3-2) 73 . 0 75 . 0 1 1000 9 . 11 1000 08 . 9 1
f i f Zn , (3-3)where L is the snow depth. The Eq. (3-3) is the well known one-parameter impedance model developed by Delany and Bazley [6].
101 102 103 104 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 Z(Nordtest) - Z(Nicolas) real imag
Figure 3-1. The difference of the model impedance values for a snow covered field of the
thickness 5 cm and the flow resistivity 16 kPa s m-2. The two models are described in [2, 4].
This modified one-parameter impedance model differs slightly from a more general two-parameter impedance model [2, 3]. However, calculation shown in Figure 3-1 indicates that the difference between two models’ outputs is negligible above 100 Hz and still very small under 100 Hz. Therefore, it is considered proper to use the modified one-parameter model to handle sound propagation above snow.
3.2
Effective flow resistivity of a snow covered field
According to the extensive investigation on sound propagation above a finite layer of snow [2], it is concluded that the extended model in terms of a flow resistivity and a snow thickness must be used.The effective flow resistivity of a snow covered field varies with the snow condition. The case study presented in [2] suggests the following:
1. For fresh or loose snow, the effective flow resistivity is often within the range of about 10~20 kPa s m-2;
2. The hardening of snow by wind and/or by sunshine can increase the effective flow resistivity up to 150 kPa s m-2;
3. Fresh/loose snow of a thickness over 20 cm is acoustically equivalent to fresh/loose snow of infinite thickness.
4. For a layer of fresh/loose snow above a layer of hardened snow, the effective flow resistivity and the effective snow thickness can in many cases be determined by the layer of fresh/loose snow. However, there is an exception as stated in the following.
5. More complicated, if a few cm of fresh snow fell above a few cm of slightly hardened snow, the effective snow thickness will always be less than the total thickness of the two snow layers. The effective flow resistivity is between that for the fresh snow and that for the hardened snow.
For general applications, it is not reasonable to require the information of detailed snow conditions over a large area. Therefore, the complicated situation discussed above, the last point, will not be considered.
The relevant Harmonoise proposal [7] is: (1) the range of effective flow resistivity for snow is between 1.3 and 50 kPa s m-2; (2) the average is 29 kPa s m-2.
The method to determine the effective flow resistivity of a snow covered field is described in the following:
1) To use the modified one-parameter impedance model (sub-section 3.1). 2) Except the complicated case mentioned above, the snow depth is the real snow
depth if only one layer of snow is concerned, or the depth of the upper layer of fresh/loose snow if there is also an under layer of hardened snow.
3) To adjust the effective flow resistivity to predict SPL best comparable to the measured SPL at the receiving position(s).
4) For the complicated case mentioned above, following the method described in [2], one should first determine the effective snow depth by positioning the first minimum in spectrum (of the SPL at the receiver) at the right frequency, which always gives the effective snow depth between the total one of the two snow layers and the one of the fresh/loose snow layer. The next step is to adjust the effective flow resistivity to predict the level at this minimum. Thus, the effective snow depth and the effective snow flow resistivity are determined (while only valid for this receiver position).
3.3
Sound propagation of the point source at the
measurement site
The standard sound source, Brüel & Kjær HP 1001 type, was used as the point sound source, see Figure 3-2. Its acoustical centre is estimated to be located at its geometrical centre. Three source heights, 0.1 m, 0.5 m and 1 m (above the snow covered ground), have been used for the point source propagation measurement. The 4 microphone positions are (d/h, in meter): 7.5/1.2, 15/1.2, 25/1.2 and 25/3.5, as shown in Figure 3-3.
Figure 3-3. The setup for the sound propagation measurement using a point sound
source (Röbäcksdalen, Umeå, February 2012).
Figure 3-4. A crusty snow layer of about 6 cm (left) is under the loose snow of 7 cm.
During the measurement, the wind speed is about 3~5 m/s, roughly in 45 degrees from the travel path downwards to the microphone positions. The temperature is 1 oC.
Currently, there are two methods to determine the effective flow resistivity of a ground of finite impedance: (1) the level difference method as described in [4]; (2) the data fitting method as shown in [2], or described in sub-section 3.2 of this report. As the level
difference method described in [4] is based on the one-parameter impedance model which has been shown not proper for a snow covered ground [2], the effective flow resistivity of the measurement site will be determined by the data fitting method.
(Note: SP has tried several times using the Nordtest method [4], which does not work at low frequencies, to determine the effective flow resistivity of a snow covered field. However, none of the results has been shown to be valid, although these data all suggest an effective flow resistivity about 10~16 kPa s m-2.)
102 103 104 40 50 60 70 hr1: 7.5m/1.2m hs: 1 m sigma: 20000 Pa*s/m2 point sound source
model-hr1 data-hr1 no excess background 102 103 104 30 40 50 60 70 hr2: 15m/1.2m hs: 1 m sigma: 20000 Pa*s/m2 model-hr2 data-hr2 no excess background 102 103 104 30 40 50 60 hr3: 25m/1.2m hs: 1 m sigma: 20000 Pa*s/m2 model-hr3 data-hr3 no excess background 102 103 104 30 40 50 60 hr4: 25m/3.5m hs: 1 m sigma: 20000 Pa*s/m2 model-hr4 data-hr4 no excess background
Figure 3-5a. Sound propagation above snow at the measurement site in Umeå, using the
point sound source (Figure 3-2). The nominal sound power is given within the octave bands 125 ~8k Hz, which corresponds to the one-third octave bands of 100 ~ 10k Hz. Therefore, below 100 Hz, the sound power drops quickly.
