See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/226059351

Multiphysics modeling of building physical constructions

Article  in  Building Simulation · March 2011

DOI: 10.1007/s12273-011-0020-0

CITATIONS READS

15 146

1 author:

Jos Van Schijndel

ASML
218 PUBLICATIONS   1,226 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Microclimate for cultural heritage View project

Multi-physics Multi-scale Computational Modeling View project

All content following this page was uploaded by Jos Van Schijndel on 07 October 2016.

The user has requested enhancement of the downloaded file.

Multiphysics Modeling of Building Physical Constructions

Abstract

The paper presents an overview of Multiphysics Applications using Comsol for building physical

constructions simulation. The overview includes three main basic transport phenomena for building

physical constructions: (1) Heat transfer; (2) heat and moisture transfer and (3) heat, air and moisture

(HAM) transfer. It is concluded that full 3D transient coupled HAM models for building physical

constructions can be build using Comsol. Regarding the heat transport, neither difficulties nor

limitations are expected. Concerning the combined heat and moisture transport the main difficulties

are related with the material properties but this seems to be no limitation. Regarding the HAM

modeling inside solid constructions, there is at least one limitation: The validation is almost

impossible due to limitation of measuring ultra low air velocities of order µm/s.

1. Introduction

Building spaces are separated from each other and from the outdoor climate by partitions: inside walls

and the building envelope, facades, roofs, floors. The building envelope is subjected to a strongly

fluctuating outdoor climate: sunshine, rain, wind and air temperatures. The performance demands

required for these structures depend on the requirements for comfort in the rooms of a building, the

energy needed to realize the desired indoor climate, the air quality and air humidity of this climate, the

durability, maintenance, use of materials and the recyclability of these structures. In the past the design

of these structures was lead by experience. Due to the more rigid requirements on performance and the

enormous increase of new building techniques, new materials and new building shapes this reliance on

experience is often not applicable any more. The result may be building damage, a bad indoor climate

and an unnecessary high energy consumption. Therefore the knowledge of heat and moisture transport

through building structures and joints is increasingly important for building design. A clear illustration

of this fact is the building code with its abundant regulations related to building physics. The

knowledge, insight and prediction models of building physics are indispensable for the realization of
high quality buildings that satisfy the required performances [1-3]. Topics related with building

physics published in previous International Journals of Engineering Education are provided in

references [4-7].

Many scientific problems in building physics can be described by PDEs [8]. There are a lot of

software programs available in which one specific PDE is solved. They are developed in order to get

the simulation results in a short time and most often a lot of effort has been put into the simplicity of

input of data, e.g. geometrical data. A disadvantage is that they often are not very flexible when the

user wants to change or combine models. Another drawback is that they most often act as black boxes.

Another category of commercially available software like Comsol [9] is developed specifically for

solving PDEs where the user in principle can simulate any system of coupled PDEs. Practical

physics/engineering problems, related with building physics [10, 11], in the area of heat transfer [12-

21], moisture transport [22-29], fluid dynamics [30-39] and structural mechanics [40-43] can be

solved with the software. One of the main advantages of Comsol is that the user can focus on the

model (PDE coefficients on the domain and the domain boundary) and does not have to spend much

time on solving and visualization. The scientist can concentrate on the physics behind the models and

the engineer can calculate details for designing purposes using validated models.

How Comsol works

Comsol solves systems of coupled PDEs (up to 32 independent variables). The specified PDEs may be

non-linear and time dependent and act on a 1D, 2D or 3D geometry. The PDEs and boundary values

can be represented by two forms. The coefficient form is as follows *:

∂u
da − ∇ ⋅ (c∇u + αu − γ ) + β ∇u + au = f in Ω ( 1a )
∂t

n ⋅ ( c∇ u + α u − γ ) + q u = g − λ on ∂Ω ( 1b )

hu = r on ∂Ω ( 1c )

*
The symbols provided by the Comsol modeling guides are also used here.
The first equation (1a) is satisfied inside the domain Ω and the second (1b) (generalized Neumann

boundary) and third (1c) (Dirichlet boundary) equations are both satisfied on the boundary of the

domain δΩ. n is the outward unit normal and is calculated internally. λ is an unknown vector-valued

function called the Lagrange multiplier. This multiplier is also calculated internally and will only be

used in the case of mixed boundary conditions. The coefficients da , c, α ,ß, γ, a, f, g, q and r are

scalars, vectors, matrices or tensors. Their components can be functions of the space, time and the

solution u. For a steady state system in coefficient form da = 0. Often c is called the diffusion

coefficient, α and ß are convection coefficients, a is the absorption coefficient and γ and f are source

terms. Other forms of the PDEs and boundary conditions available in Comsol are the general and

weak forms. In this paper only the coefficient form is used.

