| Author |
 Topic  |
|
|
EB@ars
Junior Member
USA
2 Posts |
 Posted - 07/27/2010 : 17:41:03
|
I am simulating some infiltration scenarios in furrows. I have been able to successfully set up some simulations but am having difficulties with others. The cases I am testing are all alike. THe domain is 2D, 100 X 100 cm, consisting of a single material. I define the furrow shape using a simple trapezoid (0.2 bottom width, side slope = 2). I am only simulating water flow, using the van Genuchten model, and with no hysteresis. The upper boundary condition is variable head, the bottom is free drainage, and the sides are no flux. The variable head is actually constant - 10 cm water depth measured from the furrow bottom. I ran this case successfully first with a constant initial water content (0.15), and then with a water content varying linearly with depth (0.1 at the surface, 0.15 at the bottom). I then tried to repeat this simulation with water content decreasing linearly in the range 0.1 to 0.15 the from top to the bottom of the furrow, and then with a constant water content from the furrow bottom to the free draining boundary. At this point, the simulation has failed to calculate beyond 30 min, when the inputs are provided for a 360 min simulation. The program doesn't crash, it just stops at 30 min. I have tried several changes to the inputs (boundary and initial conditions, calculation times), but still the scenario will only simulate for 30 min. The only way in which I can get the simulation to work is to reenter all of the data in a new file.
I have also tried to simulate an infiltration scenario with redistribution, with water ponding during the first six hours, and zero water depth after that. The total simulation time is 3 d. I have been able to successfully complete the simulation with the furrow geometry domain described above. When I used instead a rectangular domain (1D infiltration), I cannot get the program to apply the specified time-variable upper boundary condition. Instead, the program appears to apply a constant value at that boundary.
Is there any way to overcome these program quirks. Thanks.
|
|
|
Jirka
Moderator
USA
1788 Posts |
Posted - 07/27/2010 : 18:08:44
|
When the simulations stops at exactly 30 min, then there is obviously a reason for that. Two reasons I can think of are: A) The final print time that you request is at 30 minutes. The code does not go beyond the final requested print time (there is no reason for that). B) Your last time-variable boundary condition is specified for 30 minutes (the code would not go beyond that either, having undefined BC).
When you have time variable BC with specified heads, then both positive and negative values are imposed at the boundary (as they should). Zero is not interpreted as no water in the furrow. For that you would need to use the special "BDRC Options" from the Tools bar. In 1D you sould need to use instead a "Variable Pressure Head/Flux" BC and specify KodTop=1 for head BC and KodTop=-1 for flux BC.
Jirka
|
 |
|
|
EB@ars
Junior Member
USA
2 Posts |
Posted - 07/28/2010 : 21:51:19
|
With respect to the simulation that stops at 30 min, the simulation time and boundary conditions are both set for 360 min while the print times are set at every 10 min. I have tried several modifications - increasing the duration of the simulation, changing the print times, boundary conditions, but without success. I rebuild the example from scratch and everything works fine. As I indicated before, the example worked correctly at first. I ran into trouble only after I made some changes to the initial conditions.
The other issues have been resolved. I had set the BDRC options correctly for some of my examples but not for others. |
|
|
| |
Topic |
|
|
|
Hydrus quirks?
