Tuflow - User contributions [en]

1D Pumps

2026-04-21T03:32:37Z

ShuangGao: /* 2D-2D Configuration */

=Introduction=
This post provides a modelling example for a 1D pump using a pump curve. For this example we will set up a pump in two common situations (2D-2D & 1D-2D).

=Pump Attributes=
A pump needs to first be digitised in a 1d_nwke layer. The direction of the polyline must go from inlet to outlet as a pump is unidirectional. The attributes required for a pump in your 1d_nwk layer can be found in the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]. 
In the 1d_nwk layer, the following attributes are required: 
#ID = ID of the pump channel. 
#Type = "P" or "PO". 
#US_Invert = Intake elevation of the pump. 
#DS_Invert = Outlet elevation of the receptor. 
#Inlet_Type = Used to specify the pump curve in the Depth Discharge database. 
#Width_or_D = Diameter of the pump’s outlet pipe/hose. 
#Number_of = Number of (identical) pumps.
 
[[File:1d_nwk_pump_pipe.PNG|border|300px]] 

=2D-2D Configuration=
Because pumps are zero length channels they do not create automatic nodes at the upstream and downstream end. If you ran the model with just a pump polyline and SX connection you will get [[TUFLOW_Message_1353| ERROR 1353]]. To remove this error, the most efficient schematisation is to digitise a 1d_nwk 'NODE' at the upstream and downstream end of the pump (no need for a separate 2d_bc SX layer). Unlike NODEs connected to pipes and channels, NODEs connected to zero length pump required the following attributes: 
#Type = "NODE". 
#Len_or_NA = The 'NODE' requires a nominal storage amount. This can be estimated from the pipe length and diameter attached to the pump. 
#US_Invert = The upper elevation of the automatically created NA table. Make sure these values are set higher than the expected water levels at the intake and outlet of the pump. 
#DS_Invert = The bottom elevation of the pump nodes. As the pump do not create automatic nodes, the bottom elevation of the pump nodes must be specified. Note that this does not change the intake or outlet elevations of the pump, but only sets the bottom elevation of the nodes for storing water. 
#Conn_1D_2D = Sets to "SX" to connect the 1D pump with the 2D domain. Without the SX connection, water will build up within the node and cause an instability. 

See the example below. 

[[File:1d_nwk_pump_SX_node.png|border|600px]] 

A simple 2D-2D pump configuration will look like the below schematisation. 

[[File:Pump_schematic.PNG|border|600px]] 

=1D-2D Configuration=
Connecting a pump from a 1d network to the 2d domain or vice versa is similar to the configuration above, the only difference is that the connection with a 1d structure does not require a 1d nwk ‘Node’. A storage chamber in the 1d network can also be modelled using a 1d_na node with an elevation vs area .csv assigned to the node. 
[[File:1D-2D_pump_schematisation.PNG|border|600px]] 

=Estry Control File Setup=
Within the *.ecf the following commands and files are required to run a pump with no logical controls: 
<tt>Read GIS Network</tt> <tt>==</tt><tt>..\model\mi\1d_nwke_xxxxx.MIF</tt> 
<tt>Depth Discharge Database</tt> <tt>==</tt><tt>..\bc_dbase\xxxxx.csv</tt> 
If you do not specify a Depth-Discharge database then you will be faced with [[TUFLOW Message 1118 | ERROR 1118]]. 

=TUFLOW Operating Control File (.TOC)=
For guidance on setting up the operating controls for pumps, refer to the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]. 

.ecf command required: 

<tt>Read Operating Controls File</tt> <tt>==</tt><tt> xxxxx.toc</tt> 

=Depth Discharge Database=
The depth discharge database is set up in the same way as a pit inlet database (refer to the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]). Each pump ‘Inlet_type’ must reference a name within the depth discharge database, otherwise [[TUFLOW_Message_1118 | ERROR 1118]] - Could not find pit inlet type ",a," in the pit inlet database. The ‘Area (m2)’ column is the area of the pump offtake and ‘Width (m)’ column is the width of the pump offtake. Without information in the Area(m2) or Width(m) columns in the depth discharge database [[TUFLOW Message 1092|ERROR 1092]] and [[TUFLOW Message 1093|ERROR 1093]] will appear. 

