Ponding between layers

A discussion forum for users of the new HYDRUS 2D/3D. HYDRUS is a software package for simulating water, heat and solute movement in two- and three-dimensional variably saturated media. Happy Posting!
Post Reply
lpom
Posts: 10
Joined: Wed Mar 14, 2018 10:27 am
Location: United Kingdom

Ponding between layers

Post by lpom » Sun Feb 16, 2020 9:32 am

Dear all,

I have been having some trouble calibrating some experimental results. The experiment consists on simulating a railway drainage system, and I'm traying to calibrate the pressure heads at different depths in the subballast layer. The main problem that I'm having is that I'm not being able to match the pressure heads during the wetting period (at a constant flow rate). I have been working on Van Genuchten parameters and I just achieve changes during the drying period. I believe this is because the water is accumulating in the interface between the ballast and subballast (extremely high Ksat on the surface followed by a lower Ksat value) because the inflow is larger than the infiltration capacity of the subballast layer. In general, results are better at the bottom of the experiment and the closer they get to the ballast layer, the worse they get.

Does anybody know how could I lower down the pressure heads during the wetting period? If this is result of water reaching the ballast layer and ponding in the interface, how could I repreduce this in HYDRUS?

Please find attached the file as a reference,

Many thanks and kind regards,

LP
Attachments
HYDRUS doubts heads.zip
(1.18 MiB) Downloaded 10 times

DandD
Posts: 14
Joined: Wed Jan 15, 2020 12:11 pm
Location: Germany

Re: Ponding between layers

Post by DandD » Fri Feb 21, 2020 12:55 pm

Hi,
how does your experimental data look like, if I may ask?
Looking at your model, kind of steady conditions are reached in a short period of time (3h or something like this) during the constant flow rate. Do you see these steady conditions in the measured values? Or more specifically - what is your expectation?

Another question - are the parameters in the model you uploaded the ones you used or did you upload some kind of 'basic' model and altered the parameters?
The combination of the soil parameters given seems bit strange to me. E.g. relatively high Qs with an extremely high ksat and then a small n value...

If you have measured data - have you tried fitting with an inverse solution? (I would not suggest fitting all soil specific parameters at the same time, though...have you measured any in some way?)

I just had a quick look into your simulation and some questions popped up (see above) - let's see if this is helpful in any way.

Best regards
Frieder

lpom
Posts: 10
Joined: Wed Mar 14, 2018 10:27 am
Location: United Kingdom

Re: Ponding between layers

Post by lpom » Sun Feb 23, 2020 10:17 am

Hi Frieder,

Please find the results I'm trying to replicate in this paper https://www.dropbox.com/sh/kx7gtim32hp7 ... efYRa?dl=0
I am able to achieve steady state conditions, but I'm being unable to find a better fitting to the head pressure values and that'is why I'm wandering if there might be any physical effect that I'm missing in the model.

At first I based my VG parameters values on the paper attached but since they don't match very well the head values in the upper observation points, I started modifying them with little success.

I haven't tried inverse modelling the results, since I only have head pressure values for the lower layer.

Then, should I try to obtain the Van Genuchten parameters for the lower layer first using inverse modelling? I'm not sure how to proceed next.

Many thanks,

LP

DandD
Posts: 14
Joined: Wed Jan 15, 2020 12:11 pm
Location: Germany

Re: Ponding between layers

Post by DandD » Mon Feb 24, 2020 9:37 am

Hi,
I used your uploaded model and only changed the n values to 2 for ballast and pea gravel (according to the articel) and 1.5 for the subgrade. If I got it all right, the results in the observations show a similar or even better fit compared to Fig. 12 a & b in the article (looking at the wet steady conditions).
If you want to improve the fit by changing the van Genuchten parameters - you should think about which parameter impacts which model output/result.
E.g. during the steady conditions, the subgrade is completely saturated (right?) - so changing the van Genuchten parameters will not change a thing I believe.
So you should maybe focus on the ballast layer (unsaturated conditions) and see, if changes in specific paramters actually have an effect (and what effect in particular) on your results. I am a bit worried, that they will not have a big impact because of the material used ('extreme' values due to basically stones as material). Make sure they stay in a reasonable range.

So I would suggest checking the ballast layer and if changes of the parameters actually lead to different results.

Due to the different materials, the main flow paths a really narrowed down ('bottleneck', see picture)...if changing ballast parameters does not help I would think about the conceptual model again. From looking at the articel/experiment I would have expected some kind of outflow on the left side of the ballast layer (depending on rain intensities)...not sure if they write anything about it...
Other than that, I am no expert on train track drainages and never have such high permeable soils (or rather stones - not sure how well the van Genuchten model can represent them, actually).
Good Luck!
Frieder
Attachments
streamlines.PNG
bottleneck (low permeable sediment at the bottom vs relly high permeability in top)
streamlines.PNG (204.95 KiB) Viewed 241 times

lpom
Posts: 10
Joined: Wed Mar 14, 2018 10:27 am
Location: United Kingdom

Re: Ponding between layers

Post by lpom » Fri Feb 28, 2020 9:40 am

Thank you for the suggestions. Will do the changes.

Kind regards,

L

Post Reply