Author's personal copy Experimental data set for validation of heat, air and moisture transport...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Experimental data set for validation of heat, air and moisture transportmodels of building envelopes

Tadiwos Zerihun Desta, Jelle Langmans, Staf Roels*

Division of Building Physics, Katholieke Universiteit Leuven, Kasteelpark Arenberg 40, B-3001 Leuven, Belgium

a r t i c l e i n f o

Article history:Received 2 September 2010Received in revised form15 October 2010Accepted 11 November 2010Available online 18 November 2010

Keywords:Building envelopeHAM, Light weightTemperatureHygrothermalValidation data

a b s t r a c t

This paper reports experimental studies on heat, air and moisture (HAM) transfer through a full scalelight weight building envelope wall under real atmospheric boundary conditions. The main objective ofthe article is to generate informative data so that it can be used for numerical validation of HAM models.The considered wall is a multilayered structure built up from outside to inside of external board, ventedcavity, fibreboard sheathing, mineral wool between wooden studs and interior finishing. The global wallhas a surface area of (1.80 � 2.68) m2; and is subdivided into three vertical parts. The parts differ fromeach other by the applied interior finishing. Between the different layers of each part and on the surfacesof the wall humidity, temperature and heat flux sensors are placed in a 3D matrix. At the outer surface ofthe wall, the applied sheathing is a bituminous wood board. In the board nine removable specimens areincluded. By regularly weighing the fibreboard samples, their moisture content could be quantified.Using data collected over a total time span of about two years, insight about the hygrothermal behaviourof the different envelope parts is obtained and at the same time a well-documented data set is generatedthat can be used for hygrothermal envelope model validation purposes.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Nowadays building designs need to fulfil with criteria of energyefficiency, minimum environmental impact and provide healthyand safe condition for building occupants. Light weight construc-tions are becoming a promising alternative in attaining theimposed criteria even in countries with a masonry tradition.Moreover the ease to incorporate a thick insulation layer betweenthe wooden studs makes light weight buildings relatively simple tobuild, sustainable and renewable. But while high thermal perfor-mances are easily achievable, such constructions are susceptible formoisture-related structural deteriorations [1]. Moreover excessmoisture can lead to the growth of pathogenic moulds and reducethe esthetical appearance of buildings [2]. In order to attainoptimum performance, numerical models have been developed tosimulate the hygrothermal response of building envelopes. Thesehygrothermal building envelope simulation tools (often referred toas Heat, Air and Moisture (HAM) models) evolved from the Glaser-method, a one-dimensional hand calculation method for vapourdiffusion through insulated components [3,4]. Through the yearsmodels that incorporate heat and moisture capacity, liquid water

and air transport, two- and three-dimensional aspects and differentmoisture sources such as wind driven rain, rising damp, initialmoisture, interstitial and surface condensation have been devel-oped [5e7]. Several of these HAM-models are nowadays commer-cially available for practitioners in the field and are increasinglyused for analysing the heat and moisture behaviour of buildingcomponents. Although HAMmodels are excellent engineering toolsto optimise and analyse thermal and moisture performance ofbuildings and building components, care need to be taken on theiruse. For example incorrectly defined boundary conditions andmaterial properties can lead to erroneous results [8]. Furthermorethe models need to be validated under different realistic scenariosbefore the predictions are credible.

An important effort to standardise HAM-modelling procedureswas made by the EU-initiated HAMSTAD-project (Heat, Air andMoisture Standards Development) [9]. Benchmark cases were usedformodel validation. The cases havebeen selected in such away thatvarious materials, transport mechanisms and climatic boundaryconditions were covered [10]. The validation, though, relied only onnumerical inter-modal comparisons or problems with simpleanalytic solutions since accurate and well-documented experi-mental data are scarce. In the framework of IEA ECBCS Annex 41a transient heat and moisture experiment on a porous buildingmaterial was performed for benchmarking numerical models withrespect to hygroscopic loading [11]. This kind of well-controlled

* Corresponding author. Tel.: þ32 16 321349; fax: þ32 16 321980.E-mail address: [email protected] (S. Roels).

Contents lists available at ScienceDirect

Building and Environment

journal homepage: www.elsevier .com/locate/bui ldenv

0360-1323/$ e see front matter � 2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.buildenv.2010.11.002

Building and Environment 46 (2011) 1038e1046


laboratory experiments are an important part of validating HAM-models, but of course the ultimate goal of hygrothermal buildingenvelopemodels is to predict the heat, air andmoisture response ofreal building components exposed to real climatic conditions. InNorway, an interesting large scale field study was conducted duringfour years in order to provide validation data for HAM-models [12].In total 16 different wall assemblies were investigated. The timberframe walls, insulated with 150 mm glass fibre, have variouscombinations of different types of interior lining (gypsumboard andwoodfibreboard), vapour barrier (with orwithout spunbonded foil)and exterior sheathing (wood fibreboard, gypsum board, spunbonded foil and asphalt impregnated wood fibreboard). However,the number of different test sections impedes measuring thehygrothermal behaviour of each assembly in detail.