==Pump Curve==
The performance of pumps is a function of suction head at the inlet and the level of the discharge location. The resultant total head between the water level at the inlet and outlet is what determines the flow rate through the pump. If the suction level is low the pump will need to provide more energy in the form of pressure to maintain the water elevation at the outlet, the opposite is also true if the water depth at the pump inlet is high. 

[[File:pump_fundamentals_total_head.jpg|thumb|none|500px|www.pumpfundamentals.com]]

That being the case it is important to consider what total head is required to achieve the modelling objectives and what flow rates you may require. Once you have an idea on any limits in total head you can start to research an appropriate pump and then extract the performance curve that is often incorporated as part of the technical specifications. An example is shown below. 

[[File:Manufacturer pump curve.JPG|border|600px]]

==Creating a TUFLOW pump curve==
The setup of the Depth Discharge database for a pump curve is similar to reading in inflow hydrographs, hyetographs etc, that is; a source .csv, and the two corresponding headings within the 3rd and 4th column. 

[[File:depth-discharge_pump.PNG |border|500px]] 

Once you have your manufacturer curve for your given pump it is now necessary to create the curve .csv for TUFLOW to read in. The manufacturer specifications will need to be translated into a total head vs pump rate chart. Although reading in the depth discharge database is the same process as other boundary conditions within TUFLOW, the curve itself is fundamentally different as you no longer need to start the csv file with 0,0. If the curve did start at 0,0 this would not make sense because at a total head difference of 0m the pump should effectively be operating at peak performance so the flow rate would be greater than 0 m3. The image below shows a csv file for the pump performance curve given in the previous section. 
 
[[File:Pump_curve_csv_example.png|border|600px]] 

==Using a Pump Curve in a TUFLOW Operating Control (TOC) File==

With the pump curve defined in the depth-discharge database it can either be specified within the pump 1d_nwk fields in the inlet_type field, for non-operational pumps, or it can be defined with the TOC file, for operational pumps. When defining with a TOC file, the pump curve is defined at the top of the structure control definition block and then the subsequent rules can turn the pump on/off. See the below TOC structure control definition for an example. In this case, the pump curve is used when the pump switches on once upstream water levels reach 2.75m AD. The pump curve is then used until the upstream water levels are reduced to 2.25m AD at which point the pump is switched off.

<tt>Define Pump Control</tt><tt> ==</tt> Pump_1
<tt>Pump Capacity</tt><tt> ==</tt> pump_1
HU <tt>==</tt> H1D Pump1.1
<tt>If </tt>HU <= 2.25
<tt>Pump Operation</tt><tt> ==</tt> Off
<tt>Else if</tt> HU > 2.25 <tt>AND</tt> HU < 2.75
<tt>Pump operation</tt><tt> ==</tt> <tt> No Change </tt>
<tt>Else if</tt> HU <tt>>=</tt> 2.75
<tt>Pump Operation</tt><tt> ==</tt> On
<tt>End if</tt>
<tt>End define</tt>

=1D Result File=
Although strictly not a check file, the operation of the pump can be confirmed by opening the *_1d_O.csv which is found within the csv folder where the results are written. The *_1d_O.csv monitors the operation of structures, this file can be quite useful in checking how the structure is performing with the given .toc file and GIS inputs. 
{| align="center" class="wikitable" width="75%"
! style="background-color:#005581; font-weight:bold; color:white;"| Filename prefix / suffix
! style="background-color:#005581; font-weight:bold; color:white;" width=75%| Brief Description
|-
| [[Pump_Results_1d_O | _1d_O.csv]]|| This csv displays the status of the pump, whether that is closed or fully open, results for any logic parameter specified in the TOC file and the flow through the pump if it is in operation.
|}
 
Any further questions please email TUFLOW support: [mailto:support@tuflow.com?Subject=TUFLOW%201D%20pumps%20help support@tuflow.com]
 
{{Tips Navigation
|uplink=[[ TUFLOW 1D Channels and Hydraulic Structures | Back to 1D Channels and Hydraulic Structures]]
}}

1D Pumps

2026-04-21T03:32:22Z

ShuangGao: /* 2D-2D Configuration */

=Introduction=
This post provides a modelling example for a 1D pump using a pump curve. For this example we will set up a pump in two common situations (2D-2D & 1D-2D).