Example: Comsol code and results of a 2D steady state thermal bridge

A 2D steady state thermal bridge problem is used as an example of how Comsol works. In Figure 1 the

geometry of the 2D thermal bridge problem is shown. Figure 1 represents half of a horizontal section

of an insulated external wall. B7 is the axis of symmetry. The internal wall, surrounded by b5, b6 en

b7 intersects the insulation and forms a so-called thermal bridge. In Table I the lengths and boundary

conditions of each boundary segment are given. The PDE model for the inside of the domain is given

by:

∇ ⋅ ( K∇ T ) = 0 (2)

Where K is thermal conductivity and T is temperature. Using the coefficient form (1a) and the PDE

model (2), it follows that u equals T and the coefficients of (1a) are all zero (a= da = f= α= ß = γ =0 )

except c. The c coefficient equals the heat conductivities at the sub domains concrete (Kconcr) and

insulation (Kinsul). The boundary values are heat fluxes and so the Neumann condition is applied. For

example, boundary condition b1: f = hce*(Te-T) is represented by taking q = hce , g = hce*Te in eq.
(1b). Note that the term n·c▼u in (1b) represents the heat flow into the domain and is calculated

internally and the term λ in (1b) is zero because mixed boundary conditions are not applied in this

example.

In Comsol there are generally speaking two ways to implement problems: by user interface or by

scripts. The user interface can be seen as a tool that generates scripts. In order to show how it works,

Figure 2 shows the complete Comsol script to solve the example problem. The default values of all

PDE and boundary coefficients are 0. Also some comments (%) are included for better understanding

of the code. The initial mesh is presented in Figure 3, and the solution in Figure 4. This example

shows the transparency and flexibility of PDE models in Comsol.

The reader should notice that in this Section we focused on how Comsol works. The objective of the

paper is to provide an overview that includes three main basic transport phenomena for building

physical constructions: Heat transfer (Section 2), heat and moisture transfer (Section 3) and heat, air

and moisture transfer (Section 4). General conclusions and future research is provided afterwards.

2. Heat transport

The (2D) heat transport modeling is already presented in the previous Section. The extension of 2D

steady state modeling to the ultimate 3D transient modeling of building constructions, is straight

forward and is quite easy to teach to students. This section shows therefore exemplary results obtained

by students. They implemented a 3D application from a real case from the practice and evaluated the

dynamic 3D thermal results using the related performance indicators. The real case consists of an

outdoor facade that is formed by 250 mm concrete at the outside, 80 mm insulation material at the

inside, finished by gypsum board at the inside surface. The internal floor constructions were supported

by steel IPE 300 beams. The steel beams are forming 3D thermal bridges by penetrating the inside

insulation. Furthermore the steel beams are not protected against fire. Table II shows the materials

properties and Figure 5 presents the vertical sections of the construction. The students calculate the

(steady state) 3D thermal bridge effect of the uninsulated steel beams penetrating the insulation

material. Furthermore they use the already mentioned performance indicators but now applied to a 3D
construction. They make two animations showing the most critical place during the summer and

during the winter. In the real case, it was decided that the steel beam should be insulated. The

graduates are asked to make a proposal for the thickness and kind of insulation under the condition

that the RH near the coldest surface should not exceed 70 %RH. Furthermore they need to show that

their design meets this requirement. Figure 6 shows an exemplary result obtained by the students.

We conclude that Comsol is a very efficient tool to simulate 3D transient heat transport in building

constructions for both research and education.

3. Heat and moisture transport

3.1 Modeling

The heat and moisture transport can be described by the following PDEs [11]:

∂T
CT = ∇ ⋅ ( K11∇T + K12∇LPc)
∂t (5)
∂LPc
C LPc = ∇ ⋅ ( K 21∇T + K 22∇LPc)
∂t
With

LPc =10 log( Pc)

CT = ρ ⋅ c
K11 = λ
∂Pc Mw (6)
K12 = −llv ⋅ δ p ⋅ φ ⋅ ⋅ Psat ⋅ ,
∂LPc ρ a RT
∂w ∂Pc
C LPc = ⋅
∂Pc ∂LPc
∂Pc ∂Pc Mw
K 22 = − K ⋅ − δ p ⋅φ ⋅ ⋅ Psat ⋅ ,
∂LPc ∂LPc ρ a RT

∂Psat
K 21 = δ p ⋅ φ ⋅ ,
∂T

Where t is time [s]; T is temperature [oC]; Pc is capillary pressure [Pa]; ρ is material density [kg/m3]; c

is specific heat capacity [J/kgK]; λ is thermal conductivity [W/mK]; llv is specific latent heat of

evaporation [J/kg]; δp vapour permeability [s]; φ is relative humidity [-]; Psat is saturation pressure
[Pa]; Mw = 0.018 [kg/mol]; R = 8.314 [J/molK]; ρa is air density [kg/m3]; w is moisture content

[kg/m3];K is liquid water permeability [s].