The total snow depth at the microphone positions is 53 cm. Below the loose snow of about 7 cm thick there is a layer of lightly-crusty snow of about 6 cm thick, as shown in Figure 3-4.
Based on the data fitting, the effective flow resistivity was determined to be 20 kPa s m-2.
As shown in Figure 3-5, below 100 Hz the model predicted SPLs at the receivers differ much from the respective measured SPLs, due the two reasons: (1) the nominal sound power of the point source is given in octave bands of 125 – 8k Hz, which corresponds to the one-third octave bands of 100 – 10k Hz; therefore, the sound power drops quickly beyond 100 – 10k Hz; (2) below 100 Hz background noise (measured at a different time) becomes important. 102 103 104 40 50 60 70 hr1: 7.5m/1.2m hs: 0.5 m sigma: 20000 Pa*s/m2 point sound source
model-hr1 data-hr1 no excess background 102 103 104 30 40 50 60 70 hr2: 15m/1.2m hs: 0.5 m sigma: 20000 Pa*s/m2 model-hr2 data-hr2 no excess background 102 103 104 30 40 50 60 hr3: 25m/1.2m hs: 0.5 m sigma: 20000 Pa*s/m2 model-hr3 data-hr3 no excess background 102 103 104 30 40 50 60 hr4: 25m/3.5m hs: 0.5 m sigma: 20000 Pa*s/m2 model-hr4 data-hr4 no excess background
Figure 3-5b. The caption is the same as that for Figure 3-5a. (Here hs = 0.5 m.)
Moreover, at the receiver positions of 25 m distance, some error around 200 Hz can be found. This is in fact a typical problem of the ground impedance model for a snow field with its porosity varying with depth. In such a case the position of the first minimum in
frequency spectrum together with the level can be handled only empirically, because the necessary information concerning the detailed properties of the snow is not available [2]. We decide not to care about this error, because (1) it is not important when considering the total noise sound level and (2) when handling sound propagation over a large snow covered field it is not practical to care this kind of local details which can vary a lot point by point. 102 103 104 40 50 60 70 hr1: 7.5m/1.2m hs: 0.1 m sigma: 20000 Pa*s/m2 point sound source
model-hr1 data-hr1 no excess background 102 103 104 30 40 50 60 70 hr2: 15m/1.2m hs: 0.1 m sigma: 20000 Pa*s/m2 model-hr2 data-hr2 no excess background 102 103 104 30 40 50 60 hr3: 25m/1.2m hs: 0.1 m sigma: 20000 Pa*s/m2 model-hr3 data-hr3 no excess background 102 103 104 30 40 50 60 hr4: 25m/3.5m hs: 0.1 m sigma: 20000 Pa*s/m2 model-hr4 data-hr4 no excess background
Figure 3-5c. The caption is the same as that for Figure 3-5a. (Here hs = 0.1 m.)
4
Analysis of the measurements
In section 3 the effective flow resistivity of the snow field at the measurement site in Umeå has been determined to be 20 kPa s m-2. This flow resistivity value has been used in determining the excess attenuation of sound propagation over the snow field.
In sub-section 4.1, the measurement and the data analysis of snowmobile engine noise will be described. And, based on the analysis, the rpm dependence of the engine noise will be determined. Thereafter, in sub-section 4.2, the measurement of snowmobile pass-by noise at several constant speeds will be described and the total snowmobile noise will be analysed and compared with that of the engine noise. The noise emission data of pass-by at full throttle will be analysed in sub-section 4.3, which provide useful information for understanding the North-America noise indicator, LpASmax. Finally, in sub-section 4.4, non-engine noise will be studied.
4.1
The engine noise of the two snowmobiles
By measuring snowmobile noise under a stationary condition the engine noise data was collected, which serves as the starting point in modelling snowmobile noise.
4.1.1
Measurement arrangement
As described in sub-sub-section 2.3.2, four microphone positions were used (d/h): 7.5/1.2, 7.5/2 and 7.5/3.5 m (7.5 m is the distance between the microphone position and the snowmobile axis line) for the three microphones along the normal line passing through the centre of the snowmobile; the use of three microphone heights is of an advantage to determine the relative source strength ratio if there are two or three point sources contribute. The fourth microphone has a position similar to the first one (7.5/1.2 m) but one snowmobile length (3.3 m) shifted to the rear side from the normal line. This fourth position is for checking out if there is an important effect of horizontal directivity.
Figure 4-1. The location of the exhaust outlet: left picture is for the Arctic Cat and right
picture for the Yamaha.
The engine noise of the two snowmobiles were measured with the right side facing to the microphones. The reason for this choice is (1) the total number of measurements has to be limited because of the small budget for the project; and (2) for the snowmobile Arctic Cat the exhaust outlet is located at the front-right side while for the Yamaha there are two exhausts with the outlets symmetrically located at the rear, see Figure 4-1.
4.1.2
The noise spectrum of snowmobile engine noise
The operating conditions are listed in Table 4.1. Part of the engine noise data are presented in Figures 4-2 and 4-3.The snowmobile Arctic Cat has an engine of two-stroke and two cylinders. The idling rpm is 1800, which takes 1/30 s for each working cycle (of one cylinder) and it then corresponds to 30 Hz. And, 60 Hz can be estimated when considering the two cylinders.
