User Forum of Software BASEMENT

davidz

User

Registered: 2016-02-01

Posts: 16

Problem with start of the model

Hi

I want to set up a 1D model, but there ist something wrong with my setup.
I always get the following Error. I allready tried the two options which are recomended by basement. Restarting with wet conditions is not possible because i can't get a restartfile with wetted conditions.
And in SECTION_COMPUTATION i tried several options and combinations.
At the moment I use in SECTION_COMPUTATION:

SECTION_COMPUTATION {
    type      = iteration
    precision = 0.01
}

ERROR -> Newton-iteration to calculate water surface elevation (z) as a function
   of the wetted area (A) failed for cross section '1'. Maximal number of
   iteration (1000) was reached.
   Requested A = 5.891e+018
   Current z   = 0
   A-Anew      = 1000
   [z0,z1]     = [443.97, 456.31]
   [A0,A1]     = [0, 108.262]
   Please try to (1) start simulation with wetted initial condition, (2) check
   tags in HYDRAULICS -> PARAMETERS -> SECTION_COMPUTATION.

I think there is something with my geometry file.. When use an simple geometry (regularly trapezoidal crosssections) then this problem does not happen.
Is it possible that small distances between crossections or even between the crosssection-Points can create instabilities?

Thanks for your help.

David

sajjad

User

Registered: 2014-10-01

Posts: 102

Re: Problem with start of the model

from my experience, some x-sections may not be correct which causes non-convergence. Check your sections for abrupt slope changes laterally. Maybe, the d/s boundary cond. is not correct (if u have used hqrelation as bc). Try 'zero-gradient' to see whether u get the same error or not.

Regards

René Kaufmann

User

Registered: 2016-01-28

Posts: 22

Re: Problem with start of the model

Did you try the type "TABLE" in Section_Computation?

Try with min_intervall = 0.01 and max_intervall = 0.05

davidz

User

Registered: 2016-02-01

Posts: 16

Re: Problem with start of the model

thanks.. now it works.

I had to delete one "Strange" crosssection with an island in the middle of the river.

And i changed from itteration to table.

davidz

User

Registered: 2016-02-01

Posts: 16

Re: Problem with start of the model

I got the next problem with my model... It's still the same model... I can start it now under wetted conditions. for just hydraulic conditions it runs fast and without any problems.

But as soon i enter a gate as an inner boundary to model a bridge i get the same error as above..

@sajjad: What do you mean with "abrupt slope changes laterally". I have a small river with a trapezoidal crosssection but sometimes there is also a wall and a few coulverts below bridges...

I tried to delete a few crossections so that the distance between the crossections gets more regular.

Thanks for your help.

David

ingridsenn

User

Registered: 2016-11-28

Posts: 10

Re: Problem with start of the model

Hi David

I am not an expert myself, but two thoughts on your problem(s):
a) For the newton iteration error: now that you added an inner boundary, you may need to re-adjust the values min_interval, max_interval as suggested by René or be less restrictive in the precision parameter within your block Parameter (if you can, according to your project goals).
b) regarding the cross sections, I would rather not delete any cross sections. Your simulation is fast (if it works), so you can afford to have many. I think what Sajjad means is that it may cause a problem if the geometry changes abruptly. Random example: Cross section CS3 is 10 m wide and the embankment is sloping 3:2, cross section CS4 is 5 m wide with vertical walls. If you have such cases, I would try to have more cross sections instead of less. Make a reasonable assumption and construct some in-between profiles or try the experimental interpolation in Basement's Cross Section editor tools.

I hope you can manage to resolve your problem.

Ingrid

René Kaufmann

User

Registered: 2016-01-28

Posts: 22

Re: Problem with start of the model

It seems, there is a problem with the gate inner boundary. Please refer to the post http://people.ee.ethz.ch/~basement/foru … php?id=478

davidz

User

Registered: 2016-02-01

Posts: 16

Re: Problem with start of the model

René solved my problem... the model runs in Basment 2.6. there must be a problem with the implementation of the Gate in version 2.7.
Thanks.