Jump to content
TUFLOW Forum

tuflow support

Administrators
  • Content Count

    142
  • Joined

  • Last visited

Everything posted by tuflow support

  1. Q: I have been unable to input a 12D TIN (in 12da format) from 12D version 10 into TUFLOW. A: After reviewing the .12da format it would appear that when exporting ANSI code is the default format for previous 12D versions. However, UFT-16 code is the default format for all output files for V10. The code can be changed in the file box (not in the output dialog). When inputting the file name for the output file, we need to open the file box twice: one for file name and the other for file format. Alternately, if the file is opened in UltraEdit, it can be saved in ANSI/ASCII format which should overcome the issue. We will investigate updating this in future releases of TUFLOW. If you have any issues with the above, please contact support@tuflow.com. Regards TUFLOW Support Team
  2. Q: How can I generate maps for specific duration of inundation? E.g. how can I map areas where there is greater than 24 hours inundation. A: The dat_to_dat utility has a duration option. To use this you will need to specify the -dur switch. A description is: "-dur<cutoff_value>" determines the duration in hours the cutoff value is exceeded The command line for the dat_to_dat.exe would look something like: Dat_to_dat.exe -b -dur0.1 results_d.dat Once this is processed, a newly created .dat file will contain the duration at timestep 111111. You should be able to convert this into GIS format using TUFLOW to GIS. For an ESRII grid (suitable for MapInfo, QGIS, ArcGIS) the command line will look something like: TUFLOW_to_GIS.exe -asc -t111111 -2dm results.2dm results_d(dur0.1).dat Note: By default TUFLOW_to_GIS looks for the .2dm file with the same name as the .dat file. The “-2dm results.2dm” in the command line above specifies the name of the .2dm file to be used. Once this has been loaded into GIS, this can be contoured or queried. For example, if you contoured the data from 24 to 96 this would highlight the area where the duration of inundation was between 24 and 96 hours. Regards TUFLOW Support Team
  3. Q: Does ESTRY assume glass-walling when the water level exceeds the top of the cross-section? If it assumes glass walling, then what is the default value of the height of the glass wall? Can it be changed? A: Yes, ESRTY allows the flow to exceed the top of the inputted section. Cross-section hydraulic properties above the highest elevation are calculated based on the flow width remaining constant at the width of the highest elevation in the table. If the hydraulic properties are calculated from a cross-section profile, it uses the effective flow width as shown in the .eof file (it does not use the storage width) – this preserves the effect of any variation in relative roughness across the cross-section. All other hydraulic property sources use the storage width, and any relative roughness effects are ignored once the water level exceeds the highest elevation. Also note that the wetted perimeter remains constant above the highest elevation; ie. it is not increased on the vertical as the flood level rises. The height varying channel data can be found in the _1d_ta_tables_check.csv. The _xsl_check.mif and _hydroprop_check.mif contain more information on the properties at the top of the section. The current default is for the TUFLOW to terminate a simulation with a unstable error if the water level exceeds ten times the depth of the CS or NA table. To allow the water level to exceed 10 times the channel depth the .ecf command can be used: Depth Limit Factor == <value, default is 10> Note: this command does not apply to bridge channels, where the highest elevation in the section is used as the obvert of the structure. Regards TUFLOW Support Team
  4. Q: I am having problems with message error 2001 – 2d_bc layer outside 2d domain. I have checked that all my GIS layers have the same coordinate reference system (projection) and that the 2d_bc layer is in fact inside the 2d domain as set by my 2d_code layer, I can find where the problem is. A: The TUFLOW model domain is defined by the location (or origin and orientation angle) and the dimensions and not the code polygon. For example: Origin == 0,0 Orientation Angle == 0 Grid Size (X,Y) == 1000,2000 Defines a model that extents from coordinates 0,0 (X,Y) and is 1000m in the X direction and 2000m in the Y. Therefore, the maximum coordinates are 1000,2000. Even if the code layer (2d_code) extends beyond this, the area is not included in the model domain. A boundary (2d_bc) at coordinates 4000,3000 will cause this error. Importing the domain check and the messages location will help you find where / why this is occurring. See also: http://wiki.tuflow.com/index.php?title=TUFLOW_Message_2001 Regards TUFLOW Support Team
  5. Q: When I try to use the cutoff depth (-cd) option in TUFLOW_to_GIS I get an error that the file cannot be found. The command line I am using is: TUFLOW_to_GIS.exe –asc –cd<0.05> -t99999 1000ARI_d.dat A: The brackets <> indicate an option, these should not be included in the command line. The correct command line is: TUFLOW_to_GIS.exe –asc –cd0.05 -t99999 1000ARI_d.dat
  6. Q: If there is a downstream boundary condition where I set the water surface elevation below the ground level, how does it work? Will TUFLOW assume a free outfall or a normal depth? A: Where the outflow is not downstream controlled by the water level specified at the boundary, free outfall will apply. Alternatively, a 2D HQ boundary can be specified which uses Manning’s equation for a user-defined slope (essentially normal flow). We are looking into an automatic normal flow option for a future release.