If we assume that, for the engine of the Arctic Cat, the fundamental is 60 Hz, then, according to the data, the sixth and sixteenth harmonics are also important in the noise emission, under stationary.
Table 4.1. The operating conditions for measuring the engine noise of the snowmobiles.
Operating conditions Arctic Cat Yamaha Note constant rpm 1800 1400 Idling rpm constant rpm 3000 2000 Moderate rpm
constant rpm 4000 2500 High rpm, which is under but close to the engagement rpm
Acceleration Sweep from rpm 1800 to rpm 4000
Sweep from rpm 1400 to rpm 2500
Acceleration was gently made to avoid the engagement 102 103 104 35 40 45 50 55 60 65 70 75 80
Arctic Cat stationary
rpm 1800 rpm 3000 rpm 4000 rpm 1800-up
Figure 4-2. The Leq values of the engine noise of the Arctic Cat, recorded at 7.5/1.2 m
position (on the normal line), where rpm 1800-up means that the rpm value is gently raised up from rpm 1800 to rpm 4000 (repeated six times).
The snowmobile Yamaha has an engine of four-stroke and three cylinders. It has an idling rpm 1400, which corresponds to 12 Hz if one cylinder is considered and about 35 Hz if the three cylinders are considered.
If we take 35 Hz as the fundamental for the engine noise, we then see that the second harmonics dominates at idling rpm while the fourth and twelfth harmonics become also important at rpm 2500 where the fundamental becomes less predominant.
Approximately, in the total A-weighted level, this Yamaha engine of four-stroke and three cylinders (88.4 kW) is about 3 dB quieter than the Arctic Cat engine of two-stroke and two cylinders (113 kW).
102 103 104 20 30 40 50 60 70 80 Yamaha stationary rpm 1400 rpm 2000 rpm 2500 rpm 1400-up
Figure 4-3. The Leq values of the engine noise of the Yamaha, recorded at 7.5/1.2 m
position (on the normal line), where rpm 1400-up means the rpm value is gently raised up from rpm 1400 to rpm 2500 (repeated six times).
The acceleration can only be made gently because one should be careful to avoid the engagement otherwise the snowmobile will move. Thus, by making a gentle acceleration, it in fact becomes a sweep operating condition (repeated six times when recording). It is found that, for the engine noise under this sweep operating condition, each frequency component of the resulted engine noise is mainly determined by the one of the highest level among the spectra of respective rpm values.
A common feature of the engine noise of the two snowmobiles is that, by increasing rpm value, the spectrum levels at the fundamental and low number of harmonics will be reduced while the spectrum levels at high frequencies will be raised; and, not only the total sound power will raise but also the sound power at high frequencies will become more and more important.
4.1.3
The rpm dependence of snowmobile engine noise
For engine noise it is extremely difficult, if not impossible, to properly handle the sound power distribution among the fundamental and harmonics. While, it is possible to describe the total sound power level of engine noise, at least in an empirical way.
Based on the collected engine noise data, a rpm-dependence of 25 was proposed for the two engine types. The examples of using this rpm-dependence are given in Figures 4-4 and 4-5. As can be seen, this rpm-dependence works well except around those harmonics peak levels. 102 103 104 30 40 50 60 70 80 Leq(rpm 1800) Leq(rpm 3000)+25*lg(1800/3000) 102 103 104 40 50 60 70 80 Leq(rpm 4000) Leq(rpm 3000)+25*lg(4000/3000)
Figure 4-4. Examples of rpm-dependence of the engine noise at stationary, for the Arctic
Cat. A rpm-dependence of 25 is applied. As discussed in the text above, the spectrum levels at the fundamental and low number of harmonics are difficult to model.
102 103 104 20 40 60 80 Leq(rpm 1400) Leq(rpm 2000)+25*lg(1400/2000) 102 103 104 20 40 60 80 Leq(rpm 2500) Leq(rpm 2000)+25*lg(2500/2000)
4.1.4
The source height of snowmobile engine noise
For the Arctic Cat, the exhaust pipe outlet is located at front-right side, 200 mm high. And, for the Yamaha, there are two exhausts located each side at the rear, 500 mm high. Moreover, for the two snowmobiles the height of engine centre is 400 mm.
As described in sub-sub-section 4.1.1, four microphone positions have been used when measuring the snowmobile noise at stationary. Since the results shown in the following figures are given in channel nr, it is necessary to describe the relationship between the microphone positions and the channel numbers; this is given in Table 4.2.
Table 4.2. The microphone positions and the corresponding channel numbers.
Channel 1 Channel 2 Channel 3 Channel 4 Note 8.2 / 1.2 7.5 /1.2 7.5 / 2 7.5 / 3.5 d / h (m)
For checking out if there are more than one important point source, and for determining the corresponding source heights, different possible source heights of 0.1:0.1:0.6 (m) have been tested. Each time only one point source of a source height was considered. The sound propagation attenuation from the point source to one of the four receiver positions was calculated using the Nord2000 propagation model, the latest version compro18e whilst has been revised for being able to handle sound propagation above snow (see section 3). Thus, the sound power level can be determined according to
LW Leq
r
Aexcess 24 lg
10
, (4-1)where Aexcess is the excess attenuation, r the distance between the point source and the receiver, Leq the equivalent SPL recorded at the receiver position.
For the Arctic Cat the results are shown in Figures 4-6 and 4-7, for a source height 0.2 m or 0.4 m respectively. The results for other possible source heights (0.1 m, 0.3 m, 0.5 m) are not presented because they do not provide constructive/extra information.
