Results 1 to 6 of 6

Thread: 1D overland flow discharging to 2D mesh

  1. #1

    Join Date
    Aug 2015
    Posts
    3

    1D overland flow discharging to 2D mesh

    Hello all,

    I am currently working with a 1D/2D coupled hydraulic model of a river in Infoworks ICM. My model includes the sewer network (combined and storm systems), the river that runs through the city and its surroundings, and the urban ground surface for which a 2D mesh was built with LiDAR data. These three systems are coupled and flows from a system to another are taken into account.

    My aim is to estimate the velocities and the water depths on the ground surface during a severe flood that caused significant damage in the urban area (up to 2 meters of water depth).

    During 1D/2D coupled simulations of the storm, I noticed that not enough water is present on the 2D zone compared to some measurements, because many of the manholes that are flooded and that release flood water on the ground surface are located outside of the 2D mesh, where LiDAR data is missing. I would therefore like to take that flood water and put it back in the 2D zone to observe the velocities and depths, using only 1D links and nodes.

    I have tried several options that have already failed:

    - using 2D points sources referring to inflow events calculated from the flood volumes : no water observed on the 2D zone despite matching 2D point sources with inflow events
    - using overland conduits (representing the roads where flood water runs) from manholes outside of the 2D zone to manholes inside the 2D zone

    The last option, using overland conduits to represent roads as channels for the flood water, seems to be the most promising. I however have issues regarding the discharge of the overland conduit to the 2D mesh.

    I tried using Oufall 2D nodes to release the water into the 2D mesh but only in some very rare cases I was able to see water on the 2D zone, despite the fact that the ground level and the flood level of the Oufall 2D node are the same as the mesh element ground level in which it is located.
    I checked that the cumulative DS volume of the conduit is the same as the flow volume into the 2D zone of the outfall 2D node.

    I would really like to know why some outfall 2D nodes do not discharge any water on the surface. I saw in the log result file that some flows from nodes to 2D mesh were limited. Do you think it might cause that problem ?

    If you have alternative methods to put that excess flood water into the 2D mesh, I would like to hear it too.

    Cheers,

    Nicolas

  2. #2
    Forum Moderator

    Innovyze Employee



    Join Date
    Feb 2013
    Posts
    133
    Nicolas,

    I'm a little concerned by the suggestion that the 2D Point sources option doesn't work, I see no reason why it shouldn't.

    Also, based on your description, I see no reason why the 2D Outfalls shouldn't pass flow to the 2D domain.

    How are you assessing the flow on the 2D surface?

    It might be easiest to send your model to support@innovyze.com.

    Duncan

  3. #3

    Join Date
    Aug 2015
    Posts
    3
    Quote Originally Posted by Duncan Kitts View Post
    Nicolas,

    I'm a little concerned by the suggestion that the 2D Point sources option doesn't work, I see no reason why it shouldn't.

    Also, based on your description, I see no reason why the 2D Outfalls shouldn't pass flow to the 2D domain.

    How are you assessing the flow on the 2D surface?

    It might be easiest to send your model to support@innovyze.com.

    Duncan
    Duncan,

    Thanks for the quick reply. First, I want to mention that I don't assess flows quantitatively for now, but rather qualitatively. Assessing flow more accurately will come after I calibrate the 2D zone parameters (roughness, mesh element and triangle size etc). Given the number of runs I have been doing these last days, I could not spend too much time building complex result analysis objects, and a simple visual check was necessary.

    To check if the nodes pass flows to the 2D surface I first use Themes for Velocities and Depths on the GeoPlan Window for the simulation, while making sure that even small depths and velocities are visible. Then, I check the maximum depth and maximum speed of the elements around the nodes that are supposed to pass flow. In many cases I could not see any velocity nor depth on the surface, and maximum depth and maximum speed in surrounding elements were zero.

    I have been trying again using 2D Outfalls and today it worked, as I increased the minimum size of the elements of the mesh around the nodes!
    It appears that too small mesh elements prevent any water from discharging from a node and running on the 2D mesh.
    I had the same issue with a storage node that was clearely supposed to spill water on the 2D mesh. I am trying with larger mesh elements and I will see if it solves this problem too.

    Cheers,

    Nicolas

  4. #4
    Forum Moderator

    Innovyze Employee



    Join Date
    Feb 2013
    Posts
    133
    OK, it would be good to get a copy of your model with the 2D Point sources which don't convey flow onto the 2D mesh to investigate further.

    The 2D outfalls will have the amount of flow limited if the mesh element is small, I would imagine this is what you are seeing. Hence the larger mesh element conveying more flow.

    With regards the storage nodes, these won't automatically spill onto the 2D mesh, unless you have some sort of connection, for instance an inline bank.

    Only manholes with a 2D flood type (2D, Gully 2D and Inlet 2D) and 2D Outfalls will interact with the 2D mesh.

  5. #5
    Quote Originally Posted by Nicolas View Post
    Duncan,

    too small mesh elements prevent any water from discharging from a node and running on the 2D mesh.

    Nicolas
    That's my experience too. Use a number of mesh zones to vary the mesh size sensibly.

    In concept what you're trying to do (use 1D links to transfer the flood flows onto the 2D mesh) should work, but it might be easier in the long run to extend the 2D domain to include all your flooding nodes. At the moment you have to assume where the flow is going to go and how fast its going to get there, which kind of defeats the object of 2D modelling.

  6. #6

    Join Date
    Aug 2015
    Posts
    3
    Ducan,

    The mesh element size seems to matter a lot regarding the flow limiting mechanism. It appears that it also affects 2D manholes discharging to the mesh.
    For the storage node, I implemented inline banks and it worked. Now my retention basin can discharge to the 2D mesh.

    About sending a copy of the model, I will have to check with the project manager first. I am not sure I am allowed to do that.


    Martin,

    Thanks for sharing your experience.

    I would really like to extend the 2D zone, unfortunately the span of LiDAR data I have is limited (and getting more LiDAR coverage is costly and time consuming) and I have already extended the 2D zone to the maximum. The manholes that flood are much further than the border of the LiDAR covered area.

    The second constraint is the simulation running time. It already takes two hours to run, though it is not optimized. My client will want the model to run as fast as possible. Thus the simpler, the better. Some manholes don't flood that much thus using a mesh over a big area just for these nodes increases the calculation time without improving much the interpretation of the simulation results.

    Actually, using overland conduits has shown good results. I can now see water discharging from these conduits (which represent the roads carrying the flood water from manholes upstream to the ground surface).

    I however still have doubts about the ability of the model to exchange water between 2D manholes and the 2D zone. The log report of the simulation shows "Warning 1053: One or more nodes had their flow limited from the 2d zone; see node results grid for more details." Looking at the node results window, I only found the duration of flow limiting but not the amount of flood water that could not reach the 2D zone, which is critical information.

    I am a bit concerned by that since I only used nodes of type "Manhole" and of flood type "2D" and since the help file mentions that only Inlet 2D nodes can have their flow limited. "When modelling 2D inlets, flow capping will be applied to limit exchange between the 2D and 1D network in situations where inflow at the node would exceed volume contained in the 2D element at any given timestep. This may occur if the 2D element that the inlet node is located within is too small."

    I will try increasing the element size and see if it gets better.

    Anyway, thanks for the advice.

    Cheers,

    Nicolas

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •