Elv

RE: Problem to find equilibrium under variable discharge

2023-03-09T06:24:20Z

Hi,

The space marching method is something for Liselot to answer... I will reach her. The crash occurs because the size of `Mak` does not match with the number of flownodes (`dx`) and the number of effective size fractions (`nef`). There must be something murky there. Maybe you could stop in debug mode and check what are the dimensions of `Mak`.

Regards,

V

Issue running the base bifurcation example provided in the repository

2023-03-08T15:37:46Z

Hi,

I am exploring the bifurcation options in ELV. Unfortunately, the input file for the bifurcation is not working. Can someone help me with the error codes---

***
Size inputs must be integers. in check_bra, at line 185
The parents are:
1)    check_input, at line 109
2)    ELV, at line 85
3)    run_ELV, at line 53
4)    oh_ELV_please_run, at line 157
5)    main_run_single, at line 37

***

The input file is the same as in svn. But I have attached it here.

Thanks in advance.
Best wishes,
Kifayath

Problem to find equilibrium under variable discharge

2023-03-08T15:38:44Z

Hi,
I have the following question-

1. I have been trying to calculate using the space marching method (initype=52) for unisize sediment single branch problem. For initype 2, the code runs smoothly but with 52 it crashes with the following set of messages-

***
Number of elements must not change. Use [] as one of the size inputs to automatically calculate the appropriate size for that dimension. in sediment_transport, at line 145
The parents are:
1)    march_step_uni, at line 62
2)    ini_equi, at line 130
3)    initial_condition_construction, at line 259
4)    condition_construction, at line 85
5)    ELV, at line 106
6)    run_ELV, at line 53
7)    oh_ELV_please_run, at line 157
8)    main_run_single, at line 37

***

I tried to follow the error code, but could not solve it. What is the problem in this case? I have attached the input file.

Thanks in advance.
Best wishes,
Kifayath

RE: Issue with unisize runs in ELV

2023-02-09T11:15:01Z

In addition, I am using flowtype 1 (steady), and steady solver type 1 (energy euler)

Issue with unisize runs in ELV

2023-02-09T11:13:16Z

Hi! I am trying to run a very simple unisize, constant discharge, single branch model in elv to find the equilibrium solution. The inputs look like the following:
input.sed.dk =[2]./1000; input.ini.Fak = [1]; input.ini.fsk = [1];
input.ini.slopeb = 1*1e-4; input.ini.initype=1;

input.bcm.type=1; input.bcm.NFLtype=1; input.bcm.NFLparam=[1e-4,0.05]; Qbk0_base = 0.05; input.bcm.Qbk0_tot = 0.05;

input.bch.uptype=1; input.bch.Q0=2000; input.tra.cr=2 %Engelund Hansen;

input.mor.gsdupdate=7;

input.grd.L=100e3; input.grd.dx=2000; input.frc.Cf=0.004; input.mdv.dt=86400

When I run this (along with the other inputs) input set up to calculate the equilibrium with initype=1, the get_equivals_V gets an error - "Error using optimoptions
Invalid solver specified. Provide a solver name or handle (such as 'fmincon' or @fminunc).
Type DOC OPTIMOPTIONS for a list of solvers."

if I choose initype=2, then the substrate runs out. I am not sure which parameter I am putting in wrong. Any help would be much appreciated.

RE: Normal Flow initial condition not found

2021-06-11T13:18:41Z

Hi

I guess you are getting the error in the log-file:
-ATTENTION!!! Normal flow initial condition not found. Maximum relative error

The essence is that the non-linear solver does not find a solution. It is difficult to know what is the problem without the input. What is your input? In general, I would follow the following strategy to find the issue. First, double check that all input make sense. If you are using a lot of size fractions, consider reducing the number of size fractions to make life easier for the solver and see whether it also has a problem in that case. To check that what you want to do is indeed happening under the hood, consider running in debug mode (there is a flag for it). This will allow you to stop the simulation, for instance, in <ini_normalflow_L> before calling <get_equivals_V>. Finally, you can directly call this functions outside Elv to make it even simpler to debug.

