Modélisation WUFI2D - Incohérences

Post Reply
jpair68
WUFI User
WUFI User
Posts: 7
Joined: Thu Apr 24, 2014 2:07 am -1100

Modélisation WUFI2D - Incohérences

Post by jpair68 » Tue Jun 17, 2014 2:50 am -1100

Bonjour à tous,

Je suis nouveau sur le forum (et sur Wufi !). Je souhaite modéliser à l'aide de WUFI 2D la migration de vapeur d'eau d’une paroi d’un bâtiment situé dans le Haut-Rhin, dans le sud de l’Alsace. C’est une paroi constituée de 70 cm de grès et d'une couche de plâtre traditionnel, avec une poutre en bois s'insérant dans le mur, et d’un enduit extérieur clair. Le bâtiment n'a pas de pathologie lié à l'humidité.
Concernant la modélisation, j'ai réduit le maillage (50*50) afin de réduire les erreurs de convergence, renseigné l’absorption à 0,4 et l’émission à 0,9. J’ai également pris le fichier climat de Fribourg et fait la simulation sur 3 ans, en commençant début octobre.
Pour les conditions initiales, un autre détail me pose problème. On peut renseigner une valeur d'humidité relative, mais laquelle ? Sur une longue période (5 ans), ça ne devrait pas poser de problème, la paroi devrait retrouver un équilibre. Pourtant, j'ai testé une humidité relative de 0,3 et 0,8, les résultats ne sont pas du tout les mêmes : bien qu'il devrait y avoir un équilibre, la teneur en eau est nettement supérieure pour une humidité relative initiale plus importante.
Le bâtiment est actuellement sain, mais les résultats de Wufi me donnent une teneur en eau qui augmente au fil des années dans la paroi… Pourriez-vous m’éclairer sur ce point ? Qu’est-ce qui vous semble incohérent dans ce que j’ai renseigné ?

Image


J'ai également d'autres questions concernant l'utilisation de WUFI.
Comment se fait-il que, malgré des erreurs de convergence, le logiciel arrive tout de même à sortir des résultats ? Sont-ils exploitables ?
Pourquoi le choix du maillage est important ? En le réduisant, j'obtiens des résultats bons vis-à-vis des erreurs de convergence. Mais moins précis non ?
Peut-on se permettre d’ignorer les chaleurs latentes et de fusion lors des simulations ?
Une dernière chose : je suis obligé de relancer le calcul pour pouvoir utiliser WufiGraph. Vous aussi ?

Je vous remercie de m'avoir lu,

Arthur

Daniel
WUFI SupportTeam IBP
WUFI SupportTeam IBP
Posts: 132
Joined: Tue Jul 19, 2005 2:50 am -1100
Location: Fraunhofer IBP, Holzkirchen
Contact:

Re: Modélisation WUFI2D - Incohérences

Post by Daniel » Sun Jun 22, 2014 11:25 pm -1100

Dear Arthur,

for existing buildings I'd propose to start with an Initial Content around the average outdoor RH. In Europe somewhat between 75 and 85 %. This should be a good Approach. But in case of extensive rain water Absorption or known higher water Contents - higher values could / should be used. Aas you already mentioned before: it doesn’t influence the long term moisture level but normally accelerates to reach it.

For older masonry walls rain water absorption is an important influencing factor – so if the water contents are higher than expected in your case you maybe have to check driving rain offer and water absorption of the surface…

If there is no convergence, the last result is used anyway. So you have to decide in fact on your own if result is accurate enough. This can be done by the help of the result analzis – here you get information about the last residua…

Concerning the grid resolution – normally you have to refine it step by step until you don’t get any more differences for the results compared to the previous solution. By experience you will get a feeling what is appropriate (depends also on structure, materials, loads (with or without rain), gradients etc.).

Latent heat should not be ignored – this possibility only allows for example to quantify the influence of such effects…

Result evaluation is theoretically possible while running the simulation – but in practice i’ts too slow. So we do the same – pause the simulation and restart it afterwards.

Best regards
Daniel
Dr.-Ing. Daniel Zirkelbach, Deputy Head of Department Hygrothermics, IBP Holzkirchen

jpair68
WUFI User
WUFI User
Posts: 7
Joined: Thu Apr 24, 2014 2:07 am -1100

Re: Modélisation WUFI2D - Incohérences

Post by jpair68 » Thu Jul 17, 2014 1:04 am -1100

Dear Daniel,

Thank you very much for your answer and your explanations. But I have some questions to ask again !
Concerning the grid resolution – normally you have to refine it step by step until you don’t get any more differences for the results compared to the previous solution
Do I have to start with a coarse grid and then, reach an accurate grid? Or do I have to do the opposite? Which calculation is right if the two are different?
Moreover, where can I find the driving rain offer and water absorption of the surface? I can't find it…
How many convergence failures are acceptable? And iteration interrupted? Is the calculation right if I don't have convergence failure but some iteration interrupted?
I made the same simulation as the first one, by adding 12 cm PU insulation. I was expecting to see the water content increases in the boundaries between rock and insulation, but no. The PU is a non-hygroscopic material, so, theoretically the water should be stopped by the insulation layer. The water content even decreases at the boundaries. Do you have an idea of what it happened?

Thank you.
Best regards,

Arthur

Daniel
WUFI SupportTeam IBP
WUFI SupportTeam IBP
Posts: 132
Joined: Tue Jul 19, 2005 2:50 am -1100
Location: Fraunhofer IBP, Holzkirchen
Contact:

Re: Modélisation WUFI2D - Incohérences

Post by Daniel » Thu Jul 17, 2014 5:58 am -1100

Dear Arthur,

in practice you'll try to use an adequate grid already in the first approach. However, as this is not always sure so you can check by going from coarse to fine - if there are no differences you can leave the coarser versoin as it can be solved faster.

You can Change the rain water load under surface climate - where you also choose your Location. Here you can adapt the value R2 (wind depending load). Normal values are 0,07 for a single Family house, 0,05 for the lower part of a high rise building, 0,1 for the middle and 0,2 for the hightes part of the same buidling. However - the values are for not sheltered buildings on the free field. In the city the values can be reduced or in some cases (at the end of a street for example) also increased (please check also the online help).

There is no clear limit concerning number of convergence errors / you have to check if the result accuracy is OK für your specific structure.

Daniel
Dr.-Ing. Daniel Zirkelbach, Deputy Head of Department Hygrothermics, IBP Holzkirchen

jpair68
WUFI User
WUFI User
Posts: 7
Joined: Thu Apr 24, 2014 2:07 am -1100

Re: Modélisation WUFI2D - Incohérences

Post by jpair68 » Thu Jul 17, 2014 8:36 pm -1100

Thank you very much. Your answer is so helpful !

Arthur

Daniel
WUFI SupportTeam IBP
WUFI SupportTeam IBP
Posts: 132
Joined: Tue Jul 19, 2005 2:50 am -1100
Location: Fraunhofer IBP, Holzkirchen
Contact:

Re: Modélisation WUFI2D - Incohérences

Post by Daniel » Sun Jul 20, 2014 8:51 pm -1100

Dear Arthur,

please also check the follwoing topics in the forum which may help:

http://www.wufi-forum.com/viewtopic.php?f=15&t=439
http://www.wufi-forum.com/viewtopic.php?f=15&t=768

best regards
Daniel
Dr.-Ing. Daniel Zirkelbach, Deputy Head of Department Hygrothermics, IBP Holzkirchen

Post Reply