null

Message Boards

simulation of Gravel/Cobble sized sediment

SM
Sam McWilliams, modified 5 Months ago.

simulation of Gravel/Cobble sized sediment

Youngling Posts: 1 Join Date: 6/13/19 Recent Posts

Hello,

 

I am simulating a shallow river system  in Delft3D-Flow which includes the transport of silts, sands, gravel and cobble (16 and 100 mm  gravel and cobble respectively). I am defining the larger sizes classes as sands to implement in operation->discharges upriver rather than as boundary constituents to remove the issue of settling at the inflow boundary. This seems to work well and sediment moves downstream as expected into the rest of the system. This is confirmed by checking the mass across a cross section near the outfall, confirming the expected mass is entering the system. The rest of the system has 0 mass defined in the bed.  

 

However, when comparing input mass of sediment with total mass within the system for each size class, the gravel and cobble size classes show an order of magnitude increase in sediment mass within the system.  A series of the following messages appeared in the tri-diag file but were suppressed due to their frequency. The gravel and cobble size classes were both referenced (Fraction :#) 

 

Sediment erosion shortage at NM 23936 Fraction: 2 Mass available   :          -0.1358E-44 Mass to be eroded:          -0.1358E-44 

My questions are thus: 

1. Is the message above indicating the sediment is being eroded from the bed despite no mass in the bed to begin with? Is this to satisfy an internally imposed concentration within the system? If, so how do you force a no erosion condition at the bed to eliminate this problem? 

2. Do you have to define grain sizes greater than 2 mm as SedTyp ="bedload" or can the model handle SedTyp="sands" with D50 greater than 2mm? The Van rijn equations referenced in the model do not indicate an upper limit. 

 

Note, There are warnings about velocity change too high but no Courant messages indicate timestep is right on the cusp, i.e. advised timestep is <2-3% of applied value. 

 

Thank you for your time,

-Sam

Richard Measures, modified 5 Months ago.

RE: simulation of Gravel/Cobble sized sediment

Jedi Knight Posts: 178 Join Date: 3/23/11 Recent Posts

Hi Sam,

In answer to your questions:

Q1.

That warning message occurs when, in a single timestep, more erosion is calculated than there is sediment available for a particular sediment fraction in a particular cell.

Delft3D restricts the entrainment/transport based on the availability of sediment within the surface layer. For example if there is only 1mm of sediment on the bed, but the surface layer is set to 100mm thick the transport rate will be limited to 1% of what it would be if there was 100mm or more of sediment. This rate limiting is the same mechanism which prevents erosion where no sediment is available.

Delft3D handles the excess erosion by remembering the sediment deficit For example, if 2kg of a particular fraction erodes but there is only 1kg available then the cell would be left with a deficit of 1kg of that fraction. No further erosion of that cell with then happen until the defecit is filled. This defecit tracking means that while the error will introduce a small mass balance error in the short term the mass balance error should eventually resolve itself (or at least cease to get bigger).

This warning is more likely if morfac is set higher or the surface layer is set thinner - so changing these settings may provide a way to reduce the frequency with which it occurs. I hope that makes sense?

Q2.

I think it is fine to define sizes greater than 2mm as sand, and I have done this in the past with no issues.

Cheers,

Richard