Based on the results shown in Figure 4-6, we can see that when using only one single point source of 0.2 m high
it works well below 200 Hz;
compared with channel 2 result, channel 1 result suggests a light effect of horizontal directivity;
channel 3 result is well close to channel 2 result, up to 2 kHz.
If also consider the results shown in Figure 4-7, it can be understood that (, based on a general understanding of the excess attenuation of sound propagation,) with a higher source position and a higher receiver position, the excess attenuation will contain more ripples at high frequencies. As a few dB error in spectrum is not strange in handling sound propagation above snow, such limited errors above 2 kHz are then considered tolerable.
As the result of a single point source of 0.4 m high does not improve the case at high frequencies, it is then concluded that a point source of a height between 0.2 m and 0.4 m can equivalently describe the engine noise of the snowmobile Arctic Cat.
The horizontal directivity will be neglected because (1) it does not show a strong effect; and (
2)
to work out a directivity function requires an extensive investigation which is out of the purpose of this project. Moreover, the sound power of the engine noise will be the mean of the resulted sound powers determined based on respective recordings at the four receiver positions. 102 103 104 60 65 70 75 80 85 90 95 100 105 hs = 0.2 (m) Arctic Cat Lw (linear) cha 1 cha 2 cha 3 cha 4 meanFigure 4-6. The sound power levels of the engine noise of the snowmobile Arctic Cat at
the idling rpm, determined using respective recordings at the four receiver positions. Assuming a single point source of 0.2 m high.
102 103 104 65 70 75 80 85 90 95 100 105 hs = 0.4 (m) Arctic Cat Lw (linear) cha 1 cha 2 cha 3 cha 4 mean
Figure 4-7. The caption is the same as that for Figure 4-6, while a different source height
102 103 104 50 60 70 80 90 100 110 hs = 0.5 (m) Yamaha Lw (linear) cha 1 cha 2 cha 3 cha 4 mean
Figure 4-8. The sound power levels of the engine noise of the snowmobile Yamaha at the
idling rpm, determined using respective recordings at the four receiver positions. Assuming a single point source of 0.5 m high.
102 103 104 50 60 70 80 90 100 110 hs = 0.4 (m) Yamaha Lw (linear) cha 1 cha 2 cha 3 cha 4 mean
Figure 4-9. The caption is the same as that for Figure 4-8, while a different source height
of 0.4 m.
For the engine noise of the snowmobile Yamaha, discussions are similar to those for the snowmobile Arctic Cat, while the source height is between 0.4 m and 0.5 m. Relevant results are shown in Figures 4-8 and 4-9. (It seems there is no important horizontal directivity.)
4.2
Snowmobile pass-by noise: under a constant
speed
The purpose of the pass-by measurements is to collect the noise emission data of total snowmobile noise. Thus, as expected, by comparing the spectrum characteristic of the total pass-by snowmobile noise and that of the engine noise, it will be possible to justify if non-engine noise should be considered or not.
4.2.1
Measurement arrangement
Four microphone positions were positioned along a line normal to the travel path, the same as those used in the sound propagation of a point sound source (sub-section 3.3). These four positions are (d/h, in meter): 7.5/1.2, 15/1.2, 25/1.2 and 25/3.5 (see Figure 3-3). By this setup, it is sure how well the sound propagation has been handled.
The reason to choose these four receiver positions has been described in sub-sub-section 2.3.3. The measurement conditions were described in sub-section 3.3.
During a pass-by, a recording would start from 40 m before and would end 40 m past the normal line. Relative to the nearest measurement distance 7.5 m where recordings will be used to determine the sound power level of the noise emission, this travel section is about 5 times the measurement distance at each side then guarantees an accurate recording of the pass-by SEL value. One pass-by example was shown in Figures 4-10.
Figure 4-10. The video clips of a recorded pass-by of the snowmobile Yamaha under a
For each snowmobile, at each direction of drive, and at each of three constant speeds of 25, 50 and 75 km/h, the pass-by noise has been recorded only once. (Note: The number of pass-bys is controlled due to the small budget of this project, while measurement quality is guaranteed by a proper design of the systematic measurements. Therefore for each pass-by measurement, 1 or 2 dB variations in recording are considered tolerable.)
4.2.2
The noise spectrum of snowmobile pass-by noise
The recorded pass-by sound exposure level, SEL, can easily be transformed into the corresponding equivalent sound pressure level, Leq,Leq SEL10lg
T SEL10lg
80/V
, (4-2)where the length of the travel-path is 80 m and the pass-by speed V is in m/s.
In Figures 4-11 (L2R) and 4-12 (R2L) the 25 km/h pass-by Leq was compared with those of the engine noise, for the snowmobile Arctic Cat. (The recordings at the microphone position 7.5/1.2 m were used.) “L2R” means that the snowmobile passes from the left to the right, as exampled in Figure 4-10; and, “R2L” means from the right to the left.
102 103 104 35 40 45 50 55 60 65 70 75 80
Arctic Cat stationary compared with cruise (L2R)
rpm 1800 rpm 3000 rpm 4000 25 km/h
Figure 4-11. The comparison of the pass-by Leq under 25 km/h and those of the engine noise at different rpm-values, for the snowmobile Arctic Cat. “L2R” means that the snowmobile passes from the left to the right, as shown in Figure 4-10.