Other examples of comparison of experimental data withnumerical simulations about hygrothermal performance of buildingwalls can be found in literature, for example [13e16]. But, in mostcases the experiments were only intended to illustrate specificcapabilities of the simulationmodels and lack generality. Due to theabsence of details and sometimes the simplicity of the test config-urations, they can hardly be considered as a validation data.

This paper describes a carefully planned full scale experimentwith a goal of generating valuable data that can be used for vali-dating HAM simulation models under real climatic conditions.Considering the extent of reports based on numerical simulation ofhygrothermal performance of building envelopes, there is anobvious need in the research community for such a comprehensivedata set for model validation purposes.

The test case in this report is rather unique in its complexity,configuration and follow-up. The experimental set-up is a lightweight building envelope wall. The global wall is subdivided intothree sections which differ in air and vapour tightness of the inte-rior finishing. The outside and interior environmental conditionsare identical which makes it possible to compare the dynamichygric behaviour between the three configurations. The experiment

ran for about two years. During this time, the air pressure differenceacross the wall and the temperatures, humidity and heat fluxeswithin the construction were continuously logged. In addition, theevolution of the moisture content of the exterior sheathing speci-mens, a bituminous impregnated soft wood fibreboard, is quanti-fied by regularly weighing nine removable specimen of the board,which are placed in the global sheathing wall layer.

The first part of this paper describes the global experimentalset-up and provides measured hygrothermal material properties ofthe used building components. Then the measured data are pre-sented and analysed.

2. Experimental setup

2.1. Test setup

The experimental set-up is situated in the VLIET test building,located at K.U.Leuven, Belgium (50.88�N, 4.7�E). The test buildingitself was constructed for the comprehensive study of the hygro-thermal behaviour of building components under real climaticconditions [17]. In the building, a compartment has been isolatedconnecting the SW and NE façade (Fig. 1). The room has a length of6.5 m, a width of 1.8 m and a height of 2.68 m. Except the test wallin the NE façade all other walls, roof and floor are made air andvapour tight by using plywood in combination with Polyethyleneplastic foil.

The test wall at the NE-side is a light weight wooden construc-tion. The global wall has a surface area of 1.80 � 2.68 m2; and issubdivided into three vertical sections, each with a similar config-uration but different characteristics for the heat, air and vapourtransport.

The basic configuration of all three sections consists of woodenstuds (25 � 200 mm) with mineral wool insulation (200 mm) inbetween. At the outside this load-bearing construction is sheathedwith a bituminous impregnated soft wood fibreboard (18 mm),followed by a cavity (25mm) andwater resistant woodenmultiplexboard (18 mm). The cavity is ventilated naturally through openingsat the top and bottom of the outer board (Fig. 2). So far, the threeparts of the wall (i.e., left, central and right parts) are identical. Theinterior finishing, however, varies. The left part is finished withawooden interiorfinishing (12.5mm),which is relativelyair open. Itwill be referred to as ‘wooden section’. The right part has uncoatedgypsumboard (12.5mm) that is air tight but vapour open as interiorfinishing. This part will be referred to as ‘gypsum section’. Themiddle part has an air and vapour barrier (0.2 mm), unventilatedcavity (25 mm) and interior uncoated gypsum board finishing(12.5mm). The latter canbe seen as an air and vapour tight referencecase (see Fig. 4) and will be identified as ‘reference section’.

The choice of the exterior sheathing follows from the fact thatthe bituminous mixed wood fibreboard is very hygroscopic andallows capillary water transport. As a result, the fibreboardsheathing can buffer large amounts of moisture (see Fig. 7) andavoids in this way drainage of eventual moisture condensation inthe winter seasons. The fibreboard was constructed in such a waythat each part of the wall contained three removable specimens ofthe board (Fig. 2). A compressible polymer is fitted along the innerperimeter of the specimen in order to avoid unaccountable vapourand air leakage. The specimens are used to quantify the moistureevolution of the fibreboard in time. Theweight increase/decrease ofthe board has been measured on a regular basis.