=Pump Attributes=
A pump needs to first be digitised in a 1d_nwke layer. The direction of the polyline must go from inlet to outlet as a pump is unidirectional. The attributes required for a pump in your 1d_nwk layer can be found in the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]. 
In the 1d_nwk layer, the following attributes are required: 
#ID = ID of the pump channel. 
#Type = "P" or "PO". 
#US_Invert = Intake elevation of the pump. 
#DS_Invert = Outlet elevation of the receptor. 
#Inlet_Type = Used to specify the pump curve in the Depth Discharge database. 
#Width_or_D = Diameter of the pump’s outlet pipe/hose. 
#Number_of = Number of (identical) pumps.
 
[[File:1d_nwk_pump_pipe.PNG|border|300px]] 

=2D-2D Configuration=
Because pumps are zero length channels they do not create automatic nodes at the upstream and downstream end. If you ran the model with just a pump polyline and SX connection you will get [[TUFLOW_Message_1353| ERROR 1353]]. To remove this error, the most efficient schematisation is to digitise a 1d_nwk 'NODE' at the upstream and downstream end of the pump (no need for a separate 2d_bc SX layer). Unlike NODEs connected to pipes and channels, NODEs connected to zero length pump required the following attributes: 
#Type = "NODE". 
#Len_or_NA = The 'NODE' requires a nominal storage amount. This can be estimated from the pipe length and diameter attached to the pump. 
#US_Invert = The upper elevation of the automatically created NA table. Make sure these values are set higher than the expected water levels at the intake and outlet of the pump. 
#DS_Invert = The bottom elevation of the pump nodes. As the pump do not create automatic nodes, the bottom elevation of the pump nodes must be specified. Note that this does not change the intake or outlet elevations of the pump, but only sets the bottom elevation of the nodes for storing water. 
#Conn_1D_2D = Sets to "SX" to connect the 1D pump with the 2D domain. Without the SX connection, water will build up within the node and cause an instability. 

See the example below. 

[[File:1d_nwk_pump_SX_node.png|border|300px]] 

A simple 2D-2D pump configuration will look like the below schematisation. 

[[File:Pump_schematic.PNG|border|600px]] 

=1D-2D Configuration=
Connecting a pump from a 1d network to the 2d domain or vice versa is similar to the configuration above, the only difference is that the connection with a 1d structure does not require a 1d nwk ‘Node’. A storage chamber in the 1d network can also be modelled using a 1d_na node with an elevation vs area .csv assigned to the node. 
[[File:1D-2D_pump_schematisation.PNG|border|600px]] 

=Estry Control File Setup=
Within the *.ecf the following commands and files are required to run a pump with no logical controls: 
<tt>Read GIS Network</tt> <tt>==</tt><tt>..\model\mi\1d_nwke_xxxxx.MIF</tt> 
<tt>Depth Discharge Database</tt> <tt>==</tt><tt>..\bc_dbase\xxxxx.csv</tt> 
If you do not specify a Depth-Discharge database then you will be faced with [[TUFLOW Message 1118 | ERROR 1118]]. 

=TUFLOW Operating Control File (.TOC)=
For guidance on setting up the operating controls for pumps, refer to the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]. 

.ecf command required: 

<tt>Read Operating Controls File</tt> <tt>==</tt><tt> xxxxx.toc</tt> 

=Depth Discharge Database=
The depth discharge database is set up in the same way as a pit inlet database (refer to the [https://docs.tuflow.com/classic-hpc/manual/latest/ TUFLOW Manual]). Each pump ‘Inlet_type’ must reference a name within the depth discharge database, otherwise [[TUFLOW_Message_1118 | ERROR 1118]] - Could not find pit inlet type ",a," in the pit inlet database. The ‘Area (m2)’ column is the area of the pump offtake and ‘Width (m)’ column is the width of the pump offtake. Without information in the Area(m2) or Width(m) columns in the depth discharge database [[TUFLOW Message 1092|ERROR 1092]] and [[TUFLOW Message 1093|ERROR 1093]] will appear. 