The material and boundary functions are calculated using Matlab. These functions are used to convert

measurable material properties such as K, φ, δp and λ which are depent on the moisture content into

PDE coefficients which are dependent on the LPc and T. This is schematically shown in Figure 7. The

results for two materials (based on HAMstad benchmark 1) are presented in Figure 8. For each

material and at each point the vapour pressure can be calculated using a similar corresponding

function.

3.2 Verification

Benchmarks are important tools to verify computational models. In the research area of building

physics, the so-called HAMSTAD (Heat, Air and Moisture STAnDardization) project is a very well

known reference for the testing of simulation tools. This Section summarizes the results for the

HAMSTAD benchmark no.1: ‘Insulated roof’. The roof structure is analyzed in 1D regarding dynamic

heat and moisture transport. The thermal insulation is facing the interior and a there is moisture barrier

facing the exterior. The structure is perfectly airtight. Figure 9 shows the structure. The governing

PDE equations are given by (1). Figure 8 provides the PDE coefficients for load bearing material a and

insulation material b. The boundary conditions are at the (i)nternal (connected with material a):

q = hi ⋅ (Ti − T ) + llv ⋅ β ⋅ ( pi − p )
(7)
g = β ⋅ ( pi − p )

and (e)xternal (connected with material b):

(8)
q = he ⋅ (Te − T )
g =0

Where q is heat flux [W/m2]; h is convective heat transfer coefficient [W/m2K]; β is vapour transfer

coefficient [kg/Pa.m2s]; p is vapour pressure [Pa]; g is moisture flux [kg/m2s]; The boundary
conditions are hourly based values provided by benchmark. The results are shown in Figures 10 and

11. Confronting these results with the benchmark, it is concluded that the results are within the

provided bandwidths. Other benchmarks will be evaluated in near future. Moreover, the ‘benchmark

1’ model using COMSOL [9] and companying report are published at the HAMLab [45] research and

education website.

3.3 Validation

The data set is part of the measurement program at the Hunting Lodge St. Hubertus site, performed

during 2006-2007 by Briggen [46]. Details of this project can be found in Briggen et al [47]. One of

problems seemed to be high moisture contents at the inside surface of the façade of the tower. The

construction of this façade is shown in Figure 12. The outside climate conditions were measured by a

weather station within 50m from the building. The inside air temperature and relative humidity were

measured using standard equipment (see Figure 13). A representation of inside surface conditions

were obtained by placing a small box (5cm x 5cm x 1cm) against the wall and measure the air

temperature and relative humidity inside. The estimation of the measurement error of this method is

left over for future research. The input data consists of the measured time series of the indoor and

outdoor climate as presented in Figures 14 and 15. Figure 16 and 17 show the results. Figure 16 shows

that the simulated inside surface temperature is already quite close to the measured one. At this point it

is very important to consider possible errors (due to measurement methodology) of the measured data.

We conclude that Comsol is an efficient tool to simulate 2D transient heat and moisture transport in

building constructions for research. The most difficult part is to calculate the PDE coefficients from

given material properties as shown in equations (6). At this moment the present material library in

Comsol is not capable to calculate these PDEs coefficients automatically. This is a disadvantage for

education. Although not presented in this paper, the extension to 3D heat and moisture transport is

expected be straightforward.
4. Heat, air and moisture transport

4.1 Modeling

The starting point of this Section is the earlier work of van Schijndel [48,49]. The HAM2D

construction model is presented in Figure 18. Where λ is the heat conduction coefficient; d the

thickness; U the U-value; R the heat resistance; ρ the density; Cp the heat capacity; K the air

permeability. Tables III and IV provide detailed material properties of the construction. The internal

and external conditions are provided in Figure 19. The changes compared to the reference HAM2D

model of van Schijndel [48,49] were twofold:

(1) Internal and external boundary conditions (Figure 19) are now based on a typical Dutch

climate instead of the more or less extreme climate of Denver;

(2) In the reference model the wind induced pressure was steady, thus also the airflow through

the construction was steady during the simulation period. However, in this Section, the wind induced

pressure is dynamic (Figure 19, bottom)) and thus also the airflow through the construction.