102 103 104 35 40 45 50 55 60 65 70 75 80
Arctic Cat stationary compared with cruise (R2L)
rpm 1800 rpm 3000 rpm 4000 25 km/h
Figure 4-12. The caption is the same as that for Figure 4-11, while the pass-by data in
this plot is for the R2L pass-by.
By these results, one may conclude that, for the snowmobile Arctic Cat, its left side (R2L, 67.6 dBA) is a little bit more noisy than its right side (L2R, 65.7 dBA). However, it is not proper to draw this conclusion considering only one pass-by has been recorded for each direction and 1 or 2 dB variation in measurement is common. In fact, by checking the spectrum, the first maximum for the pass-by R2L is located at a slightly higher frequency than the one for the L2R pass-by, which may suggest a higher drive speed.
The more important issue is to compare the spectrum characteristic, at high frequencies where the non-engine noise may contribute. However at high frequencies, obviously, the 25 km/h pass-by data and the engine noise data are, roughly, of the same characteristic.
Let us further check the pass-by data of higher speeds. (For reducing redundancy, the noise emission data of the pass-by at full throttle has been put together, while it will be discussed in next sub-section.)
All the pass-by data (the recordings at the microphone position 7.5/1.2 m) are presented in Figures 4-13 (L2R) and 4-14 (R2L). One can see they are of the same characteristic, at high frequencies in spectrum. This then suggests that the non-engine noise sources are not important both in the total level and in frequency components. Thus, we conclude that it is enough to consider the engine noise only, for the snowmobile Arctic Cat on a soft trail.
For the noise of this two stroke engine, it seems that the spectrum peak levels will slightly “shift” to upper frequencies when the rpm value increases. However, as will be shown next, this phenomenon has not been found for the four stroke engine. (And, opposite situations are shown in Figures 4-2 and 4-3.)
The extra information released by these data is that for speed 25 km/h the corresponding rpm is close to the engagement rpm.
102 103 104 40 45 50 55 60 65 70 75 80
Arctic Cat pass-by (L2R)
25 km/h 50 km/h 75 km/h acceleration
Figure 4-13. The pass-by Leq under a constant speed of 25, 50 or 75 km/h, or at full throttle, for the snowmobile Arctic Cat.The pass-by direction is L2R.
102 103 104 40 45 50 55 60 65 70 75 80 85
Arctic Cat pass-by (R2L)
25 km/h 50 km/h 75 km/h acceleration
Figure 4-14. The pass-by Leq under a constant speed of 25, 50 or 75 km/h, or at full
102 103 104 20 30 40 50 60 70 80
Yamaha stationary compared with cruise (L2R)
rpm 1400 rpm 2000 rpm 2500 25 km/h
Figure 4-15. The comparison of the pass-by Leq under 25 km/h and those of the engine noise at different rpm-values, for the snowmobile Yamaha.
102 103 104 20 30 40 50 60 70 80
Yamaha stationary compared with cruise (R2L)
rpm 1400 rpm 2000 rpm 2500 25 km/h
Figure 4-16. The caption is the same as that for Figure 4-15, while the pass-by data in
this plot is for the R2L pass-by.
For the Yamaha four-stroke engine (88.4 kW) its spectrum characteristic differs from that of the Arctic Cat two-stroke engine (113 kW). As shown in Figure 4-15, the sound level
of the Yamaha engine noise is, compared with that of the Arctic Cat engine noise, about 5 dB less below 100 Hz, comparable at medium frequencies and 5~10 dB less above 1 kHz.
For the Yamaha, the spectrum characteristic of 25 km/h pass-by noise is slightly different from that of the engine noise, above 3 kHz (comparing the curve slopes). This may suggest some contribution of non-engine noise at the high frequencies. However, because of the low level, the non-engine noise is still not important in this situation.
102 103 104 35 40 45 50 55 60 65 70 75 80 85 Yamaha pass-by (L2R) 25 km/h 50 km/h 75 km/h acceleration
Figure 4-17. The pass-by Leq under a constant speed of 25, 50 or 75 km/h, or at full
throttle, for the snowmobile Yamaha. The pass-by direction is L2R.
102 103 104 35 40 45 50 55 60 65 70 75 80 Yamaha pass-by (R2L) 25 km/h 50 km/h 75 km/h acceleration
Figure 4-18. The pass-by Leq under a constant speed of 25, 50 or 75 km/h, or at full
The Yamaha pass-by data were presented in Figures 4-17 (L2R) and 4-18 (R2L). As can be see, the spectrum level of the pass-by noise develops with the rpm value in the same rate, above 100 Hz. Therefore, if the non-engine noise sources are relatively not important at 25 km/h, then they are still relatively not important at a higher speed.
Thus, it can be concluded that, for the tested two snowmobiles, only engine noise is important if they drive on a soft trail.
4.2.3
The speed dependence of snowmobile pass-by noise
In sub-sub-section 4.1.3 a rpm-dependence of 25 has been proposed for the engine noise. As the dominance of engine noise during the pass-bys has been proved, therefore, a speed dependence of 25 was naturally also proposed for snowmobiles pass-by noise. The example calculations were given in Figures 4-19 ~ 4-22. As can be seen, this speed dependence works reasonably well. (No intension to handle those peak levels at harmonics.) 102 103 104 40 50 60 70 80 Leq(25 km/h) Leq(50 km/h)+25*lg(25/50) 102 103 104 50 55 60 65 70 75 Arctic cat (L2R) Leq(75 km/h) Leq(50 km/h)+25*lg(75/50)Figure 4-19. Measured and calculated Leq for the pass-bys, for the Arctic Cat. A speed-dependence of 25 is applied. No intension to handle those peak levels at harmonics.
