Jump to content
TUFLOW Forum

rjones

Members
  • Content Count

    48
  • Joined

  • Last visited

Community Reputation

0 Neutral

About rjones

  • Rank
    Advanced Member

Contact Methods

  • Website URL
    http://
  • ICQ
    0

Profile Information

  • Gender
    Male
  • Location
    Des Moines, Iowa

Recent Profile Visitors

1668 profile views
  1. rjones

    HPC Souce Inflows

    Hi, Hopefully someone else will have better suggestions. I can't think of a quick/full solution to your problem, but have some ideas on how to look at it. I would start by using the bcc_check file to see how often this happens. You can use your GIS to do a spatial join for this file and the "grid_check" file. This will give you the cell FID or n, m values. Try counting how many cells have more than 4 inflows applied. If there aren't too many, you could create combined versions of inflows that commonly overlap (SA1 + SA2) and use the combined curve where the polygons overlap (determine overlap using GIS files). I hope this gives you some ideas. If not, hopefully someone else chimes in with other ideas.
  2. For really large models, I use the raster output with ArcGIS and it generally works well. When using SMS with really large models, make sure you run the 64-bit version and consider purchasing additional RAM if you are maxing out the available RAM.
  3. ***Sorry, I didn't realize someone had already provided this advice. FYI - At least in Word 2010 there is an easy way to add the ability to go back easily. At the very top of the screen there are some little icons representing the "quick access toolbar". The button on the furthest right is a menu that you can use to customize the options that are seen. Choose the "more commands" option that brings up a dialog. In the dialog, under "choose commands from" change it to "All commands". Then scroll down and select "Back". Add this to the toolbar and you have the primary functionality you are interested in. I'm not saying this is superior to online but I find this very handy when using the TUFLOW manual. I hope this helps.
  4. rjones