The experiment ran from October 7th, 2008 till April 30th, 2010.During this period, the setup is adapted three times in order to studythe effect of several internal and external conditions on the pre-vailing HAM transport. The first modification on the test configu-ration was performed on November 20th, 2008 by introducing

Nomenclature

d building material thickness [m]r density [kg m�3]cp specific heat capacity at constant pressure

[J kg�1 K�1]u moisture content [kg kg�1]l thermal conductivity [W m�1 K�1]F relative humidity [e]q temperature [�C]sd water vapour diffusion-equivalent air layer

thickness [m]sd,dry water vapour diffusion-equivalent air layer

thickness at 33% relative humidity [m]lq, lo, lw, v, m, n, c1, c2, a, b model parameterskair air permeance [kg m�2 h�1 Pa�1]gair density of air flow rate [kg m�2 h�1]Δpair air pressure difference [Pa]

AbbreviationsHAM heat, air and moistureFS fibreboard sheathingMW mineral woolVR vapour retarderGB gypsum boardWF wooden finish

T.Z. Desta et al. / Building and Environment 46 (2011) 1038e1046 1039


a moisture source in the room and controlling the lower roomtemperature at 20 �C by means of infrared bulbs. These changesmade the test configuration a representative dwelling room.

After about three months of data collection on March 5th, 2009a ventilation system was introduced in order to study the effect ofair transport on the hygrothermal performance of the buildingwall.In this stage, measurement of inlet air velocity together with thepressure drop across the NE wall was started. The ventilationsystem shown in Fig. 3, consists of a small fan with constant

revolution rate installed in a polyethylene tube. The ventilationtube is extended into the room so that the air flow in the tubebecomes fully developed before it is supplied to the room avoidingabrupt air flow rate and pressure fluctuations in the room. Theventilation system creates a rather constant overpressure in theroom, increasing the pressure drop over the NE test wall. Howeverno significant air flow through the wall could be attained due to the

Fig. 2. Outside view (left) and inside view (right) of the test wall. The outside viewshows how the nine fibreboard sheathing specimens can be removed to measure theirweight increase/decrease on a regular basis.

Fig. 1. The VLIET test building at K.U.Leuven, Belgium and the schematic view of the compartment with the test wall in the NE façade.

Fig. 3. The ventilation system that penetrates through the South-West (SW) wall.

T.Z. Desta et al. / Building and Environment 46 (2011) 1038e10461040


low air permeance of all three sections of the test wall. Finally,about seven months after the ventilation system was activated, onOctober 11th, 2009 the interior finishing of the wooden section ofthe wall was made more air open to investigate the effect ofincreasing air flow rate on the hygroscopic response of the wall.Increasing the air permeance of the wooden finishing was achievedby drilling very small holes (1.5 mm in diameter) in an equidistantfine grid through it (328 holes/m2).

On the surfaces and within the three sections of the test walla total of fifty-seven thermocouples, eighteen humidity sensors andtwelve heat flux sensors are mounted in a structured manner(Fig. 4). The sensors cover the most important positions of the walland are placed in three rows i.e., top (t), at 45 cm from the ceiling,middle (m) at 136 cm from the floor and bottom (b), at 45 cm fromthe floor. In the room nine thermocouples and three relativehumidity sensors are hanged on a grid. Moreover the pressuredrops across the NE and SW wall together with the air velocity inthe ventilation duct are measured. Outside conditions are loggedwith a weather station and meteorological mast [18].

All full scale data, i.e., temperature, relative humidity, heat flux,pressure drop and air flow velocity are recorded every 10 min.

2.2. Material properties

One of the key requirements for successful validation ofnumerical predictions is the availability of correct material prop-erties [8]. Therefore all the essential hygrothermal properties of theapplied materials were measured except the heat capacity of all the

materials and moisture properties of the mineral wool which istaken from [19]. All other properties have been determined at thedivision of Building Physics at K.U.Leuven on different samples andmean values are reported. When possible also informationregarding the spread of measured values is given. Table 1 shows themeasurement accuracy and other details of sensors and balancesthat may help to estimate the equipment error.

2.2.1. Heat transfer properties2.2.1.1. Thermal conductivity. The thermal resistance of thebuilding components is measured with the heat flow meter appa-ratus as described in the standard guidelines [20]. The results areplotted in Fig. 5

In Fig. 5 the linear curve is given by:

l ¼ lqqþ lo (1)

Parameters lq and lo are supplied in Table 2.The highly hygroscopic exterior fibreboard can absorb a lot of

moisture in the winter season. Consequently the thermal conduc-tivity in the setup can deviate for this material from the dry statevalue. Therefore the thermal conductivity of the board at variablemoisture content is measured and presented in Fig. 6.