The Partial Differential Equations (PDEs) are [48]:

∂T
Heat : ρC p + ∇ ⋅ (−λ∇T ) + ρC p u ⋅ ∇T = 0
∂t
∂P
Air : + ∇ ⋅ (− K∇P ) = 0; u = K∇P (9)
∂t
∂p
Moisture : v + ∇ ⋅ (− D∇pv ) + u ⋅ ∇pv = 0
∂t

Where u is air velocity [m/s]; P is (scaled) atmospheric pressure [Pa]; pv is vapor pressure [Pa]; D =

diffusion coefficient [m2/s] equals (pvsat.δa)/(μ.ξ); pvsat is saturation vapor pressure [Pa]; δa is vapor

permeability coefficient (1.8 10-10 s);μ is vapor diffusion resistance factor [-]; ξ is specific moisture

capacity related to RH [kg/m3]; K = permeability [kg m-1 s-1 Pa -1]. In Figure 20 the sub-domains are

presented. Table V provides the PDE coefficients of the sub-domains.

The boundary values are:

Heat : Flux : n ⋅ (λ∇T ) = h(Tinf − T ); Insulation : n ⋅ (λ∇T ) = 0

(10)
Air : Pressure : P = P0 ; Insulation : n ⋅ K∇P = 0

Moisture : Flux : n ⋅ ( D∇pv ) = β ( pv inf − pv ); Insulation: n ⋅ ( D∇pv ) = 0

Where n is normal vector of surface [-]; h is surface coefficient of heat transfer [W/m2K]; β is surface

coefficient of vapor transfer [s/m]. In Figure 21 the boundaries are presented. Table VI provides the

boundary coefficients. Where Ti, Te airsolar, pi, pe and Pwind are time dependent input signals as provided

in Figure 19.

4.2 Simulation results

Three cases are considered, each with different wind induced pressure difference between internal and

external (∆P):

(1) ∆P = 0 (reference case)

(2) ∆P = Pwind (see Figure 19 Bottom) and

(3) ∆P = - Pwind

The simulation results are visualized in two ways:

- Firstly, by movies. This probably the best way to analyze the results.

- Secondly, by time series at six locations: Two points at the inner surface (Figure 22), two points at

the inside wood-insulation surface (Figure 23) and two points at the outside wood-insulation surface

(Figure 24).

The first three sub-figures each show six variants (two positions , each with 3 different wind induced

pressures at the external boundary) of temperature (top), vapor pressure (second) and RH (third). The

sub-figure at the bottom shows the wind induced pressure. The results of indoor surface temperatures

(Figure 5 top) show a quite steady difference of about 1 oC between the continuous lines representing

position a, and the dotted lines representing position b. This is due to the (corner) thermal bridge

effect. Only at high wind induced pressure differences (for example day 28) the airflow due the wind
has some significant effect on the temperatures. This means that thermally, at the internal surface the

multi-dimensional effect is far more dominant than the effect of airflow through the construction. The

results of the surface vapor pressures are very different. The vapor pressures are almost not influenced

by both the 2D effect and the wind speed (four curves: 2a; 2b; 3a & 3b coincide during the whole

simulation period). The RH has roughly the same shape as the vapor pressure. The temperatures of

Figure 23 (top) show the same pattern as Figure 22. However the vapor pressure (second sub-figure) is

influenced by the wind induced pressures (see difference between 1a, 2a & 3a). This clearly shows the

effect of air movement. Again, the RH (third sub-figure) has roughly the same shape as the vapor

pressure. The temperatures profiles of Figure 24 (top) are almost the same (all curves coincide). Thus

there is no effect of airflow. The vapor pressures of Figure 24 (second) have roughly the same shapes

as the previous Figure 23, again showing clearly the effect of air flow. The RH (third sub- figure)

shows the dependency of the temperature as well as the air flow.

We conclude that Comsol is capable of simulating 2D transient heat, air and moisture transport in

building constructions. The main problem is that the simulated air velocity inside solid constructions is

of order µm/s. It is almost impossible to measure this directly and therefore it is very difficult to

validate these types of models. Experimental results are needed to confirm or contradict our

computational findings on the impact of airflow. Furthermore, the moisture transport in this section is

based on vapor transport, so rain penetration and condensation are not included. This means that the

results are not accurate for RH above 90%. A model using the moisture potential LPc instead of vapor

pressure should fix this problem. Although we are quite able to build such a model it would be even

more complicated to validate it.

6. General Conclusions

The paper presents an overview of Multiphysics Applications using Comsol for building physical

constructions simulation. The overview includes three main basic transport phenomena for building

physical constructions: (1) Heat transfer; (2) heat and moisture transfer and (3) heat, air and moisture

transfer. It is concluded that full 3D transient coupled HAM models for building physical
constructions can be build using Comsol. Regarding the heat transport, neither difficulties nor

limitations are expected. Concerning the combined heat and moisture transport the main difficulties

are related with the material properties but this seems to be no limitation. Regarding the HAM

modeling inside solid constructions, there is at least one limitation: The validation is almost

impossible due to limitation of measuring ultra low air velocities of order µm/s.

