Community header

Forum

Bars in rivers

HA
Abdelrahman Abdou, modified 5 Months ago.

Bars in rivers

Keen Forecaster Posts: 6 Join Date: 9/2/20 Recent Posts
Hi, I have a question. I built up my model as a straight channel 2DH simulation. Then I put perturbation in the upstream boundary (discharge per cell) to trigger instability and bar formation. The perturbation was with intensity 1% for each cell and I added noise to discharge along with the time series as well (change in discharge amplitude every 60 minutes (my time step is 0.5 minutes)) and my model begins to produce bars but after around 30 years of the simulation the bars tends to disappear and I did not reach the equilibrium state with my model. what could be my mistake?
2- changes in bed began to appear after more than 5 years of simulation I think it is a bit slow, what could be the reason ??
3- When I used the equation of E-H there are morphological changes but when I tried Van Rijn 1984 with the recommended value (Acal = 1, AKS = 1, Ws(settling velocity calculated) =0.04) the bed morphology is the same, what could be the reason?
Victor Chavarrias, modified 5 Months ago.

RE: Bars in rivers

Famous Fortune teller Posts: 48 Join Date: 4/24/20 Recent Posts
Hi Abdelrahman, 
Your question may be more related to bar theory than to the software. It raises several questions. There exist several types of river bars. Which type of bar do you intend to reproduce? For instance, to reproduce free bars no instability at the upstream boundary condition is needed. If the width-to-depth ratio is above the critical value, bars will emerge in your model and will travel dowsntream (under certain condition they travel upstream, but it is rare). What is your width-to-depth ratio? is it above critical? 
You mention that you impose a perturbation at the upstream as a 1% difference in discharge varying with time. That makes me think that you want to have forced bars. Is the discharge the same across the upstream boundary? Or is there a variation in the lateral direction? 
You talk about equilibrium. That reinforces the feeling that you want to have forced bars. In that case, check whether you may have them imposing a variable discharge with time. 
The time for bars to emerge depend also on the type of bars but 5 years seems a long time. What is the sediment transport rate? That is the main factor controlling the growth rate of bars.
In case of free bars, the critical width-to-depth ratio depends on the sediment transport relation. It may be that the bed is stable when using Van Rijn (1984) but unstable when using Engelund and Hansen (1967). Most important, check the bed-slope-effect closure relation. This has a very large impact. Check also which kind of sediment you are considering. Possibly you have sediment type 'sand'. This means that it can be both transported as bedload and suspended load. The fraction in each transport mode depends on the sediment transport relation. For the case of EH67, it is all transported as bedload unless you specify a certain input parameter. Consider setting everything as bedload if the processes related to suspended load are not relevant by changing the sediment type to 'bedload'.
HA
Abdelrahman Abdou, modified 5 Months ago.

RE: Bars in rivers

Keen Forecaster Posts: 6 Join Date: 9/2/20 Recent Posts
Thank you for your reply. I just realized that it was moved to a new section.*from bar mode calculation I have m = 4.96 so I expect multiple bars with the discharge and dimensions I provide.(Is the discharge the same across the upstream boundary? Or is there a variation in the lateral direction? )
*There is variation in discharge value along the transverse section with amplitude(+ or -) 1% in other words i have 30 cells on the transverse section and i have a total discharge with 1331 m3/s, so i divide (1331/30 = 44.37 m3/s per cell then I use code to make a perturbation along the transverse cross-section with amplitude (+ and -1%) so the values of discharge per cell are randomly between ((44.37+0.01*44.37) and (44.37-0.01*44.37)).
*my channel is 655m width*12000 m length (cell size (22.83 m*40m )) and the bed slope 3.5 cm/Km
*Horizontal eddy visc = 1 , Hz eddy diffusivity = 10* I use only sand with (d50 =0.378 mm) and I did not impose sediment transport concentration, I use equilibrium sand concentration profile at the inflow boundaries*Initial sediment layer thickness at bed =10 m*bedslope effect (Koch and flokstra)alfabs =1, Ash=9, Bsh=0.5, Csh=0.3, Dsh=1*Secondary flow Espir =0.5(That reinforces the feeling that you want to have forced bars. In that case, check whether you may have them imposing a variable discharge with time.)
*I did not get this point completely, do you mean if I impose variable discharge with time, I would not reach equilibrium bar mode (bed formation)? so if I want to run like a real case with different daily discharges along with the whole time series, you mean I will not reach equilibrium? 
*In my simulation, I imposed variation in discharge with amplitude (+or - 1%) of Qtotalalong the whole time series.
Thank you for your help!
Regards, 
HA
Abdelrahman Abdou, modified 5 Months ago.

