Table Of ContentKTHEngineeringSciences
Predicting flow-generated noise from HVAC
components
Oscar Kårekull
LicentiateThesis
Stockholm,Sweden
2015
AcademicthesiswithpermissionbyKTHRoyalInstituteofTechnology,Stockholm,tobesubmitted
forpublicexaminationforthedegreeofLicentiateinTechnicalAcoustics,Wednesdaythe27thof
May,2015at10.15,inroomMunin,Teknikringen8,KTH-RoyalInstituteofTechnology,Stockholm,
Sweden.
TRITA-AVE2015:22
ISSN1651-7660
ISBN978-91-7595-564-3
(cid:13)c OscarKårekull,2015
Postaladdress:
KTH
FarkostochFlyg
SE-10044Stockholm
Visitingaddress: Teknikringen8
Contact: [email protected]
Abstract
Moreenergyefficientfans,i.e. largersizesrunningatlowerspeeds,inHeatingVentilationandAir
Conditioning(HVAC)systemsdecreasethefannoiseandincreasetheimportanceofflowgenerated
noiseinothersystemcomponents,e.g.,dampersandairterminaldevices.
In this thesis, an extended prediction model, using semi-empirical scaling laws, for flow noise
predictioninHVACsystemsatlowMachnumberflowspeedsispresented. Thescalinglawscanbe
seenasacombinationofageneralizednoisespectrumbasedonexperimentaldataandconstriction
flow characteristics, where the latter can be gained from Computational Fluid Dynamics (CFD)
simulations. Theflowgeneratednoisecanbepredictedbysemi-empiricalscalinglawstoavoida
timeconsuming,fullyresolvedsimulationormeasurement.
Here,anapproachissuggestedwherethegeneralnoisespectraarecombinedwithturbulentdata
obtainedfromReynoldsAverageNavierStokes(RANS)simulations. Amodelisproposedusing
amomentumfluxassumptionofthedipolesourcestrengthandafrequencyscalingbasedonthe
constrictionpressureloss.
Toevaluatetheapplicabilityofthesemi-empricalscalinglawondifferentHVACgeometriesboth
literaturedataandnewmeasurementdataareconsidered. Focusisatcomparinggeometriesof
highandlowpressurelossbutalsotodiscussthedifferencesinotherproperties, e.g. radiation
characteristics. A general noise reference spectrum is determined by a best fit calculation of
measurementdataincludingorifice,damperandbendgeometries.Airterminaldevicesattheendof
aductarealsoevaluatedandcomparedtoconstrictionsinsideducts. Theexpectedaccuracyofthe
suggestedmodelanditschallengesasatoolforflownoisepredictionofnon-rotatingcomponents
inHVACsystemsarediscussed.
Keywords: flownoise,noiseprediction,HVAC
ii
Sammanfattning
Pågrundavökadeenergieffektivitetskravharstörrefläktarsomroterarmedlägrehastighetbörjat
användasibyggnadersventilationssystem(HVAC).Delägrehastigheternaharminskatljudnivån
från fläkten och ökat betydelsen av strömningsalstrat ljud från andra systemkomponenter, t.ex.
spjällochluftdon.
I denna avhandling presenteras en förbättrad prediktionsmodell, utifrån semi-empiriska skal-
ningslagar,förströmningsalstratljudiventilationssystem. Skalningslagarnakansessomenkom-
binationavgenerellaljudspektraochstrypningensspecifikaflödesegenskaper,därdetsenarekan
fåsfrånComputationalFluidDynamics(CFD)simuleringar. Semi-empiriskaskalningslagarärett
alternativförattundvikatidskrävandemätningarellerfulltupplöstasimuleringar.
Ett tillvägagångssätt presenteras här där det generella spektrat, bestämt utifrån experimentell
data,kombinerasmeddatafrånReynoldsAverageNavierStokes(RANS)simuleringar. Enpre-
diktionsmodell föreslås där källstyrkan hos dipolkrafterna definieras utifrån rörelsemängd och
frekvensskalningenutifrånstrypningenstryckfall.
FörattutvärderavilkaHVACgeometriersomkaningåidengenerellamodellenanalyserasbåde
resultatfrånlitteraturensamtnyamätningar. Avhandlingsarbetetfokuserarpåattjämförageo-
metrieravhögtochlågttryckfallmenocksåpåattdiskuteraskillnaderiandraegenskapersåsom
strålningskarakteristikt.ex. genomattjämföraluftdonislutetavenkanalmedstrypningarinuti
kanalen. Ettgenerelltljudspektrumföreslåsutifrånenanpassningavmätdataförstrypningar,spjäll
ochböjar. Modellensförväntadenoggrannhetochdessutmaningarsomprediktionsverktygför
icke-roterandekomponenteriventilationssystemdiskuteras.
Nyckelord: flödesalstratljud,bullerprediktion,HVAC
iii
Acknowledgements
Thecreativityinaresearchprojectbutalsoitsdifferentchallengesareagreatexperience,especially
foranengineer. IwouldliketothankmyemployerFläktWoodsforenablingmethisopportunity.
I would also like to express gratitude to my supervisors, Mats Åbom and Gunilla Efraimsson,
forallsupportbutalsoforintroducingmetoyourrealmofscience.
Tomycolleagues,thankyouforsupportandcompanionship.
Tomyfamily,thankyouforloveandpatience.
OscarKårekull
Stockholm,8thMay2015
iv
Thislicentiatethesisconsistsofanintroductionandsummaryoftheresearchinadditiontothe
followingappendedpapers
PaperA
Predictionmodelofflowductconstrictionnoise.