Currently we are working on the extension of HAM models with induced mechanical stresses and salt

transport. The accompanying paper of Schellen and van Schijndel in this Journal presents how

mechanical stresses can be incorporated into the HAM models. The important issue of including salt

transport is left over for future research.

References

1. M.H. de Wit, Heat and moisture in Building Envelopes, Lecture notes, 200 pages available at

http://archbps1.campus.tue.nl/bpswiki/index.php/Hamlab (2010)

2. C.E. Hagentoft, Introduction to Building Physics, Studentlitteratur AB, Sweden, (2008)

3. H. S. L. C. Hens, Building Physics - Heat, Air and Moisture: Fundamentals and Engineering

Methods with Examples and Exercises, Wiley, Belgium, (2008)

4. W.M.J.A. Hochstebach, Quantitative Modeling of an Air Conditioning Problem, International

Journal of Engineering Education, 21 (4), (2005), pp.750-757

5. N.M.M. Ramos, J.M.P.Q. Delgado and V.P. de Freitas, Modelling and Solving Building Physics

Problems using MATLAB/Simulink, International Journal of Engineering Education, 21 (5),

(2005), pp.784-789

6. A. Katz, Towards ‘Remote’ Learning of Building Materials, International Journal of Engineering

Education, 15 (5), (1999), pp.390-395

7. Chung- Yau Lam and F.H.Alan Koh, A Partial Differential Equation Solver for the Classroom,

International Journal of Engineering Education, 22 (4), (2006), pp. 868-875

8. W.B.J. Zimmerman, Multiphysics Modelling with Finite Element Methods, World Scientific, UK,

(2006)
9. COMSOL: http://en.wikipedia.org/wiki/COMSOL_Multiphysics

10. A.W.M. van Schijndel, Modeling and solving building physics problems with FemLab, Building

and Environment 38, pp 319-327, (2003)

11. A.W.M. van Schijndel, Integrated Heat Air and Moisture Modeling and Simulation, PhD thesis,

Eindhoven University of Technology, Netherlands, (2007)

12. G. Mattarolo, Bard, J. , Theoretical Modeling of a Thermophotovoltaic System Proceedings of the

COMSOL Multiphysics User's Conference 2005 Frankfurt, 7 pages, available at

http://www.comsol.com/ (2005)

13. J. Reiter, Pierer, R. Thermo-mechanical simulation of a laboratory test to determine mechanical

properties of steel near the solidus temperature, Proceedings of the COMSOL Multiphysics

User's Conference 2005 Frankfurt, 7 pages available at http://www.comsol.com/ (2005)

14. I. Bombard, P. Laurent, and J. Lieto, Radiative Heat Transfer in an Infrared Oven, Proceedings of

the COMSOL Users Conference 2006 Paris, 7 pages available at http://www.comsol.com/

(2006)

15. J.V. Pearce, Improved Temperature Standards Using COMSOL Multiphysics, Proceedings of the

COMSOL Users Conference 2006 Birmingham, 6 pages available at http://www.comsol.com/

(2006)

16. V.S.S. Srinivas, and G.K. Ananthasuresh, Analysis and Topology Optimization of Heat Sinks with

a Phase-Change Material using a COMSOL Multiphysics Platform, Proceedings of the

COMSOL Users Conference 2006 Bangalore, 7 pages available at http://www.comsol.com/

(2006)

17. V. Le Sant, and G. Failleau, Thermal Modeling of Devices Realizing Temperature Fixed-Points,

Proceedings of the COMSOL Users Conference 2007 Grenoble, 7 pages available at

http://www.comsol.com/ (2007)

18. G. Jóhannesson, Using COMSOL Multiphysics to Model Ground Heat Exchange, Proceedings of

the COMSOL Users Conference 2007 Grenoble, 8 pages available at http://www.comsol.com/

(2007)
19. W. Joppich, N. Kopp, and D. Samokhvalov, Numerical Behavior of Different COMSOL Solution

Methods for a Heat Transfer Problem Coupled with a Structural Mechanics Problem,

Proceedings of the COMSOL Users Conference 2007 Grenoble, 7 pages available at

http://www.comsol.com/ (2007)

20. T. Zeineldin, Modeling the Process of Drying Stationary Objects inside a Tumble Dryer Using

COMSOL Multiphysics, Proceedings of the COMSOL Conference 2008 Hannover, 6 pages

available at http://www.comsol.com/ (2008)

21. H.L. Schellen, van Aarle, M.A.P., van Schijndel, A.W.M., The use of FEMLAB models to protect

a monumental church organ in case of hot air church heating Proceedings of the COMSOL