  7. Q: I have a simulation in which the flow in the model continues after the end of my input boundary. The final value in the time-series is at 15 hours but the flow continues at a steady rate after this. See attached image. A: If the model run extends past the last point in the boundary condition, the last value is used. This should be set to zero if you want no further inflow into your model. If you are using the convert_to_ts1.exe to convert the data the -e0 can be used to add a zero value to the end of the time-series. Likewise, if the simulation start time is before the boundary condition data starts, the first value in the time-series will be used until the first time in the time-series.
  8. Now that the latest version of TUFLOW supports imperial units, the values of some inputs may exceed the single precision limits. As discussed here in another topic on the TUFLOW forum, when storing floating point values on a computer, the number takes a certain number of bytes to store. Single precision numbers use 4 bytes and double precision numbers use 8 bytes. The computer sets aside a certain number of bits to handle the exponent and the rest to handle the significant digits. Single precision will give about 7 digits of precision and double precision about 16. Care needs to be taken if the values in the model inputs become very large, in the hydrograph below the model had very large 1d nodal area (1d_NA) values manually assigned for modelling of a dam storage in 1D (numbers larger than 100,000,000). When running in single precision the numerical precision of the NA was impacting the flow results through the dam. The same model in double precision version of TUFLOW yields expected results. Regards TUFLOW Support Team
  9. Q: I have setup a Matrix flow channel, as per Section 4.7.4.7 of the TUFLOW manual (2010 version), but when reviewing the results, there appears to be no flow through the channel. A: The final flow matrix read by TUFLOW is outputted in the log file (.tlf) for the run, (tip: search for "M Channel Matrix"). An extract for the model provided is below: Final M Channel Matrix Data for 85 Flows: h2, h1 0.0000, 30.760, 31.260, 31.760, 32.260 30.760, 0.00, 0.00, 0.00, 0.00 31.260, 0.00, 0.00, 0.00, 0.00 31.760, 0.00, 0.00, 0.00, 0.00 32.260, 0.00, 0.00, 0.00, 0.00 32.760, 0.00, 0.00, 0.00, 0.00 33.260, 0.00, 0.00, 0.00, 0.00 33.760, 0.00, 0.00, 0.00, 0.00 As can be seen the flows are all zero, which is why there is no flow in the channel. This is occurring because the "Number Of" attribute of the channel is set to 0 (flow and area matrices are multiplied by this value). There is an error in the manual that says if a value of zero is entered then this parameter is set to one. Investigation found this this is not case. We will fix this for the next update to TUFLOW (or update the manual to clarify). In the interim if the "Number_of" attribute in the 1d_nwk GIS is set to 1, the correct values come through (extract below). Final M Channel Matrix Data for 85 Flows: h2, h1 1.0000, 30.760, 31.260, 31.760, 32.260 30.760, 0.00, 150.64, 185.22, 467.69 31.260, -150.64, 0.00, 256.80, 393.51 31.760, -185.58, -256.80, 0.00, 314.65 32.260, -467.70, -395.70, -314.65, 0.00 32.760, -503.13, -523.68, -667.23, -628.07 33.260, -984.63, -961.53, -873.00, -876.94 33.760, -1126.56, -1109.00, -1001.45, -1020.57 Regards TUFLOW support team.