In Fig. 6 the linear curve fit is given by:

l ¼ lwwþ lqqþ lo (2)

Note that in Eq. (2) the sumof the last two terms is they-interceptof Fig. 6 (the thermal conductivity of the dry fibreboard as describedin Eq. (1)). The model parameters lw, lq and lo are given in Table 2.

rosnesytidimuHrosnesxulFtaeH

selpuocomrehTerbifnoitalusnI

nidliuB g sllaw

draobmuspyGdraobrebiF

gnihtaehs

xelpitluMdnegeL

reirrabruopaV

gdraobnedooW

Exte

riorytivaC

oom

inte

rior

itdW itfR itG

R

oo cesne no e cesecnere no cesmuspy no

Fig. 4. Section view of global wall configurations and positioning of the measurement sensors.

Table 1Specification and accuracy of sensors and measuring instruments.

Test type Instrument Manufacturer Model Measurement range Accuracy

Full scale experiment at the test building Thermocouple (type T) Thermo-electric Belgium P-26-TT-IEC Max. 105 �C �0.1 �CRelative humidity sensor Honeywell, Belgium HIH-4000 0e100% �0.5%Heat flux sensor Hukseflux Thermal sensors,

The NetherlandsHFP01 �2000e2000 W/m2 �5% of readings

Pressure sensor Halstrup walcher Gmbh, Germany P92 0e25 Pa �4% of readingsVelocity sensor(hot wire anemometer)

TSI incorporated, U.S.A. Model 8475 0.05e2.5 m/s �3% of readings

Balance Mettler Toledo, Switzerland PB1502-L 0e1.51 kg �0.01 gSmall scale laboratory setup for material

property measurementBalance Mettler Toledo, Switzerland XS1003s 0e1.01 kg �0.001 gPressure sensor Energy conservatory, U.S.A. 4DG-700 0e1400 Pa �1% of readingsAir flow meter Vögtlin, Germany V-100 0.02e0.9 m3/h �2% of readings



The variations of the thermal conductivity measurements are verysmall; therefore the error bar plots are not shown in Fig. 6.

2.2.1.2. Specific heat capacity and density. In Table 2, the densities ofall materials are measured by the Laboratory of Building Physics atK.U.Leuven and the heat capacities are taken from literature [19]and product manual of the suppliers.

2.2.2. Hygric properties2.2.2.1. Moisture capacity. The moisture capacity of the buildingcomponents is measured by determining the hygroscopic sorptioncurves according to the guidelines described in [21]. The results areshown in Fig. 7.

In Fig. 7 the curve fit between moisture content and relativehumidity is given by:

u ¼ v�1þ ðm$lnðfÞÞn�1�n

n (3)

The model coefficients of Eq. (3) are presented in Table 2.

2.2.2.2. Water vapour transmission. The determination of watervapour transmission properties of the building components wasperformed according to the international standard procedure at23 �C [22].

Fig. 8 shows the water vapour diffusion-equivalent air layerthickness (sd) value for the building components. The curve fit inFig. 6 has a form of:

sd ¼ 11

sd;dryþ c1

dec2f

(4)

The model parameters of Eq. (4) for the test wall componentsare provided in Table 2.

2.2.2.3. Liquid water transport. Liquid water transport in the inte-rior finishing and mineral wool insulation is unlikely to occur.Therefore only the water transport parameters for the moisturebuffering fibreboard sheathing material were measured. Thecapillary moisture content, thewater adsorption coefficient and thevacuum saturation moisture content of the fibreboard were foundto be 0.6 kg/kg, 0.0052 kg/m2/Os and 3.7 kg/kg respectively.

2.2.3. Air transport propertiesThe air permeance of the different sections of the test wall is

determined by measuring the permeance of its components ina small scale laboratory setup. The laboratory setup which is

Fig. 5. Thermal conductivity as a function of temperature for building components ofthe test wall (FS ¼ fibreboard sheathing, MW ¼ mineral wool, GB ¼ gypsum board,WF ¼ wooden finish).

Table 2Overview of the material properties.

MaterialParameter

FS MW VRa GB WF(undrilled)

WF(drilled)

d 0.018 0.2 0.0002 0.0125 0.015 0.015r 274 20 1000 690 400 419cp 2068 840 1400 870 2000 2000sd,dry 0.14 0.3 20 0.13 2.20 2.11lq 1.275E-4 7.25Ee5 0 2.382E-4 1.884E-4 1.884E-4lo 0.0469 0.0347 0.15 0.206 0.0909 0.0909lw@ 1.5 �C 0.021 0 0 0 0 [email protected] �C 0.0273 0 0 0 0 0v 1 0.1 e 0.2 0.26 0.26m �1888.4 �3000 e 0.1 �19.48 �19.48n 1.4 1.5 e �334.40 1.55 1.55c1 7.18E-13 2.3E-6 e 5E-6 1.12E-6 2.98E-6c2 27.61 15.3 e 10.35 14.44 12.83a 0.16 4.66 e 0.0022 0.02 0.47b 1 0.92 e 0.97 0.89 0.84

a The vapour barrier, polyethylene foil is an air tight material with very lowhygroscopic and vapour transmission potential.

250.0

450.0 C°5.1tatnemerusaemC°5.1tatifevruc

C°5.11tatnemerusaem511tatifevruc °C

50.0

250.0

W.m

-1.K

-1]

evruc tif ta 11 . C5

6400

840.0 [W

0 50.0 1.0 51.0 2.0 52.0 3.0440.0

0. 640

]k/k[w k[ g k/ g]

λ

Fig. 6. Thermal conductivity as a function of moisture content for the fibreboardsheathing.

draob nedoowgnihtaehs draoberbif

20

4.0

kg/k

g] derusaemtif evruc

derusaemtif evruc

0

0.2u [

4.0/kg]

draob muspyg

derusaemif

loow larenim

derusaemtif

0 5.0 10

2.0u [k

g evruc if t

0 5.0 1

evruc tif

]-[ ]-[ φ φ

Fig. 7. Sorption isotherm of building components of the wall measured at 23 �C.



described in [23], is an air tight chamber inwhich one of the faces isdesigned to be perfectly sealed with a test sample specimen (35 cmby 35 cm). The sealing is attained by placing a compressible rubberbetween the sample and the chamber. Later on, the casing thatholds the sample is screwed with the rest of the chamber (Fig. 9).After sealing, compressed air with variable pressure is injected intothe chamber and the corresponding pressure drop and air flow rateacross the sample are recorded simultaneously.

The permeance of the building components is obtained byfitting the experimental data to Eq. (5).

gair ¼ kairDp; with kair ¼ aDpb�1 (5)

Parameters a and b are determined from the measured air flowrate and pressure drop across the sample as presented in Fig. 10,and are provided in Table 2. The measured air flow rate through the

gypsum board sample was very small and is not shown in Fig. 10.However its curve fit parameters are provided in Table 2.

The variations of the airflow measurements are very small;therefore the error bar plots are not shown in Fig. 10.

3. Hygrothermal measurement results

Fig. 11 plots the evolution of the measured boundary conditionsduring the experiment i.e., the indoor and outdoor temperatureand relative humidity together with the measured pressure dropacross the NE wall. As can be seen, in summer time the roomtemperature deviates from the set point value of 20 �C, since onlya heater and no cooler is installed in the test room. After thewooden section of the test wall is drilled, the humidity in the roomdecreases, demonstrating the enhanced moisture transport fromthe room to the outside.

2

5.2

50

1

5.1

s d [m]

)dellirdFW(tifevruc)dellirdFW(tnemerusaem

)dellirdnuFW(tifevruc

0

. )dellirdnuFW(tnemerusaem

3.0

2.0

52.0evruc tif )SF(

)SF(tnemerusaem)BG(tifevruc

)BG(tnemerusaem

10

51.0

s d [m] )WM(tifevruc

)WM(91.fermorf

50.0

.

0 2.0 4.0 6.0 8.0 10

]-[φ

Fig. 8. Water vapour diffusion-equivalent air layer thickness of building componentsof the test wall (FS ¼ fiberboard sheathing, MW ¼ mineral wool, GB ¼ gypsum board,WF ¼ wooden finish).

Fig. 9. Laboratory setup to determine the air permeance of building materials.

003]

001

002

003

r [kg.

m-2

.h-1

]

WM

0g air

51

02

]

)dellirdnu(FWFW ( dellird )

01

51

ir [kg.

m-2

.h-1

)(SF

0 02 04 06 080

5g a

]P[p ria P[ a]Δ

Fig. 10. Measurement results of air flow rate versus pressure drop for the differentbuilding components of the test wall.

ed

urc

e

Le

ft w

all

nterio

r d

rille

Mo

istu

re so

u

sta

rted

Ve

ntila

tio

n

ac

tiva

te

d

in

M V

a

01

02

[°C]

edisni0

edisniedistuo

9.0

6.07.08.0

φ [-]

edisniedistuo5.0 edistuo

01

r [Pa]

11/91 10/21 30/60 50/40 70/92 01/90 10/60 30/20 40/820

etad /yad[ month]

p air

llawENehtssorca

90028002 0102

Δθ

Fig. 11. Inside and outside temperature (mean values), inside and outside humidity(mean values) together with the air pressure drop across the wooden section of thewall at the middle row.