Users Conference 2007 Grenoble, 7 pages available at http://www.comsol.com/ (2007)

22. H. L. Schellen, A. van Schijndel, and P. Briggen, The use of COMSOL for Building Constructions

Engineering Regarding Heat and Moisture Transport Proceedings of the COMSOL Conference

2008 Hannover, 6 pages available at http://www.comsol.com/ (2008)

23. U. Krause, Schmidt, M., Lohrer, C., Computations on the coupled heat and mass transfer during

fires in bulk materials, coal deposits and waste dumps Proceedings of the COMSOL

Multiphysics User's Conference 2005 Frankfurt, 7 pages available at http://www.comsol.com/

(2005)

24. M.E. Cakmak, Y. Zevi, V.L. Morales, B. Gao, J.L. Nieber, and T.S. Steenhuis, Pore Scale

Modeling of Colloid Transport in Unsaturated Porous Media with COMSOL from the

Proceedings of the COMSOL Users Conference 2006 Boston, 4 pages available at

http://www.comsol.com/ (2006)

25. M.V. De Bonis1,, G. Altieri, and G. Ruocco, Combined Transfer Phenomena in 3D Modeling of

Packaged Foods Proceedings of the COMSOL Users Conference 2006 Milano, 6 pages

available at http://www.comsol.com/ (2006)

26. A. W. M. van Schijndel, The Effect of Air Movement on the Heat and Moisture Characteristics of

Building Constructions, Proceedings of the COMSOL Users Conference 2007 Grenoble, 7

pages available at http://www.comsol.com/ (2007)

27. T. Fong, M. Chui, and D. L. Freyberg, The Use of COMSOL for Integrated Hydrological

Modeling Proceedings of the COMSOL Conference 2007, Boston, 7 pages available at

http://www.comsol.com/ (2007)

28. A.W.M. van Schijndel, Heat and Moisture Modeling Benchmarks using COMSOL Proceedings of

the COMSOL Conference 2008 Hannover, 5 pages available at http://www.comsol.com/ (2008)

29. M. de Bonis, and G. Ruocco, A Multiphysics Approach to Fundamental Conjugate Drying by

Forced Convection Proceedings of the COMSOL Conference 2008 Hannover, 7 pages available

at http://www.comsol.com/ (2008)

30. G. Cammarata, and G. Petrone, A Numerical Investigation on Active Chilled Beams for Indoor Air

Conditioning Proceedings of the COMSOL Conference 2008 Hannover, 7 pages available at

http://www.comsol.com/ (2008)

31. Barletta, A., Lazzari, S., 2D Free Convection in a Porous Cavity Heated by an Internal Circular

Boundary Proceedings of the COMSOL Multiphysics User's Conference 2005 Stockholm, 6

pages available at http://www.comsol.com/ (2005)

32. Lee, Y., Finite Element Simulation of the Oscillatory Flow in a Channel with the Heat Transfer

through a Hot Bump, Proceedings of the COMSOL Multiphysics User's Conference 2005

Boston, 6 pages available at http://www.comsol.com/ (2005)

33. B.A. Finlayson, Using COMSOL Multiphysics to Model Viscoelastic Fluid Flow Proceedings of

the COMSOL Users Conference 2006 Boston, 9 pages available at http://www.comsol.com/

(2006)

34. S. A. Osman, and M. T. Ismail, Wind Loads Prediction Using Three-Dimensional Simulation of k-

ε Turbulence Model, Proceedings of the COMSOL Users Conference 2007 Kuala Lumpur, 5

pages available at http://www.comsol.com/ (2007)

35. C. Balocco, Investigation of Natural Ventilation Techniques in Historical Buildings, a Case Study

in Palermo, Proceedings of the COMSOL Users Conference 2007 Grenoble, 5 pages available at

http://www.comsol.com/ (2007)
36. J. Frassy, C. Lecot, M. Murariu, C. Delattre, and A. Soucemarianadin, Modeling of Drops

Spreading on Random Surfaces, Proceedings of the COMSOL Users Conference 2007

Grenoble, 5 pages available at http://www.comsol.com/ (2007)

37. C. Balocco, Analysis of Ancient Natural Ventilation Systems inside the Pitti Palace in Florence

Proceedings of the COMSOL Conference 2008 Hannover, 7 pages available at

http://www.comsol.com/ (2008)

38. T. van Goch, and A. van Schijndel, Validation of DNS Techniques for Dynamic Combined Indoor

Air and Constructions Simulations Using an Experimental Scale Model Proceedings of the

COMSOL Conference 2008 Hannover, 8 pages available at http://www.comsol.com/ (2008)

39. M. Diaz-Viera, D. Lopez-Falcon, A. Moctezuma-Berthier, and A. Ortiz-Tapia, COMSOL

Implementation of a Multiphase Fluid Flow Model in Porous Media Proceedings of the