  10. Q: I have a 12da file that I would like to convert to mid/mif using the '12da_to_from_mif.exe' tool. The 12da file is on GDA coordinates as I have have double checked by opening it up in 12da. However, when I run the tool by changing the file command in the .bat file to what is shown below the tool comes up with: WARNING - No MI Projection of mif header file ..\model\Header.mif found. Non-Earth Coordinates will be used. Command in .bat file is: 12da_to_from_mif.exe ..\model\road_proposed.12da Can you please assist as I would like to have mid/mif files that are on GDA coordinates. A: You need to have a file called “Header.mif” in the same folder as your .12da file. Header.mif is used to set the projection of the .mif file created, so simply copy any .mif file with the correct projection into the folder and rename as Header .mif (there is no need to have a .mid file).
  11. Q: I am interested in knowing whether or not I can run ESTRY 1D by itself. To give a little background, I have a TUFLOW model in which floodplains are modeled in 2D environment and channel in 1D. First, I set up a 2D model and then carved a 1D channel and linked it to the 2D domain along the banks of my channel. The model runs fine with 1D/2D linkage but the run time is about 9 hrs. I am interested in evaluating some alternatives within the channel for which I don’t need to see what is happening in the floodplain. So, if I could just run the 1D piece of the model that would be great. A: We used to issue an ESTRY.exe so that users could run 1D only models, however, when we moved to the new WIBU dongles ESTRY was left behind. You can download past versions, but they won’t recognise the new WIBU dongles, nor will they have new features. Our plan this year is to reinstate this capability of 1D only models but via the TUFLOW.exe (ie. if you start TUFLOW with a .ecf file it will be a 1D only run). In the meantime, the best option is to create a dummy 2D model of say 4 cells off to the side. You can leave the .tbc file blank (you have to have a .tbc file, but it can be empty), and the .tgc file would simply contain the location of the dummy grid and the commands “Set Code == 1”; “Set Zpts == 0.”; and “Set Mat == 1”. Your dummy 2D model and your production model can both access the same .ecf file and 1D input layers, so there won’t be any data duplication.
  12. Q: The .XF file generated by the new Read GRID Zpts command; how does TUFLOW know when it needs to update this? A: The XF file generated by a Read GRID Zpts or Read TIN Zpts command is different to the XF file generated by the Read RowCol (MID) Zpts command in that it contains info in regarding the origin, orientation and number of rows and columns. If the TUFLOW 2D domain is changed in terms of its origin, orientation or size, this will be incompatible with the XF file and TUFLOW will automatically regenerate the XF file from the GRID or TIN. The XF filename also includes the 2D domain's cell size and domain index, therefore, if the 2D domain's cell size changes or its order within the control files changes (if using multiple 2D domains), the XF file will be automatically be regenerated. The above should hopefully cover all eventualities. However, if you're using the same DEM for more than one model, and the models have the same cell size then there is a potential conflict if accessing the same GRID/TIN file. It is therefore recommended that the DEM GRID or TIN file is copied to a folder under each model's tuflow folder (eg. model_1\tuflow\model\dem) so that this situation does not occur. Alternatively, given that these commands aren't particularly slow, switch the writing/reading of XF files off by including XF OFF (eg. Read GRID Zpts XF OFF ==). In addition to the above, if the DEM GRID or TIN file is rexported and therefore has a save date later than the XF file, TUFLOW will automatically regenerate the XF file.
  13. For the benefit of other TUFLOW users who may be reading this. The issue has been resolved, the problem was occurring because the Ignore attribute on two of the culverts was set to true "T". These culverts were not being used in the simulation and the associated SX points were generating the error. This issue can be resolved by setting the ignore attribute to "F" (false) - the culverts are then included in model or by removing the SX connections in which case no culverts are modelled where the Ignore flag is set to "T". Regards TUFLOW Support Team
  14. Q: What are the units for the Stream Power output of TUFLOW (in SMS 11)? Are they Watts, Watts per square metre, watts per grid cell size (i.e. if I have a 10m grid cell, are they in "watts per 100m2" (i.e. 10m x 10m)? Any help would be much appreciated. A: Good question, the manual is a bit ambiguous in this area (this will be fixed for the next version). The bed shear stress is calculated with the formula shown in the attachment and represents the bed shear. The units of the bed shear are: kg.m-1.s-2 The stream power (Appendix page A-13 of the 2010 TUFLOW manual) is calculated as Bed Shear Stress times the Velocity Magnitude and has the units of kg.s-3 or Watts per m2.