==Pump Curve==
The performance of pumps is a function of suction head at the inlet and the level of the discharge location. The resultant total head between the water level at the inlet and outlet is what determines the flow rate through the pump. If the suction level is low the pump will need to provide more energy in the form of pressure to maintain the water elevation at the outlet, the opposite is also true if the water depth at the pump inlet is high. 

[[File:pump_fundamentals_total_head.jpg|thumb|none|500px|www.pumpfundamentals.com]]

That being the case it is important to consider what total head is required to achieve the modelling objectives and what flow rates you may require. Once you have an idea on any limits in total head you can start to research an appropriate pump and then extract the performance curve that is often incorporated as part of the technical specifications. An example is shown below. 

[[File:Manufacturer pump curve.JPG|border|600px]]

==Creating a TUFLOW pump curve==
The setup of the Depth Discharge database for a pump curve is similar to reading in inflow hydrographs, hyetographs etc, that is; a source .csv, and the two corresponding headings within the 3rd and 4th column. 

[[File:depth-discharge_pump.PNG |border|500px]] 

Once you have your manufacturer curve for your given pump it is now necessary to create the curve .csv for TUFLOW to read in. The manufacturer specifications will need to be translated into a total head vs pump rate chart. Although reading in the depth discharge database is the same process as other boundary conditions within TUFLOW, the curve itself is fundamentally different as you no longer need to start the csv file with 0,0. If the curve did start at 0,0 this would not make sense because at a total head difference of 0m the pump should effectively be operating at peak performance so the flow rate would be greater than 0 m3. The image below shows a csv file for the pump performance curve given in the previous section. 
 
[[File:Pump_curve_csv_example.png|border|600px]] 

==Using a Pump Curve in a TUFLOW Operating Control (TOC) File==

With the pump curve defined in the depth-discharge database it can either be specified within the pump 1d_nwk fields in the inlet_type field, for non-operational pumps, or it can be defined with the TOC file, for operational pumps. When defining with a TOC file, the pump curve is defined at the top of the structure control definition block and then the subsequent rules can turn the pump on/off. See the below TOC structure control definition for an example. In this case, the pump curve is used when the pump switches on once upstream water levels reach 2.75m AD. The pump curve is then used until the upstream water levels are reduced to 2.25m AD at which point the pump is switched off.

<tt>Define Pump Control</tt><tt> ==</tt> Pump_1
<tt>Pump Capacity</tt><tt> ==</tt> pump_1
HU <tt>==</tt> H1D Pump1.1
<tt>If </tt>HU <= 2.25
<tt>Pump Operation</tt><tt> ==</tt> Off
<tt>Else if</tt> HU > 2.25 <tt>AND</tt> HU < 2.75
<tt>Pump operation</tt><tt> ==</tt> <tt> No Change </tt>
<tt>Else if</tt> HU <tt>>=</tt> 2.75
<tt>Pump Operation</tt><tt> ==</tt> On
<tt>End if</tt>
<tt>End define</tt>

=1D Result File=
Although strictly not a check file, the operation of the pump can be confirmed by opening the *_1d_O.csv which is found within the csv folder where the results are written. The *_1d_O.csv monitors the operation of structures, this file can be quite useful in checking how the structure is performing with the given .toc file and GIS inputs. 
{| align="center" class="wikitable" width="75%"
! style="background-color:#005581; font-weight:bold; color:white;"| Filename prefix / suffix
! style="background-color:#005581; font-weight:bold; color:white;" width=75%| Brief Description
|-
| [[Pump_Results_1d_O | _1d_O.csv]]|| This csv displays the status of the pump, whether that is closed or fully open, results for any logic parameter specified in the TOC file and the flow through the pump if it is in operation.
|}
 
