Sunday, May 14, 2023

Approaches for representing a culvert within the 2D domain (ICM)

 From the Innovyze Blog 

Approaches for representing a culvert within the 2D domain (ICM)

This article details some of the methods available to the user when they are trying to represent culverts in ICM which accept flow from the 2D mesh. These approaches will extend to underpasses and similar structures, although they may be best represented using Base Linear Structures (2D).
Mar 8, 2022Knowledge



In this article we’ll give recommendations on how to model culverts which will interact with the 2D mesh in our ICM model. For this purpose our culvert is assumed to be a structure which has a depth which may impact its hydraulic performance. For structures which are less than a 2D element in depth the "Base linear structure (2D)" line may provide the best form of representation. More information can be found in the Help files: LINK

When modelling a culvert in the 2D domain it is different from a normal 1D pipe configuration; culverts don't have manholes at either end, they open directly to the channel. The wingwall arrangement, if present, is likely to control the hydraulic performance at the inlet.

When considering the “ground level” it is very important to match the culvert inlet and outlet levels with the bed of the channel. That is where the linkage is going to be taking place and where the calculation points will be taken from.


2D Conduit

2D conduits are recommended if stability is a concern. This is because the flow does not have to be exchanged between the 1D and 2D engine at the coupling points (nodes). It will also ensure the continuity of mass and momentum in the calculations. However, at present the 2D Conduits do have limited functionality in terms of shapes and inlet / outlet headlosses cannot be modelled directly.
Connect 2D > 2D Conduit > Connect 2D

This approach is best avoided for small pipes as the conduit is descritized into elements. This could lead to the generation of small internal elements and significant impact model run times.

1D Conduit

If you need to model the culvert using the FHWA/CIRIA method, things can get tricky because the method is only developed for 1D. In theory, we should model a section of the channel and the culvert in 1D. Therefore, we’ll need to model the channels as 1D river, and connect the 1D river to the 2D surface. The system could look like this:
Inline bank > River reach > Culvert inlet > Conduit >  Outfall 2D

We might need to a Culvert outlet link and another section of downstream river reach with an inline bank if reverse flow is expected.

As you can see this approach can be quite tedious, and it is still prone to stability issues with the complication of 1D river sections and inline banks. And without calibration data, it can be hard to judge which method is more accurate.

In practice, you might want to try different variations of this setup. For example, you can simplify this setup without adding the river reaches and the outlet link but it would be sensible to check the stability and validity from the model results:
Outfall 2D > Culvert Inlet > Break node > Culvert > Outfall 2.

1D-2D linkage basis

When linking a culvert to the 2D mesh using Connect 2D or Outfall2D, in most cases we should use the "Depth" basis because a culvert should be flush with the channel bed. There may be some very localised scour, but we can overlook that in the majority of cases. The "Elevation" linkage should only be used where the culvert entrance / exit is at a high level above the ground level (i.e. channel bed level).
If the entrance to the culvert is below the ground level of the terrain model, then the mesh elements at the entrance will have to be lowered to that level to reflect the topology of the channel bed. This can be achieved using a mesh zone or mesh level zone. It can also be a very good idea to use a mesh zone to increase the element size where the connection is made, this can help to increase the element area(volume) and reduce drying or oscillations that may occur on small elements.


Can I use node types other than Connect 2D for 2D conduit? If so, which type should I use?

When using conduit 2D, you can choose to use types other than connect 2D, but it will turn the node into a manhole and transfer the flow into the 1D engine. This would negate the reason for using the 2d conduit.

When modelling a 1D culvert, what type of node to choose from? Outfall 2D, manhole, connect 2D which one to choose from?

If you are modelling using a 1D conduit, use Outfall 2D to connect the culvert to the 2D. You should not use a manhole type node to connect a culvert to 2D. The node type between a river reach and any link is always a "break" node type.

What is the difference between 1D-2D linkage basis elevation vs depth?

Users should really follow the rule that “elevation” applies when the conduit invert is above the element level i.e. the pipe is significantly raised above the bed. Minor differences due to scour can typically be ignored.

For Outfall2D/Connect 2D when using elevation basis, what elevation is used?

ICM will look at the elevation of water in the element and the elevation of water in the node/conduit to calculate a head difference and flow rate.

How to model blockage and flap gate for culvert?

Blockages can be modelled by either adding “sediment” to the conduit or adding a “blockage” link to the appropriate location. Note that a blockage link is intended for calibration against recorded data. A flap valve can be modelled to ensure unidirectional flow. However, all of these would currently dictate employing the 1D-2D connection mentioned above.

For a more in-depth discussion of culvert design, refer to this medium post: HDS 5 .

No comments:

The Goal of SWMM5 Input Files

 🌟 SWMM5 (Storm Water Management Model 5) is a widely used urban hydrology and hydraulic modeling software developed by the United States E...