«

»

Sep 20 2011

2-D Model 101

I have had several posts that talk about and show results of 2-D hydraulic models.  Today I wanted to back up some and provide a brief overview of what goes into making a 2-D model.

In the RMA2 manual, a hydraulic model provided by the US Army Corps of Engineers, there is a helpful chart that illustrates the relative importance of different model items to the simulation.  This is a good starting point, because it is a good reminder of where you should be spending most of your time when putting together a model.

2-D Model Items of Importance

Location

The model I am going to use as an example is a tidal model of Hampton Harbor in New Hampshire, just south of Hampton.

Map picture

Mesh

A 1-D model is made up of cross sections as the main geometric input.  A 2-D model requires a mesh that is made up of nodes and elements.  The mesh is a three dimensional representation of the area being modeled.  Good elevation data is needed before a mesh can be constructed.  As seen in the chart above, the data collection effort and mesh construction will take up most of the effort getting a 2-D model setup.  The tricky part of mesh design is balancing enough refinement where you need it without creating too large of a mesh that takes a long time to solve.  Below are a few pictures of the mesh for Hampton Harbor.  This mesh has a little over 9400 elements and 28000 nodes.  All meshes kind of take on their own unique shape and character.  This one has been referred to as” the lung” or “the kidney”.  The ear/kidney looking thing on the right is the boundary to the ocean.

Nodes and Elementsnodes and elements

Entire MeshMesh

More Detail Near Harbor EntranceMesh_2

Most Detail at Bridge CrossingMesh_3

Elevation

After creating the mesh, it has to be assigned elevation values.  This is straight forward if you have some type of digital elevation model.  Below are a few images of the same mesh with elevations.

Entire MeshElevation

Harbor EntranceElevation_2

With ImageElevation_3

Mesh Quality

After the elevations are assigned it is a good idea to check the quality of the mesh.  There are some mesh conditions that could cause stability problems and slow the time it takes the model to solve or cause the model to crash.  Important quality checks include;

  • Minimum interior angle between element sides
  • Maximum interior angel between element sides
  • Concave quadrilaterals
  • Element area change
  • Number of connecting elements
  • Ambiguous gradient

Roughness

After the mesh geometry is complete it is time to assign the roughness values.  Typically these correspond to Manning’s n values, but not always.  In 1-D models the n values are assigned at various locations across the cross sections.  In 2-D models, each element gets it’s own roughness value.

Roughness Type for Entire Modelroughness

Roughness Type Zoomed Inroughness_2

Roughness Type at Bridgeroughness_3 at bridge

Boundary Conditions

We have a mesh, with elevations, mesh quality checked, and roughness values assigned.  Next we need to assign boundary conditions.  The locations of the boundaries is important to the quality of results at the location of interest.  The example that I am using is a tidal model, so I only have a varying water surface boundary.  The hydrograph below is a 100 year storm simulation of a Nor’easter.  A riverine model would include a tail water and one or multiple inflow boundaries.

Boundary LocationBoundary

Hydrographboundary hydrograph

Viscosity, Settings, Other

There are other settings that have to be determined before running a model, but they start to get more model specific, and can be application specific.  So I cannot cover all of the different settings you might encounter in a short blog post.

This is by no means an exhaustive look at what goes into a 2-D model, but at least provides a basic understanding to what goes into creating one.  Each item mentioned above has it’s own subtle nuances to learn when creating and running models, and I hope to continue to cover some of those later.

Note:  All graphics created in SMS and captured using SnagIt.

5 comments

Skip to comment form

  1. Cameron Jenkins

    Have you had a chance to work with the ADH 2-D model?

    Do you prefer using patch or paving for your mesh type?

    Thanks,

    Cameron

  2. Dusty Robinson

    I have not worked with the ADH-2-D model. Almost all of the modeling projects we work on the client specifies the model that they would like used.

    I usually use patch, but I would say it depends on the application. I have created a few meshes where we wanted a really detailed area that was going to change, so I used paving for that area so the mesh was more “flexible” to the all the scenarios we were running.

    Have you used the ADH 2-D model? If so, what are you using it for?

  3. Cameron Jenkins

    I am currently using the ADH 2-D model for a project where our client wants to create a sandbar habitat for an endangered bird. We are testing different structure sizes and locations in the mesh to find the combination that will create the largest sandbar.

    The pros for the model are that it is more stable than RMA2, has sediment transport, and it has the ability to have elements adapt during run time.

    The cons for the model are it takes longer to run than RMA2. Adaption causes the run time to increase dramatically and is more of a hassle than it is worth it seems like.

  4. Glenn Fischer

    Really great work… especially love the figures and graphics. Keep it up, fellas.

    1. Anthony Alvarado

      Thanks! I like your blog as well, especially the idea of sharing research that way and getting input. Thanks for commenting!

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>