102 103 104 40 50 60 70 Leq(25 km/h) Leq(50 km/h)+25*lg(25/50) 102 103 104 50 55 60 65 70 75 Arctic cat (R2L) Leq(75 km/h) Leq(50 km/h)+25*lg(75/50)
Figure 4-20. The caption is the same as that for Figure 4-19, while here are for the
pass-bys R2L. 102 103 104 30 40 50 60 70 Leq(25 km/h) Leq(50 km/h)+25*lg(25/50) 102 103 104 40 50 60 70 80 Yamaha (L2R) Leq(75 km/h) Leq(50 km/h)+25*lg(75/50)
Figure 4-21. Measured and calculated Leqfor the pass-bys, for the Yamaha. A
102 103 104 30 40 50 60 70 Leq(25 km/h) Leq(50 km/h)+25*lg(25/50) 102 103 104 40 50 60 70 80 Yamaha (R2L) Leq(75 km/h) Leq(50 km/h)+25*lg(75/50)
Figure 4-22. The caption is the same as that for Figure 4-21, while here are for the
pass-bys R2L.
4.3
Snowmobile pass-by noise: at full throttle
Aiming at understanding the North-America noise indicator, LpASmax, the noise emission data of snowmobile pass-by at full throttle was also collected and analyzed, as described in this sub-section.
4.3.1
The North-America method SAE J192
The North-America method J192 [5] is summarized in the following: The measurement area starts from 75 feet (22.86 m) before the centre of the travel lane (where the normal passes through the microphone stand) and ends 75 feet past the centre;
The snowmobile approaches the starting point with an initial speed of 15 mph (24 km/h) and then maintain full throttle as it travels through the measurement area;
The microphone position is 50 feet (15.2 m) from the snowmobile path and 4 feet (1.2 m) above the ground;
The noise indicator is the maximum instantaneous A-weighted SPL using time-weighting S, LpASmax, during the pass by;
It is required a minimum of 2 inch (50.8 mm) packed snow under the travel path and a maximum of 3 inch (76.2 mm) loose snow on top of the packed snow.
The result for each side is the average of three measurements differing less than 2 dB rounded to the nearest integer. The final result is the highest between the ones of the two sides.
4.3.2
Measurement arrangement
The aim of this measurement is to find the relationship between the North-America noise indicator, LpASmax, and a defined noise sound power level. As the measurement made in this project is for modelling snowmobile noise, not for certificating a snowmobile, it is not necessary to fulfil all the requirements set by SAE J192. For example, for each side of a snowmobile only one pass by has been recorded in order to reduce the measurement load – this is necessary because of the small budget.
The four microphone positions are the same as those described in sub-sub-section 4.2.1. For the North-America noise indicator, LpASmax, the value of it is determined based on the recordings of the second microphone which is located at 15/1.2 m.
The measurements of pass by at full throttle start at a speed of 25 km/h, 22.8 m before the centre and end 22.8 m past the centre.
4.3.3
The noise spectrum
The results were presented in Figures 4-13 and 4-14 for the Arctic Cat and Figures 4-17 and 4-18 for the Yamaha.
As can be seen, for a pass by at full throttle, the spectrum characteristic of pass-by equivalent SPL, Leq, is similar to that of the engine noise at a high rpm value, at medium and high frequencies. This is probably due to that, by a powerful acceleration, the engine rpm will quickly reach a high value then the engine noise of high-rpm characteristic dominates during the pass-by.
102 103 104 40 50 60 70 80 Leq(acc, L2R) Leq(75 km/h)+9 102 103 104 40 50 60 70 80 Arctic Cat 15/1.2 m Leq(acc, R2L) Leq(75 km/h)+9
Figure 4-23. The Leq of a pass-by at full throttle (with the average speed about 86 km/h) compared with that under a cruise pass-by at 75 km/h, for the Arctic Cat.
102 103 104 40 50 60 70 80 Leq(acc, L2R) Leq(75 km/h)+6 102 103 104 40 50 60 70 80 Yamaha 15/1.2 m Leq(acc, R2L) Leq(75 km/h)+6
Figure 4-24. The Leq of a pass-by at full throttle (with the average speed about 86 km/h) compared with that under a cruise pass-by at 75 km/h, for the Yamaha.
For the Arctic Cat, the measured pass-by data were shown in Figures 4-13 and 4-14. Under the acceleration the pass-by time is slightly less than 4 seconds; accordingly, estimated average pass-by speed is 86 km/h. However, the noise level is about 9 dB raised up compared with that of the cruise pass-by at 75 km/h. The comparison of the two noise emission levels was shown in Figure 4-23. (Note: considering the use of the A-weighting will reduce the importance of low-frequency components, the overestimation at low frequencies by this simple treatment is thought tolerable.)
For the Yamaha, the measured pass-by data were shown in Figures 4-17 and 4-18. The pass-by time for the pass-by at full throttle is slightly less than 4 seconds; or, the average pas-by speed is about 86 km/h. Moreover, the noise level is about 6 dB raised up
compared with that of the cruise pass-by at 75 km/h. The comparison of the two noise emission levels was shown in Figure 4-24.