As mentioned in Section 2.1 and shown in Fig. 2 nine removablespecimens are placed in the exterior sheathing. These specimensare weighed on a regular basis and their moisture content wasquantified. Fig. 12 compares the moisture content evolution of thenine fibreboard specimens with the measured boundary condi-tions, i.e., indoor and outside cavity vapour pressure together withair flow rate through the wooden section of the wall. The air flowthrough the wall is calculated by multiplying the pressure dropacross the wall with the air permeance of the wooden section ofwall. The air leakage through the reference and gypsum sections isconsidered to be negligible since the interior finishing of thesesections of thewall (gypsum board and/or vapour barrier) is knownto be airtight.

The total air permeance of the wooden section (and hence of theglobal test wall) is calculated by combining the measured per-meance of the different layers (Eq. (6)) in series arrangement.

kwall ¼�

1kWF

þ 1kMW

þ 1kWB

��1

(6)

Note that air permeance is determined on dry specimens.Therefore, in this article the effect of moisture on the air perme-ability is not reported. Furthermore, the multiplex at the façade ofthe test wall is drilled at the top and bottom for cavity ventilationpurpose. Hence the air permeance of the multiplex is not incor-porated in Eq. (6).

As can be seen in Fig. 12, the measured inside temperature istypical for a dwelling room in temperate climatic zone, while theroom vapour pressure is rather high due to the installed moisture

source (Fig. 11). This high vapour pressure was maintained in orderto study more extreme conditions.

Moreover in Fig. 12 the impact of the changes in the test set-up(the introduction of the moisture sources inside the room, theactivation of the ventilation system and the drilling of the woodensection of the wall) on the hygrothermal behaviour of the wall isclearly visible.

Comparing the evolution of the moisture content of the fibre-board specimens, distinct different results are observed for thethree wall compositions. The introduction of the moisture sourceon November 20th, 2008 combined with the winter period induceda steady moisture content rise in all specimens. The increase of themoisture content is however far more pronounced for the samplesin the gypsum and wooden section of the test wall. The gypsumsection of the test wall, which is the most vapour open, shows inthe first winter period the highest moisture increase. The increasein moisture content is very limited for the reference sectionbecause of the vapour barrier film. The ventilation system, installedon March 5th, 2009 does not seem to affect the course of moisturetransfer significantly, only a slight increase in moisture content ofthe fibreboard specimens was observed in the first winter afteractivating the ventilation system. This limited effect can be attrib-uted to the air tightness of all three parts (Table 2 and Eq. (5)).

After the first winter all the fibreboard sheathing specimensdried out again during summer 2009. The drilling of the woodensection of the wall on October 11th, 2009 significantly increases themoisture transfer by enhancing the convective mass flow ratethrough the wall. As a result, the exterior fibreboard sheathingspecimens at the wooden section gained twice and four times as

eft w

all

or d

rille

d

la

tio

n

va

te

d

ure

s

ou

rc

e

tarted

0002

0003

Pa]

Le

in

teri

Ven

ti

ac

tiv

Mo

istu

st

0001

0002

p vap [P edisni

ytivac

0203

C]

001 [°

C edisniytivac

12

[kg/

h]

llawENehtssorca

01

g air

53.0

4.0 pot-tfelelddim-tfel

2.0

52.0

3.0

53.0

w [k

g/kg

]

mottob-tfelpot-lartnec

elddim-lartnecmottob-lartnec

potthgir

50.0

1.0

51.0

w thgir - potelddim-thgirmottob-thgir

11/91 10/21 30/60 50/40 70/92 01/90 10/60 30/20 40/82]htnom/yad[etad

90028002 010290028002 0102

θ

Fig. 12. Measurement results of inside and outside cavity vapour pressure (mean values), inside and outside cavity temperature (mean values), air flow rate through the NE wall(calculated using Eqs. (5) and (6)) and moisture content evolution of the nine removable fibreboard specimens (measured).



much moisture as those of the ones at the gypsum and referencesections respectively (Fig. 12).

The specimens at the top, middle and bottom rows within thesame section of the test wall showed a difference in moisturecontent between one another especially in the winter seasons. Thismay be attributed to air looping in the outer cavity due to naturalventilation and/or buoyancy effects. This hypothesis is verifiedwhen analysing the vertical temperature profiles in the cavity. Asan example Fig. 13 plots the outer cavity temperatures at the top,middle and bottom rows for the first week of January 2010.