COMSOL Conference 2008 Boston, 7 pages available at http://www.comsol.com/ (2008)

40 . R. Ababou, Cañamón, I., Elorza, J., Thermo-Hydro-Mechanical Simulation of a 3D Fractured

Porous Rock Preliminary Study of Coupled Matrix-fracture Hydraulics, Proceedings of the

COMSOL Multiphysics User's Conference 2005 Paris, 6 pages available at

http://www.comsol.com/ (2005)

41. P. Kucera, Thermal-mechanical Analysis of Concrete Structure Exposed to High Temperature

Proceedings of the COMSOL Users Conference 2007 Grenoble, 5 pages available at

http://www.comsol.com/ (2007)

42. W. Joppich, N. Kopp, and D. Samokhvalov, Numerical Behavior of Different COMSOL Solution

Methods for a Heat Transfer Problem Coupled with a Structural Mechanics Problem,

Proceedings of the COMSOL Users Conference 2007 Grenoble, 7 pages available at

http://www.comsol.com/ (2007)

43. A. Vermeltfoort, and A.W.M. van Schijndel, Modelling of Lintel-Masonry Interaction Using

COMSOL, Proceedings of the COMSOL Conference 2008 Hannover, 9 pages available at

http://www.comsol.com/ (2008)
44. H. L. Schellen, A.W.M. van Schijndel, The use of COMSOL to solve hygro thermal building

physical problems related to insulating high rise building facades. Proceedings of the COMSOL

Users Conference 2007 Grenoble, 7 pages available at http://www.comsol.com/ (2007)

45. HAMLab, Heat Air and Moisture Laboratory:

http://archbps1.campus.tue.nl/bpswiki/index.php/Hamlab (2010)

46. P.M. Briggen, Onderzoek naar de oorzaken van de vochtschade aan de toren van Jachthuis St.

Hubertus. Master Thesis, Unit Building Physics and Systems, Eindhoven University of

Technology, the Netherlands, [in Dutch] (2007).

47. P.M. Briggen, Blocken, B.J.E., Schellen, H.L., Wind-driven rain on the facade of a monumental

tower: numerical simulation, full-scale validation and sensitivity analysis. Building and

Environment, 44(8), pp1675-1690, (2009).

48. A.W.M. van Schijndel, Heat, Air and Moisture Construction modeling using COMSOL with

MatLab, Modeling guide version 1.0, Proceedings of the COMSOL Users Conference 2006

Eindhoven (2006)

49. A.W.M. van Schijndel, HAM Construction modeling using COMSOL with MatLab, Modeling

guide version 1.0, IEA Annex 41 resesarch paper no: A41-T1-NL-06-6 (2006)
Figures

external b1

concrete b2

insulation b3
b4
b7

b5 internal

b6

Figure 1. The geometry of the 2D thermal bridge example, see Table I for boundary conditions

specifications.

%CONSTANTS
hi=7.7; %heat transfer coefficient internal
he=25; %heat transfer coefficient external
Ti=20; %air temperature internal
Te=-10; %air temperature external
KCONCR=1; %heat conductivity concrete
KINSUL=0.03; %heat conductivity insulation

%GEOMETRY: poly2(XDATA,YDATA) ; 2D polygon

CONCR=poly2([0 0 1 1 0.2 0.2],[0 1 1 0.8 0.8 0]); %concrete
INSUL=poly2([0.2 0.2 1 1],[0.7 0.8 0.8 0.7]); %insulation
fem.geom=CONCR+INSUL; %fem geometry
fem.dim=1;%fem geometry

%COEFFICIENTS OF THE PDE/Boundary problem

fem.equ.c={KCONCR KINSUL}; % fem coefficient c
fem.bnd.g={0 0 he*Te hi*Ti 0 hi*Ti 0 0 0}; % fem coefficient g
fem.bnd.q={0 0 he*Te hi 0 hi 0 0 0}; % fem coefficient q
fem.mesh=meshinit(fem); % initialize mesh
fem.xmesh=meshextend(fem); % extended mesh
fem.sol=femlin(fem); % solve, steady problem
meshplot(fem) % solve, steady problem
postplot(fem,'tridata','u','tribar','on') % plot solution

Figure 2. The complete Comsol code for solving the 2D thermal bridge problem.
Figure 3. The mesh of the 2D thermal bridge problem.

Figure 4. The solution (temperature distribution) of the 2D thermal bridge problem.

Figure 5. The design of the construction.