Table 4.3. The A-weighted noise emission data at the receiver position 15/1.2 m.
acc LpAS max (dBA)
acc LeqA (dBA)
km h
LeqA 75 / (dBA) Arctic Cat L2R 83.8 80.4 71.3 R2L 85.0 81.7 72.4 Yamaha L2R 78.9 76.9 69.8 R2L 79.2 76.6 70.5Note: The noise limit in North-America for snowmobiles is 78 dBA (LpAS max
acc ). In Table 4.3 the values of the North-America noise indicator, LpAS max
acc , and thecorresponding LeqA
acc were compared with LeqA
75km/h
(at the receiver position 15/1.2 m). For the snowmobile Arctic Cat, LpAS max
acc is about 12.5 dBA higher than
km h
LeqA 75 / . And, for the Yamaha, LpAS max
acc is about 9 dBA higher than
km h
LeqA 75 / .The North-America noise limit is based on: (1) the result for each side is the average of three measurements differing less than 2 dB rounded to the nearest integer; (2) the final result is the highest between the ones of the two sides. This requirement is important for a certification test.
In this project, the measurement load has to be reduced because of the small budget. It seems that, compared with having three qualified recordings for each situation, the data quality could be lowered if having only one recording for each situation. However, this is in fact not always true. As we have systematically measured the noise emission data of the two snowmobiles, when put all these data together, the measurement uncertainty has been reduced and the data quality has been raised up to the same good as that if having three qualified recordings for each situation (e.g. one will have the same confidence on the results). The only special factor is that, when working on the measurement data (e.g. when determining the rpm-dependence), possible 1 or 2 dB variation in the pass-by noise level should be taken into account.
4.4
Referring to some earlier measurements
For snowmobiles, non-engine noise includes transmission noise, suspension noise and track noise. For modern snowmobiles, many advanced techniques have been applied for reducing noise from these driveline components. It is then interesting to find out how important the non-engine noise will be for modern snowmobiles. With this purpose, in 2009 and in Umeå, SP Acoustics arranged a special pass-by measurement by using one snowmobile to tow one other snowmobile [8].
The towed snowmobile, Lynx Adventure Grand Tourer 600 SDI (hereafter, the Lynx), was connected to the towing snowmobile, Arctic Cat TZI Turbo LXR (hereafter, the Arctic Cat 2009) by a 40 m long rope, as shown in Figure 4-25. During the pass-by, the gear of the Lynx was put neutral and the engine was shut off. The speed is 24 km/h.
Figure 4-25. The towed snowmobile Lynx was connected to the towing snowmobile by a
The spectrum time history of the pass-by noise was recorded. Thus, both the total sound pressure level and the frequency components together with the time variations of these quantities can be studied. Three periods during the pass-by are most interesting: (1) when the towing snowmobile is most influencing; (2) when the towed snowmobile is most influencing; and (3) when the minimum SPL happens between the two most influencing time periods. The data sampling of these three periods was shown in Figure 4-26. (Note: with a short time interval the data of point-source feature has been obtained.)
(a)
(b)
(c)
Figure 4-26. Sampling the three periods during the pass-by: (a) when the towing
snowmobile is most influencing; (c) when the towed snowmobile is most influencing; and (b) when the minimum SPL happens between the two most influencing time periods.
4.4.1
Non-engine noise
For the time history of the towing-by noise emission shown in Figure 4-26, the A-weighted SPL levels at the receiver position 7.5/1.2 m is 73.4 dBA for (a) and 69.7 dBA for (c) and 57.8 dBA for (b). Because in situation (b) the A-weighted SPL level is more than 10 dBA less than those in situations (a) and (c), this suggests that it is likely that the towing snowmobile has negligible influence on the recording of the towed snowmobile noise. And, as shown in Figure 4-27, this is true even at low frequencies.
102 103 104 30 40 50 60 70 80 ArcticCat towing-by 24 km/h hr = 1.2 m d = 7.5 m / 15 m cha 1 (73.4 dBA) cha 3 (65.9 dBA) background cha 1 background cha 3 102 103 104 30 40 50 60 70 80 Lynx towed-by 24 km/h cha 1 (69.7 dBA) cha 3 (61.7 dBA) background cha 1 background cha 3 102 103 104 30 40 50 60 70 80
the minimum between the towing and towed
cha 1 (57.8 dBA) cha 3 (55.5 dBA) background cha 1 background cha 3
Figure 4-27. Comparison of SPLs at the two receiver positions: cha 1 is for 7.5/1.2 m
position and cha 3 for 15/1.2 m position. Data samplings for the towing, towed and the minimum are presented in Figure 4-26.
As shown in Figure 4-27, it clearly shows a point-source behaviour for the towing snowmobile and for the towed snowmobile, i.e. 6 dB reduction per double distance plus some due to the excess attenuation. However, for the minimum SPL between the towing and towed snowmobiles, it is determined by the two point sources located at the two sides of the normal at farther distances. Accordingly, for each of the two snowmobiles, the distances from the snowmobile to the two microphone positions are not much different, leading to a small difference between the sound immission levels during the time period.
To estimate the source height, a calculation was made. As the sound power level of the pass-by noise is not known but can be estimated using the recording at one receiver position (7.5/1.2 m position was used), the source height (in calculation of the excess attenuation) was adjusted to reach a better prediction of the SPL at other microphone position (15/1.2 m was chosen) compared with the recording. The calculation examples were shown in Figure 4-28. Accordingly, the source height is 0.4 m for the engine noise and 0.2 m for the non-engine noise.
