| Author |
 Topic  |
|
|
Nam
Junior Member
Belgium
4 Posts |
 Posted - 02/18/2010 : 17:18:57
|
Hi everyone,
I am trying to simulate with Hydrus1D a water flux through a multi-layer cap (5 different materials) which has the following properties :
Qr / Qs / Alpha(1/m)/ n / Ks (m/s) / width (m)
0.055 /0.48 /1.6 / 1.57 / 5.8E-6 / 1
0.03 /0.33 /7.44 / 2.96 / 3.33E-4 / 0.3
0.005 /0.42 /493 / 2.19 / 3.5E-3 / 2.65
0.151 /0.38 /0.8 / 1.33 / 1.8E-7 / 0.3
0 /0.44 /0.0008 /2 / 10E-10 / 2
The boundary conditions are :
upper BC : constant net rainfall constant flux (270mm/year)
bottom BC : free drainage
All I need is the bottom flux every year (step time) so I put an obervation node at the bottom.
What do I need to put for the pressure head ? Can it be constant for each node ? (knowing that this is a simulation on earth surface and so at atmospheric pression)
I supposed that the pressure head was constant but the problem is that the solution does not converge. Could you tell me what is wrong ? |
|
|
Jirka
Moderator
USA
1788 Posts |
Posted - 02/18/2010 : 18:04:30
|
If you use constant flux BC, then for precipatition, you need to use negative number. Fluxes upward are positive.
J. |
 |
|
|
Nam
Junior Member
Belgium
4 Posts |
Posted - 02/21/2010 : 18:38:50
|
Oh I see, thank you.
What about the pressure head ? Can I suppose it is constant everywhere ? |
|
|
|
Jirka
Moderator
USA
1788 Posts |
Posted - 02/21/2010 : 19:36:09
|
| sure. See our examples, e.g., Test1. J. |
|
|
|
Nam
Junior Member
Belgium
4 Posts |
Posted - 02/22/2010 : 18:56:22
|
ok, thx !
Now I got the bottom flux and I want to use it for the second part of my problem, this time with solute transport. (I simulate 300 years)
solute half life : 30 years.
Solute initial concentration in liquid phase (between -1.1m and -9.2m, considering that the top is 0m) : 3.7E-10 mole/l
everywhere else, the initial concentration is 0
Solute atomic mass : 137
solute diffusion coefficient : 1.54E-12 m^2/s
The bottom water flux for the first part was approximatively constant and equal to -10^-10 m/s so the upper BC here is a constant flux of -10^-10 m/s.
Bottom BC : free drainage.
This flux go through 3 layers with the following properties :
Qr / Qs / Alpha(1/m)/ n / Ks (m/s) / width (m) / Longitudinal dispersity (m)
0.03 /0.33 /7.44 / 2.96 / 3.33E-4 / 0.4 / 0.01 (top layer)
0.1 /0.23 /6.8E-4 / 1.39 / 2.7E-10 / 9.8 / 0.001 (middle layer)
...same material as top layer ... but width = 3 m (bottom layer)
Transverse dispersity for both material is 0.001 m
The solid density is 2650 kg/m^3 for both materials.
I consider that there is no more solute when we reach 10 times the half life (there would be left only 1/1024 of the initial solute). Knowing that, is it correct if I suppose that the pulse duration is equal to 10 times the half life ?
In "solute transport parameters", Difus. W. means ciffusion coefficient in water(in liquid state) and difus. G. in gas state, right ? If yes, can I use the same value for both ? or should I consider difus. G. equals 0 ? or I have to find the exact value ? Disp means. Dispersity right ? But is it longitudinal or transverse ?
And what means Frac and Thlm ?
SinkWater1, SinkSolid1 and SinkGas1 mean exponential decay constant in liquid, solid and gas state of water right ? But why are there also SinkWater0, SinkSolid0, SinkGas0 and SinkWater1', SinkSolid1' and SinkGas1' ?
I tried with the same value for all the Sink**** but the solution could not converge. Can you see what the problem is ?
By the way, thanks a LOT for helping me. |
|
|
|
Nam
Junior Member
Belgium
4 Posts |
Posted - 02/23/2010 : 16:44:57
|
I have some other questions.
For the first part (with the multi-layer cap) :
I have something very weird : When I change the maximum number of iteration (in Water flow - iteration criteria), I obtain different results in the file RUN_INF.OUT.
For example :
With maximum number of iteration (MNOI) = 100, I get convergency=F for TLevel 1 and 2 and T for every other TLevel.
Whith MNOI = 200, I get convergency=F for TLevel 1 and T for every other TLevel.
With MNOI = 199, I get convergency=F for TLevel 1 and 2 and T for every other TLevel.
But for MNOI = 198, I get convergency=F for every TLevel. ( It does not converge at all)
Do anyone see where the problem comes from ? (I use Water Content Tolerance and Pressure Head Tolerance equal 0.1)
Still for the first part :
When I use only 101 nodes, the solution converge for certain MNOI but when I use 1001 nodes, it never converge, no matter what I put for MNOI... How is that possible ? (Every other input datas are the same)
edit : problem solved, end of topic. |
Edited by - Nam on 03/01/2010 16:45:32 |
|
|
| |
Topic |
|
|
|
Convergence problem