Figure 6. Exemplary result of the temperature distribution

Figure 7. The conversion from measurable material properties into PDE coefficients
Figure 8. PDE coefficients CT,CLPc,Kij as functions of LPc and T calculated from the provided

HAMSTAD benchmark no.1 material properties for the load bearing (a) and insulation (b) material

(please note that this figure relies on colour, see digital version)
Figure 9. A schematic of the structure of the benchmark

Figure 10 The simulated total mass in the load bearing material (top) and insulation material (bottom)

during the first year by Comsol (left) and the reference from the benchmark (right)
Figure 11 The simulated heat flows interior (bottom) and exterior (top) to the wall during the first year

by Comsol (left) and the reference from the benchmark (right)

Figure 12. The building façade

Figure 13. Measurement of the surface temperature and relative humidity using a box (see

encirclement)
Figure 14. The measured air temperatures (top) and calculated vapour pressures (bottom, from

measured T/RH)

Figure 15. The measured solar irradiance (top) and rain intensity (bottom)
Figure 16. The measured and simulated inside surface temperature.

Figure 17. The measured and simulated relative humidity at the surface.
Figure 18. The 2D construction.

Figure 19. Top: Indoor and outside air temperatures; Middle: Indoor and outside air vapor pressures;
Bottom: The wind induced pressure; Horizontal All: number of days after January 1
Figure 20. Sub-domains & numbering

Figure 21. Boundary values & numbering

Figure 22. Time series at inner surface (see Figure 18 for locations of pos a (continuous lines) & b

(dotted lines) )
Figure 23. Time series at inside wood-insulation surface (see Figure 18 for locations of pos a

(continuous lines) & b (dotted lines) )

Figure 24. Time series at outside wood-insulation surface (see Figure 18 for locations of pos a

(continuous lines) & b (dotted lines) )

Table I. Boundary specifications of the 2D thermal bridge problem (Figure 1), where T is the

temperature on the boundary and Ti and Te respectively are the internal and external temperatures.

Boundary Boundary Boundary Boundary condition

Segment Type Segment length [m] [Wm-2]

b1 external 1.0 φ=hce*(Te-T)

b2 adiabatic 0.2 φ=0

b3 adiabatic 0.1 φ=0

b4 internal 0.8 φ=hci*(Ti-T)

b5 internal 0.7 φ=hci*(Ti-T)

b6 adiabatic 0.2 φ=0

b7 adiabatic 1.0 φ=0

Table II. Material properties of the thermal bridge problem (Figure 5)

Material thermal conductivity density specific heat

k [W/m.K] ρ [kg/m3] c [J/kg.K]

Concrete 1.6 2500 840

Mineral wool 0.037 50 840

Gypsum board 0.2 850 850

Table III. Material properties part 1

λ d U
W/mK m W/m2K
Exterior wall (inside to outside)
Int. surf. coeff. 8.29
Wood panels 0.160 0.012 13.333
Cellulose ins. 0.040 0.066 0.606
Wood siding 0.140 0.009 15.556
Ext. surf. coeff 29.300
Total air-air 0.514
Roof (inside to outside)
Int. surf coeff 8.29
Wood panels 0.160 0.010 16.000
Cellulose ins. 0.040 0.1118 0.358
Roof deck 0.140 0.019 7.368
Ext. surf. coeff 29.300
Total air-air 0.318

Table IV. Material properties part 2

R ρ Cp K
m2K/W kg/m3 J/kgK kg/msPa

Exterior wall (inside to outside)

Int. surf. coeff. 0.121
Wood panels 0.075 395 1880 10-9
Cellulose ins. 1.650 55.0 1880 5.510-5
Wood siding 0.064 530 900 10-9
Ext. surf. coeff 0.034
Total air-air 1.944
Roof (inside to outside)
Int. surf coeff 0.121
Wood panels 0.063 395 1880 10-9
Cellulose ins. 2.794 55.0 1880 5.51
0-5
Roof deck 0.136 530 1880 10-9
Ext. surf. coeff 0.034
Total air-air 3.147
 Table. V: Sub-domain PDE coefficients

Sub-domain 1-4 5-7 8-9

Λ 0.14 0.04 0.16

Ρ 530 55 395

Cp 1880 1880 1880

D = (pvsat.δa)/(μ.ξ) psatf(T)*1.8*10-10/(120*95) psatf(T)*1.8*10-10/(14*1.4) psatf(T)*1.8*10-10/(101*2.1)

K 10-9 5.5*10-5 10-9

Table VI. Sub-domain boundary value coefficients

Boundary 1, 5, 7, 9 2-3, 11, 18, 24-26 22-23

Type Flux Insulation Flux

Heat (h) 29.3 0 8.29

Heat (Tinf) Te airsolar - Ti

Moisture Inward flux 8e-8*(pe-pv) 0 2e-8*(pi-pv)

Air Pressure (P0) Pwind 0 0

View publication stats