WUFI

First, to lay out the setup:

- I have a model with 260*280 nodes
- My PC is a 1,66GHz Intel Core (2) processor, with 2 GB ram

When i had not set the thermal conductivity and short/long wave values to the boundary conditions, a calculation took about 10 hours. But since i added these boundary conditions, WUFI says that it will take med about 7-8 days! If i disable "increased accuracy" and "Adaptive convergence", the calculation takes about 24 hours.

Is this the general scope of calculation time, or have i done something wrong?

Thank you, Jeppe

The "adapted convergence" check mark changes the number of maximum iteration steps from 500 unchecked to 1000 (if checked gray) and 1500 check amrked. This is the number of steps after which the iteration is stopped. If the iteration cannot finish (because it is stopped after for example 500 steps) the result may be inaccurate.

The "increased accuracy" hook changes the convergence criteria from 5e-3 unchecked to 5e-4 gray checked and 5e-5 checked. This increases the accuracy of the difference between two results of the iteration. The lower the difference the "better" is the result.

These two changes cause a high change in the simulation time. The solution of the differential equations has to be done for every row and every column. The next factor of time is the number of elements. If you decrease the number, the calculation will be faster, but may be more inaccuracy. The aim is to find a good middle way.
You can try to reduce the number of elements by using a half symmetric system with an adiabatic system border.
But while reducing the elements beware, that to neighbored elements should not exceed a size factor of 2 (The next element is not more then double size of the considered element).

Another way of making the calculation faster is to use bigger time steps. At the moment you are calculating with hourly values. If you use 6 hour values for example, you reduce the number of time steps by the factor 6. The problem is, this only works good with the temperature, the moisture should be considered with hourly values.

Also boundary conditions like rain and radiation will rise the calculation time, because they have to be recalculated for every time step additionally.

So the increase of the calculation time is normal. Just start the calculation and lean back for a view days

Greetings, Christian

Hi Christian.

Thank you for your quick answer!

An additional question: My simulations terminates after long time, with the comment that: "inner iterations diverge". Hwo do i prevent the iterations form diverging? is the somehow linked to the vapour retarders that i have in my geometry?

Thanks, JSH