V

Normal Flow initial condition not found

2021-06-11T13:04:54Z

I am using initype=1 to set normal flow conditions but get the warning that "normal flow initial condition not found". From what I can tell, the equilibrium slope does not converge in the solver. I am using constant sediment and water discharge. The downstream boundary condition is water level. Any suggestions to resolve this?
Thanks!

RE: Node point for change in discharge

2021-05-20T14:55:30Z

Not sure which software are you using. In any case, the water discharge at a bifurcation is not presribed, as it is the results of the hydrodynamic set of equations. Worded differently, given the upstream discharge and the water level at the downstream ends, the water discharge in each branch is uniquely given, as the water level in the bifurcation must be the same for the 3 branches. If you are thinking about sediment partitioning, what you have to specify is not the sediment transport but the nodal point relation (i.e., a relation between discharge and sediment transport for the branches).

Node point for change in discharge

2021-05-20T14:35:41Z

Hi,

I am trying to model a bifurcating river. How can I change the discharge at a point other than the boundaries?
thanks!

RE: Error when modelling with mixed sediments

2021-03-28T13:23:24Z

Hi Bas,
Great to know that you are using Elv! The error message indicates that the bed level is below the lowest level of the substrate. The lowest level of the substrate is the bed level minus the sum of all the substrate layers. Worded differently, the error message indicates that the thickness of all substrate layers at one point is 0. If, for instance, initially you have 10 substrate layers of 0.5 m plus the active layer of 1 m, in total you have a 6 m substrate. In this example, if somewhere you degrade more than 6 m: ERROR.
It is strange that you say that the error appears where a sedimentation wave arises. As explained above, the error is related to excessive erosion. There must be erosion somewhere.If this happens in the first time step it indicates that from the start there is a huge amount of erosion. That is problematic and probably unreallistic. The first question I have is: are you solving steady or unsteady flow? If it is unsteady, maybe the initial flow condition is causing the large erosion. In this case, yu could consider using 'input.mor.Tstart' for preventing morpodynamic update during the start of the simulation (i.e, spin-up).
If flow is steady, there is no initial flow condition. Then, it must be related to the schematization. Indeed, by using input.mor.ThUnLyrEnd you can create a very thick last layer for preventing the error. I would suggest that you use an extreme value (1000?) just for computing a few time steps and be able to plot the output. Then we can see what is going on. There is no computational extra cost for having a very thick last layer, so it is always a good idea to have a thick one. It will not prevent the results being unreallistic, but it will allow you to prevent the crash and see what is going on.
How is your flow? Do you impose a constant discharge or a hydrograph? How large is your space step? If you impose a hydrograph with a small time step, expect some issues to arise. Let me know and we discuss it.
Regards,

V

Error when modelling with mixed sediments

2021-03-26T16:30:21Z

Hi!
I am studying the morphodynamic effects after widening a certain reach. Directly after the first time step, I get the following error:
'You are getting close to the centre of the earth since you run out of substrate material. Ask V about input.mor.ThUnLyrEnd to solve this problem. Come with coffee please '.

Unfortunately no real-life coffee for now

The location where the Ls of the top-substrate layer approaches 0 is the same location where a sedimentation shock wave rises. I believe this is a common error (based on a previous conversation with Liselot). What is happening here? Is - at the location where we observe the sedimentation shock wave - a new substrate layer created on top of the previous top-substrate layer for which a very small Ls is registered and it therefore looks like the new top-substrate layer is almost fully eroded? Is there a way we can solve this (using input.mor.ThUnLyrEnd maybe)?

Best!

Bas Gradussen

Edit: I tried running with input.mor.ThUnLyrEnd = 4 (2 times input.mor.ThUnLyr). This did not solve the problem.