RE: Bars in rivers

Keen Forecaster Posts: 6 Join Date: 9/2/20 Recent Posts
[MorphologyFileInformation]
   FileCreatedBy    = Delft3D FLOW-GUI, Version: 3.59.01.57433         
   FileCreationDate = Wed Dec 02 2020, 23:37:24         
   FileVersion      = 02.00                        
[Morphology]
   EpsPar           = false                         Vertical mixing distribution according to van Rijn (overrules k-epsilon model)
   IopKCW           = 1                             Flag for determining Rc and Rw
   RDC              = 0.01                 [m]      Current related roughness height (only used if IopKCW <> 1)
   RDW              = 0.02                 [m]      Wave related roughness height (only used if IopKCW <> 1)
   MorFac           =  1.0000000e+000      [-]      Morphological scale factor
   MorStt           =  7.2000000e+002      [min]    Spin-up interval from TStart till start of morphological changes
   Thresh           =  5.0000001e-002      [m]      Threshold sediment thickness for transport and erosion reduction
   MorUpd           = true                          Update bathymetry during FLOW simulation
   EqmBc            = true                          Equilibrium sand concentration profile at inflow boundaries
   DensIn           = false                         Include effect of sediment concentration on fluid density
   AksFac           =  1.0000000e+000      [-]      van Rijn's reference height = AKSFAC * KS
   RWave            =  2.0000000e+000      [-]      Wave related roughness = RWAVE * estimated ripple height. Van Rijn Recommends range 1-3
   AlfaBs           =  1.0000000e+000      [-]      Streamwise bed gradient factor for bed load transport
   Sus              =  1.0000000e+000      [-]      Multiplication factor for suspended sediment reference concentration
   Bed              =  1.0000000e+000      [-]      Multiplication factor for bed-load transport vector magnitude
   SusW             =  1.0000000e+000      [-]      Wave-related suspended sed. transport factor
   BedW             =  1.0000000e+000      [-]      Wave-related bed-load sed. transport factor
   SedThr           =  1.0000000e-001      [m]      Minimum water depth for sediment computations
   ThetSD           =  0.0000000e+000      [-]      Factor for erosion of adjacent dry cells
   HMaxTH           =  1.5000000e+000      [m]      Max depth for variable THETSD. Set < SEDTHR to use global value only
   FWFac            =  1.0000000e+000      [-]      Vertical mixing distribution according to van Rijn (overrules k-epsilon model)
   Espir            = 0.5                  [-]      Calibration factor spiral flow
   ISlope           = 3                    [-]      Flag for bed slope effect
   AShld            = 9.0                  [-]      Bed slope parameter Koch & Flokstra
   BShld            = 0.5                  [-]      Bed slope parameter Koch & Flokstra
   CShld            = 0.3                  [-]      Bed slope parameter Koch & Flokstra
   DShld            = 1                    [-]      Bed slope parameter Koch & Flokstra
[Numerics]
   MaximumWaterdepth = false
   LaterallyAveragedBedload = false
   UpwindBedload = true
Victor Chavarrias, modified 5 Months ago.

RE: Bars in rivers

Famous Fortune teller Posts: 48 Join Date: 4/24/20 Recent Posts
Hi, The case you are modelling seems unstable to alternate bars. That means that even without a any perturbation (in discharge, initial bed level, or width), alternate bars will start to grow, merge, and travel downstream. Do you observe this in your simulation?It seems that you want to obtain an equilibrium state (i.e., bed level not changing with time). If the bed is unstable to alternate bars, this may not happen. You may obtain a situation which is in statistical equilibrium: bar properties will remain constant but will move and travel downstream and new ones will be created upstream. It may also happen that bars grow and emerge.I am sorry to say that I do not fully grasp what your intentions are and which questions you want to answer with the model. Clarifying this may help in defining the type of simulation you need to run. 
HA
Abdelrahman Abdou, modified 5 Months ago.

RE: Bars in rivers

Keen Forecaster Posts: 6 Join Date: 9/2/20 Recent Posts
Thanks a lot for your help. I appreciate that. My study focuses on the large-scale impact of installing bridge piers on river morphology with different scenarios of piers installation and different cases. First, I want to simulate the river without any piers and run a simulation for the case with no piers, let the bed developed and bar forms (periodic bars (free or steady)) until it reaches equilibrium so that I could compare the differences afterward when I install piers in the next scenarios. what I expected that depending on the width to depth ratio and other parameters like discharge, friction, and bed slope I will witness a bar mode as calculated from the formula and after a while, it would reach the equilibrium state. but the fact is that it did not reach the equilibrium and the bars are disappearing as shown in the following figures. The figures below are the output of the simulation after 10, 20, 30, 50, and 60 years successively. what do you think is there something wrong with my approach or my simulation?
HA
Abdelrahman Abdou, modified 5 Months ago.

RE: Bars in rivers

Keen Forecaster Posts: 6 Join Date: 9/2/20 Recent Posts
Thanks a lot for your help and for Alessandra Crosato who contacted me and answered my questions.
Victor Chavarrias, modified 5 Months ago.

RE: Bars in rivers

Famous Fortune teller Posts: 48 Join Date: 4/24/20 Recent Posts
I am glad to know that your questions are answered. Still, please let me mention some points of attention. Delft3D solves the hydrostatic shallow water equation. Hence, it is not appropriate for capturing three-dimensional flow strucutures as the ones occuring in the wake of bridge piers leading to local erosion. In Delft3D, you can account for the effect of bridge piers in the flow, but this is modelled as a subgrid phenomenon in which, basically, a certain energy loss is added to the momentum equations. For this reason, the local morphodynamic impact of a bridge pier is not well reproduced. 

A second general point is that there is no need to have equilibrium conditions prior to study the impact of interventions. The concept of a river in equilibrium is theoretically perfect but unrealistic in the field.