As can be seen, for all the three sections of the test wall, sensorsat the top row have higher temperature readings followed bysensors at the middle row. Buoyancy drifts warmer air to the top,increasing the cavity temperature in upward direction. This airlooping in the cavity can also cause disproportional convectivemass transport at the surface of each row, resulting in a moisturecontent difference between the fibreboard specimens located inthe same section of the test wall but at different rows (Fig. 12).

4. Applicability and limitations of the generated data

The test setup was constructed in order to generate measure-ment data for two purposes.

1. To gain insight in the hygrothermal responses of light weightbuilding components under real climatic condition.

2. To generate a validation data set for HAM models. In thiscontext [24,25] used part of the comprehensive data set anddemonstrated its usefulness for such applications.

However there are limitations in the generated data that need tobe addressed:

(a) The generated data are intended mainly to validate dif-fusiveeconvective transport phenomena. It hardly providesinformation about radiation effects because the test wall isoriented in NE direction.

(b) As discussed at the end of Section 3, due to the buoyancy effectsand natural ventilation, the outside boundary conditions insidethe vented cavity are not the same for the top, middle and

bottom rows of the wall. Consequently, simulation of the wallconfiguration should be done in a 2D manner including thebuoyancy and natural ventilation flow in the cavity. Thishowever is a complicated approach. Since the cavity is equip-ped with temperature and humidity sensors at differentheights, the measured data inside the cavity can be used asoutside boundary conditions. In this way, it is possible tomodeland validate the HAMphenomena row by row in a 1D approachseparately. This is a more realistic and easier way of handlingthe flow problem at hand since attaining information about theeffect of the buoyancy and the natural ventilation on the cavityconditions is very difficult, if not impossible. Therefore theauthors suggest to use the bottom row of the collected data forvalidation purposes.

The data measured in the three parts of the wall at the bottomrow and corresponding sensor locations are available for researchpurposes and can be downloaded from theweb page of the BuildingPhysics section at K.U.Leuven http://bwk.kuleuven.be/bwf/Research.

5. Conclusions

A test setup was build in the VLIET test building at the Labora-tory of Building Physics of the K.U.Leuven both to generate modelvalidation data and to analyse the hygrothermal behaviour of lightweight building walls under real climatic conditions.

The test wall was subdivided in three sections which differ ininterior finishing. The wall was equipped with all necessarymeasurement sensors to follow the heat and moisture response ofthe wall in detail. Furthermore a bitumen impregnated woodfibreboard was used as exterior sheathing so that the amount ofaccumulated moisture could be measured by weighing smallspecimen of the board. It was found that as long as all walls weresufficiently air tight the amount of accumulated moisture inthe winter season was limited and it dried out in the summer forthe given configuration. Furthermore, the amount of absorbedmoisture was found to be directly proportional to the vapourpermeability of the interior finishing. As soon as one of the wallswas made more air permeable, convective flow became thedominant cause of absorbed moisture build-up.

With the notion of supplying informative data for model vali-dation purposes, the used test setup is fully described in a detailedmanner and almost all material properties of the building compo-nents of the test wall were measured except that of the mineralwool which was taken from the literature. Furthermore thehygrothermal experimental data which were collected for abouttwo years are made available. It is the author’s belief that thecompiled report is a valuable contribution that lends a hand forHAMmodellers in their model development and validation effort inparticular and for the wider research community in general.

Acknowledgement

The results in this paper have been mainly obtained withinthe framework of the project IWT, SBO 050154 “Heat, air andmoisture performance engineering: a whole building approach”and some material properties were obtained with in a Ph.D. grantfrom IWT, 81153 both funded by the Flemish Government. Thisfinancial support is gratefully acknowledged.

References

[1] Li Q, Rao J, Fazio P. Development of HAM tool for building envelope analysis.Building and Environment 2009;44(5):1065e73.

trap tfeL

4-

2-

0

[°C

]

pot elddim mottob

2-

0

[°C]

trap lartneC

4- pot elddim mottob

0trap thgiR

1 2 3 4 5 6 7

4-

2- [°C

]

pot elddim mottob

1 2 3 4 5 6 7]yad[ emiT

θθ

θ

Fig. 13. Daily averaged cavity temperature at the top, middle and bottom rows for thethree sections of the wall in the first week of January 2010.



[2] Janssen H, Blocken B, Roels S, Carmeliet J. Wind-driven rain as a boundarycondition for HAM simulations: analysis of simplified modelling approaches.Building and Environment 2007;42(4):1555e67.

[3] Glaser H. Temperatur und Dampfdruckverlauf in einer homogen Wand beiFeuchteausscheiding. Kältetechnik; 1958:174e81.