102 103 104 30 40 50 60 70
the towed Lynx
calculation cha 3 data cha 3 background cha 3 102 103 104 30 40 50 60 70
The towing Arctic Cat
calculation cha 3 data cha 3 background cha 3
Figure 4-28. The comparison between the measured and calculated SPLs at the receiver
position 15/1.2 m (cha 3), by using the measured SPL at the receiver position 7.5/1.2 m. The background noise was recorded at a different time.
The non-engine noise can roughly be learnt by this towing-by measurement, compared with the cruise-by at the same speed, as shown in Figure 4-29. It should be mentioned that one should be careful in comparing two measured equivalent SPLs – (to guarantee sampling data of point-source feature) the sampling time interval must be comparable! For example, the cruise-by Leq should be measured in the similar small time interval centred at the peak level position. This comparison shows that the non-engine noise is not important below 300 Hz. Above 300 Hz, the non-engine noise seems comparable to the engine noise (the total pass-by SPL, “cruise”, is about 3 dB higher than that of the non-engine noise, “towed”), for the tested snowmobile on the test trail which is of compacted snow.
However, the non-engine noise seems not always important. In Figure 4-30 a part of test results of snowmobile noise made in 2009 and made in 2012 are compared. This
comparison indicates that the non-engine noise of the Lynx (tested in 2009) is higher than the total cruise-by noise of the Yamaha (tested in 2012). In other words, the non-engine noise of the Yamaha (on a soft trail) is at least 2 or 3 dB lower than that of the Lynx (see those frequency components above 1 kHz). If assuming the non-engine of the Arctic Cat is comparable to the non-engine noise of the Yamaha, one will conclude that, for the Arctic Cat, the non-engine noise is not important in the total pass-by noise.
102 103 104 30 40 50 60 70 80 channel 1 towed cruise background 102 103 104 30 40 50 60 70 80 channel 3 towed cruise background (a) 102 103 104 0 10 20 30 Leq(cruise) - Leq(towed) cha 1 102 103 104 0 10 20 30 cha 3 (b)
Figure 4-29. The total noise compared with the non-engine noise of passing-by
snowmobile Lynx at a constant speed 24 km/h.(a) The total equivalent SPLs (measured in the small time interval centred at respective peak level). (b) The level difference.
102 103 104 40 45 50 55 60 65 70 75 80 passing-by at 24~25 km/h
The towed Lynx 2009 (69.7 dBA) Arctic Cat 2012 (72.8 dBA) Yamaha 2012 (69.1 dBA)
Figure 4-30. Comparison of the non-engine noise of the Lynx (tested in 2009 on a trail of
compacted snow) and the total noise of the Arctic Cat and the Yamaha (tested in 2012 on a soft trail).
4.4.2
Directivity of engine noise
An earlier study [9] on stationary snowmobile noise (i.e. the engine noise) indicated that the noise emission differs in different directions, up to 3~4 dB in difference. Roughly speaking, noise emission level at the exhaust side is higher.
5
Modelling snowmobile noise
Aiming at properly modelling snowmobile noise, a literature study was first made to broaden the knowledge. Thereafter, a simple source model of snowmobile noise was worked out, mainly based on the data analysis presented in section 4.
5.1
Literature study
It is in fact difficult to find relevant journal articles on snowmobile noise modelling – only a few articles have been found useful. The main reason could be that this noise type is limited to a few regions in the world: snowmobiles are mostly used in North-America, and secondly used in the north of the Nordic countries and in Japan. It could also be due to a commercial reason: snowmobile manufacturers may not like to tell the public how they manage to reduce the components noise, because any success in noise reduction can have a big impact on snowmobile market. Moreover, practical issues in handling
snowmobile noise are in these articles discussed more than modelling details.
However, some articles found from website are interesting. As described in [10], a chassis dynamometer was developed and built by AVL North America in the new Arctic Cat NVH facility, which is capable of testing a snowmobile under full road-load and inertia simulation conditions. The motoring capability of the dynamometer makes it possible to drive the sled track and transmission at full speed with the engine not fired. Thus, noise emitted from the driveline components can be tested. Obviously, the noise data of these snowmobile components have been collected while currently not available for the public.
All moving parts of a snowmobile can be potential noise sources [11]. For convenience, these noise types can be categorized into two groups: power-train noise and driveline noise. The power-train noise includes engine noise, intake noise and exhaust noise where exhaust is usually the primary noise source. For the driveline noise, it includes CVT2 noise, suspension noise and track noise. Moreover, due to the body vibration, the chassis and chain-case can also radiate noise. In this report the power-train noise is simply called engine noise because the measurement cannot distinguish its different contributors. The noise types other than the power-train noise are in this report called non-engine noise.
In principle an exhaust system can be made very quiet by using a very effective muffler, while this requires more space also raises weight and cost which is usually not desired - in some cases there are legal restrictions on weight and width of a snowmobile. This is the reason the exhaust has long been the primary source of snowmobile noise.
Except exhaust noise, track noise can also be important or even dominant in some cases. There are three origins of the track noise: the contact interaction when the drive cogs interface to the track, the contact interaction when the track (also the skis) hits a trail surface, and the air flow driven by the track lugs [12].
However, these track lugs by their design cannot be effective in driving air flow. And, also considering the possible top speed, the aerodynamic noise of turbulent flow induced by the track lugs is likely not comparable to the engine/exhaust noise.
Track contact noise depends much on a track design. After applying the silent track technology [13], the track contact noise produced when the drive cogs interface to the track has much been reduced. However, the track contact noise produced when the