Initial bed with a certain shape

2020-11-23T11:23:17Z

Hi Hermjan, The best way to specify a gaussian curve as initial condition it is to use a free initial condition and specify the bed level:

1	q=2;
2	sig=150;
3	mu=600;
4	etab_max=2;
5	etaw=4;

1	x=input.grd.dx-input.grd.dx/2:input.grd.dx:input.grd.L-input.grd.dx/2;
2	etab=etab_max*exp(-(x-mu).^2/sig^2);
3	h=etaw-etab;

1	input.ini.initype=2;
2	input.ini.u=q./h;
3	input.ini.h=h;
4	input.ini.etab=etab;

RE: No convergence in ELV for backwater computation

2021-01-05T08:27:20Z

Hi Hermjan,
There are several backwater solvers which vary in accuracy (1st or 4th order) and in variable solved (energy or flow depth). There are some minor details for chosing one or another one, but all of them should provide you with a good solution. The fact that a solution is not found is a symptom that a hypothesis underlying the application of a backwater solver is not fullfiled.
Concerning theta in the quasi-steady and unsteady solvers. The system of equations solved is fundamentally different in each case. Essentially, one has two unkonws in space and one in time and the other only one in time and two in space. Hence, the numerical scheme that solves one system of equations does not need to be valid to solve the other one. In particular, a fully implicit time integration is required for the quasi-steady scheme to be stable, while the unsteady scheme is stable with half implicit half explicit and one gains a second order accuracy.
Before running a morphodynamic simulation, I would test that the hydrodynamics are correct. To this end, I would recommend you to conduct a convergence study. Simply decrease the grid size and check whether the solution converges. This will also tell you how fast does it converge (i.e., order of accuracy) and it is an essential step to firmly decide on the necessary space step for your simulation. Furthermore, I would recommend that you compare with an analytical solution such as the propagation of a Gaussian flood wave. Once the hydrodynamics are clear and beyond doubt, next step is including morphodynamics. Otherwise, one is never sure if the results are genuine or the spurious due to poor numerics.

RE: No convergence in ELV for backwater computation

2021-01-05T08:27:20Z

Bijzonder... The steady model should be the most robust one, as it solved an ODE, contrary to the quasi-steady and unsteady, which solve a PDE. The flow initial condition is irrelevant in the case of the steady model. Only the initial bed level and the downstream flow boundary condition matter. I guess there is an issue with the downstream boundary condition. What are you imposing? Copy paste the log-file that the model generates.

RE: No convergence in ELV for backwater computation

2021-01-05T08:27:20Z

Sorry, I meant the steady model doesn't start (I keep up mixing up the definitions of the models). Unsteady model and quasi-steady model (with same initial conditions as for the steady model) do start.

RE: No convergence in ELV for backwater computation

2021-01-05T08:27:20Z

The quasi-steady model does not solve the backwater equation. If there is no convergence in the first time step is because the initial condition is not accurate enough.

No convergence in ELV for backwater computation

2021-01-05T08:27:20Z

I tried to start identical simulations (also identical initial conditions) with the steady model, quasi-steady model and unsteady model. The quasi-steady model doesn't start (no convergence in 1st time step). The other models do run. Time step reduction doesn't help. Any idea why the computation of the backwater curve seems to be more sensitive and any suggestions to solve it?

RE: ELV downstream normal flow condition, further to previous answer

2021-01-05T08:27:51Z

Hi,

If in both cases the intial condition you impose is normal flow and the discharge remains steady, I see no reason for finding strong erosion in one case and not in the other one. If the discharge is unsteady (it is wise to first make a run with steady discharge to check that boundary conditions behave as you want), does erosion occur before the wave gets to the dowsntream end? If yes, it seems that the initial condition is not normal flow in the case of extreme erosion. Check the bed level of the most dowsntream location. You have to adapt this to both cases.

RE: modelling concepts in ELV

2021-01-05T08:28:16Z

Hi,

There was a bug that I have just solved. The Exner equation is solved differentely for unsteady flow than for steady flow. This is because in the case of unsteady flow, flow can change direction, while it cannot if it is steady. The quasi-steady case was not dealt with.

modelling concepts in ELV

2021-01-05T08:28:16Z

Running a simple simulation with quasi-steady model approach (input.mdv.flowtype=2) : constant discharge upstream, constant water level downstream, still doesn't work. It gives following Warning: the power of the excess shield should be the same in the sediment transport relation than in the entrainement formulation. Otherwise tt may create a discontinuity in Dk because it is 0 when Qbk_st is 0 but it may cancel with the term in Ek_st
And ends with ¡¡ ERROR !!
Supposedly you do not end up here in bed_level_update, at line 137
Same simulation with input.mdv.flowtype=4 (onlu change in input) does start, although it gives the same warning. Any idea/tip?