[4] European Committee for Standardization (ECS). Hygrothermal performance ofbuilding components and building elements e internal surface temperature toavoid critical surface humidity and interstitial condensation e Calculationmethods; 2001. ISO 13788.

[5] Künzel H. Simultaneous heat and moisture transport in building components.One- and two-dimensional calculation using simple parameters. PhD-thesis,Universität Stuttgart, Germany: 1994.

[6] Grunewald J. Diffusiver und konvektiver stoff- und energietransport inkapillarporösen baustoffen. PhD-thesis, Technische Universität Dresden,Dresden, Germany: 1997.

[7] Pedersen Rode C. Combined heat and moisture transfer in building construc-tions. PhD-thesis, Technical University of Denmark; 1990. Report no. 214.

[8] Roels S, Talukdar P, James C, Simonson CJ. Reliability of material datameasurements for hygroscopic buffering. International Journal of Heat andMass Transfer 2010;53:5355e63.

[9] Adan O, Brocken H, Carmeliet J, Hens H, Roels S, Hagentoft CE. Determinationof liquid water transfer properties of porous building materials and devel-opment of numerical assessment methods: introduction to the EC Hamstadproject. Journal of Thermal Envelope and Building Science 2004;27(4):253e60.

[10] Hagentoft CE, Kalagasidis AS, Adl-Zarrabi B, Roels S, Carmeliet J, Hens H, et al.Assessment method of numerical prediction models for combined heat, airand moisture transfer in building components. Benchmarks for one-dimen-sional cases. Journal of Thermal Envelope and Building Science 2004;27(4):27e352.

[11] James C, Simonson CJ, Talukdar P, Roels S. Numerical and experimental dataset for benchmarking hygroscopic buffering models. International Journal ofHeat and Mass Transfer 2010;53:3638e54.

[12] Geving S, Uvsløkk S. Moisture conditions in timber frame roof and wallstructures, test house measurements for verification of heat-, air and moisturetransfer models. Project Report 273e2000. Oslo, Norway: BYGGFORSK,Norwegian Building Research Institute; 2000.

[13] Haupl P, Jurk K, Petzold H, Inside thermal insulation. In: Proceedings of the 2ndinternational conference on Building Physics 14e18 Sept. 2003. Antwerpen,Belgium: 2003. p. 463e469.

[14] Roels S, Depraetere W, Carmeliet J, Hens H. Simulating non-isothermal watervapour transfer: an experimental validation on multi-layered buildingcomponents. Journal of Building Physics 1999;23:17e40.

[15] Kalamees T, Vinha J. Hygrothermal calculation and laboratory tests on timber-framed wall structures. Building and Environment 2003;38(5):689e97.

[16] Geving S, Karagiozis A, Salonvaara M. Measurements and two-dimensionalcomputer simulations of the hygrothermal performance of a wood frame wall.Journal of Building Physics 1997;20:301e19.

[17] Roels S, Deurinck M. The effect of a reflective underlay on the global thermalbehaviour of pitched roofs. Building and Environment 2010;46(1):134e43.

[18] Abuku M, Blocken B, Roels S. Moisture response of building façades to wind--driven rain: comparison of field measurements and numerical simulations.Journal of Wind Engineering and Industrial Aerodynamics 2009;97(5e6):197e207.

[19] Kumaran K. Material properties e Final report, vol. 3. Belgium, KU Leuven:LaboratoriumBouwfysica,DepartementBurgerlijkeBouwkunde; 1996. pp. 135.

[20] International standard 1991. Thermal insulation e Determination of steady-state thermal resistance and related properties e Guarded hot plate appa-ratus, ISO 8302:1991.

[21] European Standards 2000. Building materials; Determination of hygroscopicsorption curves, EN 12571:2000.

[22] International standard 2001. Hygrothermal performance of building materialsand products e Determination of water vapour transmission properties, ISO12572:2001.

[23] Langmans J, Klein R, De Paepe M, Roels S. Potential of wind barriers to assureairtightness of wood-frame low energy constructions. Energy and Building2010;42(12):2376e85.

[24] Desta TZ, Roels S, The influence of air in the transport of heat and moisturethrough a building wall. In: Central European conference on building PhysicsSeptember 13e15, 2010. Cracow, Poland: 2010. p. 193e200.

[25] Desta TZ, Roels S, Experimental and numerical analysis of heat, air andmoisture transfer in a light weight building wall. In: Thermal performance ofthe exterior envelopes of whole building XI international conference, Clear-water Beach, December 5e9, 2010, Florida, U.S.A. 2010. [Accepted].


Author's personal copy Experimental data set for validation of heat, air and moisture transport...

Documents

Transcript of Author's personal copy Experimental data set for validation of heat, air and moisture transport...