Page 1 of 1

simulation with error!

Posted: Mon Aug 04, 2014 9:06 pm -1100
by WUFImem
Hello Wufi Support

I have tried to investigate a simulation with internal insulation. The construction has an internal insulation of vacuum insolation panels (VIP), EPS and a vapor control layer. The wall material hard limestone is used. The result has some convergence Failed (412 times), iteration arise Interrupted (1946 times), ATSC (0 times), Rain iteration (0 times). How much influence have these failures on the result (5 years - 43800 steps)? How can I reduce the failures and interruptions?

The second design uses the same construction and the same boundary conditions. The only exception is the used wall material. Instead of hard limestone, bricks are used. When using the same element grid X: 150 Y: 150, makes the simulation after 6 months unexplained moisture rises. In the analysis, the values of water content increase abruptly (10 times) and stay until the end of the simulation (5 years) at the same very high level. (a line)

To avoid errors, the design was re-entered, the simulation grid was changed, the wall material was changed into another brick material, etc.. Unfortunately, without success!

The only way to get a result was to reduce the simulation grid at X: 50 and Y:50. The result appears to be very questionable. The result has also some convergence Failed (412 times), iteration arise Interrupted (1296 times), ATSC (0 times), Rain iteration (0 times). Why it is not possible to simulate the same construction with a other wall material?

Best regards

Re: simulation with error!

Posted: Mon Aug 04, 2014 9:52 pm -1100
by Christian Bludau
The convergence failures as well as the interruption do not necessarily have an influence on the result. Please check your courses. If there are jumps, you can not explain, they may be caused by convergence failures.
In your second case your calculation crashed because of some convergence problems. For further information see here:
http://www.wufi-forum.com/viewtopic.php?f=15&t=494
http://www.wufi-forum.com/viewtopic.php?f=15&t=789
http://www.wufi-forum.com/viewtopic.php?f=15&t=439
http://www.wufi-forum.com/viewtopic.php?f=15&t=768
To reduce the number of grid elements usually is no good way to get the calculation to run as the accuracy of the result will decrease in most cases.

Please also go to Computational Parameters -> Enhanced and see the table "solver settings". If not yet done, try to change the value URF - RH from 0.7 to 0.3. That handles the influence of the last result to the next step and often leads to a better convergence behavior.

Christian

Re: simulation with error!

Posted: Tue Aug 05, 2014 9:55 pm -1100
by WUFImem
Have reduced the RH from 0.7 to 0.3.
The result has less problemes - iteration interrupted (269 times and only E-0 for theta, phi and rain), and Rain iteration (280 times and only E-0 for theta, phi and rain).

The attached diagram shows still a sudden rising of the water content from 8 to 20 kg / m³. This is not realistic. What is the cause?

Re: simulation with error!

Posted: Tue Aug 05, 2014 10:09 pm -1100
by Christian Bludau
Is this a sucking surface? I would guess that there is strong rain event at the time of the jump. I do not think it is caused by convergence failures, as it repeats every year at the same time.

Re: simulation with error!

Posted: Tue Aug 05, 2014 10:47 pm -1100
by WUFImem
The wall material is old brick from the IBP database. You can see it on the attachment. I wouldn`t say it is a sucking surface!

Re: simulation with error!

Posted: Tue Aug 05, 2014 10:58 pm -1100
by Christian Bludau
The Vollziegel has an a-value of about 7 kg/m²h^0,5. Thats quite a lot. Modern plasters are in the range lower than 0,1 kg/m²h^0,5.
See the material properties. If there is a transport coefficient defined for suction, then the material can suck in water from rain.