    Output Zones

    I very well could be wrong but my guess is that you need to tell which output zones are active using the command "Model Output Zones." Something like: Model Output Zones == Z1 | Z2 I hope this helps. If not hopefully someone else will have the right answer.
  5. This wouldn't be very difficult to do using your favorite GIS software.
  6. Hi Luke, I haven't used the CCA table before but it looks like it isn't a time-series output but is simply a GIS table. If you are using the 32-bit version of SMS, you can use the ArcObjects window to display ArcGIS in the window background. If you are using the shapefile output, you could then thematically map this layer. I'm no longer with Aquaveo working on SMS development. When I was there, we talked a lot about how to better support 1D channels in SMS. The only real support now is through water level lines. I am really not sure what developments have been made in this area but hope you find what you are looking for. Cheers, Rusty
  7. Hi, I am trying to use the new operating controls in the 2013 TUFLOW release. I think I understand how to setup the 1d network layer and the operating controls (.toc) file. What I can't seem to find (maybe it is there and I missed it) is how to read the Operating controls (.toc) file. I am expecting a command in the tcf/ecf file something like: Read Operating Controls == myoperations.toc Any help would be appreciated. Cheers, Rusty
  8. Hi, I am trying to use the new operating controls in the 2013 TUFLOW release. I think I understand how to setup the 1d network layer and the operating controls (.toc) file. What I can't seem to find (maybe it is there and I missed it) is how to read the Operating controls (.toc) file. I am expecting a command in the tcf/ecf file something like: Read Operating Controls == myoperations.toc Any help would be appreciated. Cheers, Rusty
  9. Hi, I'll take a stab at this. NFIP is the referring to the list of supported models on the FEMA website. Only FEMA approved models can be used to update flood maps for the National Flood Insurance Program in the US (NFIP). I have no idea what they mean. SMS is an interface and not a numeric engine and therefore not subject to FEMA approval. SMS is typically used for FESWMS (perhaps there are multiple versions but the one I am familiar with comes from the Federal Highways FHWA rather than USGS), and RMA2 which are both approved for NFIP. It also generally used to setup/run models for SRH2D also mentioned in the presentation. The only thing that I can think of for their reference to SMS is specifically the TUFLOW license part of SMS. TUFLOW is approved for use in FEMA region III but not nationwide. It is odd since the 2D engine is exactly the same as the one used in XP-2D which is approved nationwide (perhaps this approval came after this presentation was created). The approval for NFIP has nothing to do with the quality or usefulness of the model. To qualify for the NFIP there is not an extensive set of tests that non-NFIP models failed or something like this that would indicate the models are inferior. The NFIP process has more to do with the involvement and interest of other government agencies (outside of FEMA). Anyone who has used FESWMS, RMA2, and TUFLOW can attest that TUFLOW is by far more applicable for flood models because of its stability and accuracy. RMA2 and FESWMS are very temperamental, unstable in wetting/drying (flood situations), have to start completely wet which doesn't make sense for flood situations, often requires unreasonable viscosity coefficients leading to inaccuracies, slow, and difficult to work with. FLO-2D which is also approved uses a more simple set of equations than are used by TUFLOW which can lead to inaccuracies. There is no reason to believe that NFIP models are better than non-NFIP models. If you are doing work for the NFIP, you have to use an approved model. If you would prefer to use TUFLOW, you should bring it up with the FEMA reviewers. This may not help your project directly, but the more interest that is shown the more likely it will be added in the future. If you are not doing work for the NFIP, I would ignore their list of approved models because it indicates nothing of value.
  10. Output interval won't hurt you unless it is really small like seconds. It may not be affecting you at all. Anyway, good luck.
  11. Hi, I guess I don't understand why the 1D network would be causing the slowness. In my experience the slowdown is almost always the 2D portion of the model. Is the pipe network a bunch of individual sections so you end up with lots of separate 1D elements? You could represent the whole thing as a single polyline. This will affect the pipe losses but will get flow from the upstream to downstream end with less computations. I would first check to see if you can increase the timestep in both the 1D/2D portions of the model (watch for instabilities and find a timestep that works well). Also check the output interval for the 1D and make sure that it is reasonable. If it accidentally got set to output every few seconds or something, it will slow down the simulation. Lastly, perhaps you have done this but I would run with and without the pipe network to see how much it slows down computations. I hope that my responses are helpful and didn't misunderstand the situation. If not, hopefully someone else will chime in with some good ideas. Cheers, Rusty
  12. Thanks Reece for sharing your experiences and insights. I will be looking into issues A and B to see if we can make SMS smarter for these situations.
  13. Hi, We have upcoming courses scheduled that will cover using TUFLOW and the Surface Water Modeling System (SMS). You can use SMS to quickly build TUFLOW models and to visualize TUFLOW results. SMS supports 1D/2D models (both cross-sections and pipe networks), 2D models and multiple 2D domain models. SMS can be used stand alone or as a complement to your existing GIS modeling tools. These courses will benefit those who are new to TUFLOW or experienced TUFLOW users who would like to learn more about SMS. The courses will be held in Brisbane, Australia and Belfast, Ireland. More information about these courses can be obtained from the links below. Brisbane Course - June 13-15 Belfast Course - August Any questions can be directed to our training coordinator at training@aquaveo.com. Cheers, Rusty
  14. I think I know what is happening but lack much in the way of specifics. Hopefully, someone else can fill in more details later. Some of the 2D boundary types are implemented as a 1D node with a boundary condition and connected to the 2D boundary using HX/connection polylines. You could do the same thing manually but having the 2D options makes it easier to setup. I hope this helps.
  15. Hopefully someone who knows more then me will also respond, but I do have some thoughts that may help. You are trying to build a 10 Million cell grid. This is quite large. It may call for using multiple domains just due to runtimes. Are you using the 64-bit version of TUFLOW? Regardless of how much RAM you have on your machine, if you are running a 32-bit process the most it can ever access is 4 GB. Since windows should allow you to access virtual memory after your physical memory is exhausted, in theory you should be able to build very large grids with the 64-bit version. Keep in mind that if you are forced to access virtual memory this will further degrade the performance of your model and increase runtimes. I hope that this helps. Cheers, Rusty
×
×
  • Create New...