  15. Hi Blake, This response was provided by a TUFLOW user to our forum email address:
  16. It would appear that with the upgrade of the forum software some users have had their notification options change. To review and modify your notification options, please use the My Settings option in the dropbox at the top right of screen (see 1st attached image). In the My Settings page, navigate to the Notification Options section (see 2nd image). This allows you to tailor the notifications settings to suit. Regards TUFLOW Support Team
  17. An alternate error message is: The Application has failed to start because its side-by-side configuration in incorrect. Please see the application event log or use the command-line sxstrace.exe tool for more detail. Installation of Microsoft's Visual C++ 2008 Redistributable Package using the process detailed above resolved the issue. Regards TUFLOW Support Team
  18. Hi Paul, I have changed a setting in your notification settings to test if this fixes the issue of you not receiving email notification. I received notification about a response to this topic, so I suspect that it must be a setting issue. I could not see any failed delivery notifictions to your email address. Cheers TUFLOW Support Team
  19. Q: I have a model that runs fine in 2009-07 but in 2010-10 it gives the message: ERROR 1301 - No ESTRY nodes found. A: This occurred as a consequence of TUFLOW interpreting the 1D input to be in the old fixed field format. Because in the 2010-10 release 1D commands can now be embedded in the .tcf file, TUFLOW 2010-10 starts with the .tcf file when looking for 1D commands (2009-07 goes straight to the .ecf file). If there are four or more blank lines at the start of the .tcf file (as was the case here), TUFLOW 2010-10 (wrongly) assumes the 1D commands are in fixed field format (long story!). The work around is to not have four or more blank lines at the start of the .tcf file if using Build 2010-10-AA to 2010-10-AE. For 2010-10-AF or later this bug has been fixed. Note 2010-10-AF has not yet been uploaded at the time of this post. NOTE: The above also occurs for all releases and builds of TUFLOW prior to the 2010-10 release if the .ecf file has four or more blank lines at the start.
  20. Q: I am using an automatically created HQ boundary, however during the simulation I repeatedly get the following warning: WARNING 2365 - Exceeded top of HQ curve. Water Level = <value> I am using separate HQ boundaries for the main channel and the adjacent floodplain. The water level will exceed the top of bank of the main channel, is there a way to extend this upward? A: Yes, you can use the 2d_bc “d” attribute to set the maximum depth of the automatically generated HQ curve (note d is a depth, not an elevation). In the example below there are two HQ curves (check in the _2d_bc_tables_check.csv) one without a d attribute, and the other with a d (depth) attribute of 8.
  21. Windows 7 and relative paths We had a query through the support email with a simulation using relative paths starting in Windows XP, but not in Windows 7. An examination of the log files showed that three dots ...\ was used in the relative paths eg: BC Control File == ...\model\M01_5m_003.tbc Rather than: BC Control File == ..\model\M01_5m_003.tbc It would appear that Windows 7 does not recognise the ...\ syntax whereas Windows XP did. In XP the ...\ syntax performs the same as the ..\ syntax and navigates up one folder. The link below suggest that in older versions of Windows (Windows 95 and 98) ...\ would go up two folder levels. In this case ..\..\ can be used instead in Windows 7. Link with info on CD (change directory) command: http://www.computerhope.com/cdhlp.htm Regards, TUFLOW Support
  22. Answer: During investigation of the issue, a DEM was interpolated from the zpt check file (in vertical mapper), this showed that the banding occurs in the input elevations to TUFLOW. Image one and two show the banding in the DEM of the zpt check file. This banding came from the interpolation of the original dataset to a DEM. To smooth the elevations in TUFLOW the .tgc command Interpolate ZUVH ALL was added (after the zpts were read in). This command interpolates the cell side (ZU and ZV) and cell corner (ZH) elevations from the cell centres (ZC). The depth results for the interpolated zpt simulation is shown.
  23. Question: I have a direct rainfall model that has a strange banding in the depth results, the model mass error is acceptable, what could be causing this? Attached is an image of the depth results.