KårekullO.,EfraimssonG.andÅbomM.
JournalofAppliedAcoustics2014;82,45-52
PaperB
RevisitingtheNelson-Morfeyscalinglawforflownoisefromductconstrictions.
KårekullO.,EfraimssonG.andÅbomM.
Submitted
Divisionofworkbetweenauthors
Kårekullconductedthemeasurements,theanalysisandwrotethepaperssupervisedbyEfraimsson
andÅbomwhocontributedwithideas,suggestionsandtextreview.
Theworkhasbeenpresentedatthefollowingconferences
1. Kårekull O. and Efraimsson G. Comparison of RANS parameters for flow noise prediction,
Internoise2013,Innsbruck,Austria
2. KårekullO.EfraimssonG.andÅbomM.Challengesandopportunitiesforflownoiseprediction
inHVACsystems,Fan2015,Lyon,France
Theworkwassupportedby
1. TheSwedishResearchCouncilFormas,project245-2011-1615
2. FläktWoods
v
Contents
I Overview and Summary 2
1 Introduction 3
1.1 Literature review . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Objective and scope . . . . . . . . . . . . . . . . . . . . . . . 6
2 Model for flow noise prediction 8
2.1 Semi-empirical scaling laws . . . . . . . . . . . . . . . . . . 8
2.2 Force modeling by Nelson and Morfey . . . . . . . . . . . . 9
2.3 Force modeling from momentum flux . . . . . . . . . . . . . 9
2.4 Radiation resistance . . . . . . . . . . . . . . . . . . . . . . 10
2.5 Frequency scaling . . . . . . . . . . . . . . . . . . . . . . . . 12
2.6 Comparison of force definitions . . . . . . . . . . . . . . . . 13
3 RANS simulation approaches 15
3.1 Constriction pressure drop . . . . . . . . . . . . . . . . . . . 16
3.2 Downstream turbulence model data . . . . . . . . . . . . . 17
4 Evaluated geometries 19
4.1 Published data of orifice, bends and dampers . . . . . . . . 19
4.2 Measured constrictions . . . . . . . . . . . . . . . . . . . . . 19
5 Results 24
5.1 Collapse of orifices, bends and dampers . . . . . . . . . . . 24
5.2 Momentum flux model . . . . . . . . . . . . . . . . . . . . . 25
5.3 Constrictions at end of ducts . . . . . . . . . . . . . . . . . . 28
5.4 Additional HVAC components . . . . . . . . . . . . . . . . 30
5.5 Noise predictions using RANS simulations . . . . . . . . . 33
6 Scaling law generality 38
7 Conclusions and suggestions for future work 40
8 Summary of appended papers 41
Bibliography 43
II Appended papers 47
Part I
Overview and Summary
2
1 Introduction
Indoor noise levels have major negative effects in many situations e.g.
onlearningcapabilitiesinschools[1,2],recoveryinhospitals[3]andon
sleep disturbance in residential buildings [4]. Research on comfortable
indoor environment, e.g. [5], demonstrates the relations and the impor-
tanceofnoiselevelstotheindoorcomfort. Oneofthemainindoornoise
sources is the HVAC system. The HVAC system is supposed to deliver
ahealthyindoorclimatewhereapartfromtemperature,airqualityand
air velocity, a low noise level is crucial. To minimize the effects on hu-
mans,noiseregulations[6,7]demandlowindoornoiselevelswhichare
often challenging for the HVAC system to achieve.
A reduction of energy consumption in today´s society is in great fo-
cus,whereheating,coolingandairexchangeisoneofthemainelements
of energy consumption in buildings. With lowered component pressure
dropsandincreasedefficiencyofheatingandcoolingbatteriesaconsid-
erable amount of energy can be saved. A challenge for HVAC compo-
nents is to deliver a healthy indoor climate in combination with high
energy efficiency without extending the space required in the building.
On a global market, products need to be developed for a wide range
ofconditions. Forexample,inawarmclimatehighercoolingisrequired
compared to the airflow in a cooler climate. To adapt the products, dif-
ferentmeansoflimitingtheairflowareintroducedresultingintougher
noise generating conditions to control. For more flexible and energy ef-
ficient products an improved understanding and ability to predict the
acoustic properties are desirable.
Traditionally the focus of the HVAC system flow noise has been the
soundgenerationfromfans,whilethefanhistoricallyhasbeenthemain
noise source of the system. Due to increased demand of energy effi-
cientsolutions,largerfansrotatingatlowerspeedsarenowadaysused.
Lower speed decreases the fan noise level and increases the dominance
of flow-generated noise to other system components, e.g., dampers and
airterminaldevices. ThefrequencycontentsoftheHVACflownoisecan
overall be divided into fan related low frequencies and mid and high
frequencies from flow noise of other system components. In a system
wheremorenoisesourcesareofimportanceandtheenergyefficiencyis
in focus the designer has a more complex task, e.g., as discussed by the
author in [8].
When designing a HVAC system, potential flow noise generation
needs to be estimated since measurement data is not available until
the exact product and duty point is specified. It would be beneficial to
avoid system changes at a later design stage, by early identification of
the dominating noise sources, from purely the flow characteristic infor-
mation. In a product development process the lack of early prototypes
3
Description:cus, where heating, cooling and air exchange is one of the main elements . terminal devices, silencers with baffles and a heat exchanger, are eval- .. Aiki. (27) where C is a model constant and Ai is the corresponding surface.