  24. Q: I have a model with approximately 6000 channels, that is requesting 5.7Gb of memory, however a previous model with approximately 2500 channels requires only 44Mb. Is there a threshold number of channels above which the memory request increases dramatically? A: 5Gb of memory for the 1D model is definitely excessive! This is happening because TUFLOW is storing the 1D results at every 1D timestep. The 1D control file (.ecf) command Output Interval == <time in seconds> can be used to set the interval. If no output interval is specified in the 1D, the default is to output at the same times as the 2D time-series output (plot output interval). This is the default as it allows the 2D time-series output and the 1D results to be written into the same _TS.mif file for plotting in MapInfo. This can be overwritten with the .ecf command: Output Times Same as 2D == OFF. The time-series output interval in the 2D is controlled with the .tcf command: Time Series Output Interval. When no Output interval is set in the 1D control file or Time Series Output Interval in the 2D, TUFLOW will store the 1d results at every timestep, hence the very large memory requirements. Furthermore, the .tcf command Write PO Online == ON will cause TUFLOW to output the 2D time-series and 1D results at the same times as the 2D map output. Without this command, these results will be stored to the end of simulation and then written. It will slightly increase the runtime as the files are opened and closed more often (probably very minor change in runtime), however this enables you to review results while the model processes (particularly good for calibration). This does not affect the memory requirements of the 1D.
  25. Q: I have a Create TIN Zpts WRITE TIN command, that reads in contour line data to generate a TIN file. The data that I used works in the 2009 version of TUFLOW but with the 2010 release I get the following error: ERROR 2232 - Failed to complete triangulation. Check region perimeter does not cross or snap onto itself. I have checked the region doesn't snap onto itself, does the 2010 update have problems with the triangulation code? A: After reviewing the inputs files we concluded that the problem seems to lie with the new defaults in the 2010 release which control the angles allowed when creating a TIN (to prevent long thin triangles). These defaults are stricter about the angles allowed when creating triangles compared to the 2009 version. The .tgc command: TIN Angles == <point angle>, <edge angle> can be used in the 2010 version to allow the user more control over the triangulation process. By adding the TIN angles command above the Create TIN Zpts command, the model starts in the 2010 version. For more info in the command see the 2010 TUFLOW manual, this command can be used multiple times to use different angles for different inputs. The command used was: TIN Angles == 95,1 ! prior to 2010 this was 95,0 (but 0 is not recommended) Create TIN Zpts WRITE TIN == 2d_ztin_WBM05.mif However, when I look at the TINs created in SMS it would appear that the large dmax value used (dmax = 10) is causing some undesirable triangulations. The dmax value sets the maximum distance between vertices, if the space between vertices is greater than this additional vertices are inserted. For example if the vertices are 55m apart additional vertices will be inserted so that 6 segments of ~9.17m are used. With the detailed contour lines that are used the value of 10 is probably too large (the default is half the cell size, or 1m for the case of a 2m cell size such as this). TUFLOW then creates a triangulation from the vertices. The following images will hopefully clarify things. In the first the dmax is 10, the black lines are the input contour lines (from the 2d_ztin), the red squares are the vertices used for triangulation, the triangles are shown in red (one is selected – see black triangle). As can be seen the highlighted triangle is actually crossing the contour line, which isn’t desirable. The second image below has a dmax of 1m set, many more vertices are inserted and the triangulation looks much nicer (to me at least). With more vertices comes more processing time, however this should only occur the first time the input is used as the 2010 version of TUFLOW outputs a binary (xf) version. However there are some contour lines that are less than 0.1m apart, so even with a 1m dmax the triangulation may cross these contours. This occurs in the contours at the bottom right of the images below (distance between contour distance of ~0.05m. To accurately represent this in the TIN (the model has a 2m resolution, so it probably won’t make a big difference), I tried using MapInfo to create vertices with spacing of 0.1m. I extracted the coordinates of these and imported to SMS as a scatter data set. I used SMS to do the triangulation as SMS is very quick at doing this (other GIS packages such as 12D are also powerful at this). The third image shows a shaded view of the TIN created in SMS from 0.1m spaced vertices, this TIN preserves the steep gradient areas. This TIN can then be read into TUFLOW using the Read TIN Zpts == <tin file> command.
×
×
  • Create New...