Any further questions please email TUFLOW support: [mailto:support@tuflow.com?Subject=TUFLOW%201D%20pumps%20help support@tuflow.com]
 
{{Tips Navigation
|uplink=[[ TUFLOW 1D Channels and Hydraulic Structures | Back to 1D Channels and Hydraulic Structures]]
}}

2025-08-27T01:09:13Z

ShuangGao: Created page with "test"

test

File:HR Z Use Face Elevations ON.PNG

2025-08-21T01:59:22Z

ShuangGao: ShuangGao uploaded a new version of File:HR Z Use Face Elevations ON.PNG

File:HR Z Use Face Elevations ON.PNG

2025-08-21T01:51:40Z

ShuangGao: ShuangGao uploaded a new version of File:HR Z Use Face Elevations ON.PNG

2025-08-21T01:25:35Z

ShuangGao: /* Showing Face Elevation in HR_Z Check File */

=Introduction=
High Resolution (HR) Grid/Raster Map Outputs was first introduced in TUFLOW 2020-10-AB release. When an SGS model uses “<tt>SGS Approach </tt> <tt>== </tt> <tt>Method C</tt>”, the sampled elevations are retained at the end of the geometry processing. These sub-grid elevations include topography modifiers such as breaklines, and they allow a high-resolution elevation check file to be written and used for high resolution depth map outputs. Currently, ASC, FLT and TIF raster formats are supported.

For the HR output the water level at each HR output location is interpolated from the computed 2D water levels. The depth is the difference between the interpolated water level and the sub-grid elevation. This differs from the standard depth output which calculates the depths at cell centres and corners first, then interpolate the depths to the standard output grid locations (default is half the cell size). The difference in the interpolation process is illustrated below.

[[File:Sgs_std_output.png|360px]] [[File:Sgs_hr_output.png|360px]]

The advantage of the HR output is that it can retain the sub-grid detail of the terrain information even at coarse cell size. As compared in the example below, the HR depth output shows clear flow path even at 100m grid. However, nicer depth output does not mean reliable hydraulic result. In fact, the 100m cell size is too coarse to produce reliable/converged hydraulic results for any real-world flood model. It's is strongly recommended that the model convergence/benchmarking tests must be conducted based on the standard map output, and the HR output should be used for presentation purpose only.

[[File:Sgs_std_vs_hr_output.png|720px]]

=Enable High Resolution Output=
To produce high-resolution output, add “HRASC”, “HRFLT” or “HRTIF” to the “<tt>Map Output Format </tt> <tt>==</tt>” command, for example:

<tt>Map Output Format </tt> <tt>== </tt> <tt>XMDF TIF HRTIF</tt>

At the moment, only water level and depth are supported as the High Resolution Output formats, i.e.:

<tt>HRTIF Map Output Data Types </tt> <tt>== </tt> <tt>h d</tt>

The map output data types and output interval can be defined separately for the High Resolution Output formats. For example:

<tt>Map Output Interval </tt> <tt>== </tt> <tt>600</tt> <tt>! Sets all map output interval to 600 seconds</tt>

<tt>HRTIF Map Output Interval </tt> <tt>== </tt> <tt>3600</tt> <tt>! Sets HR map output interval to 3600 seconds</tt>

The output resolution for high resolution grid output is defined using the following .tcf command:

<tt>HR Grid Output Cell Size </tt> <tt>== </tt> <tt>0.5</tt> <tt>! m or ft</tt>

If this command is omitted, the default output resolution is set as the SGS sampling distance.

HR outputs are supported by HPC and Quadtree Output Zones. Below is an example of tcf commands to set up HR output in a output zone:
<tt>Model Output Zones</tt> <tt>==</tt> <tt>ZoneA</tt> 
<tt>Define Output Zone</tt> <tt>==</tt> <tt>ZoneA</tt>
<tt>Read GIS Output Zone</tt> <tt>==</tt> <tt>..\model\gis\2d_oz_S05_002_R.shp</tt>
<tt>Write Check Files INCLUDE</tt> <tt>==</tt> <tt>dem_Z</tt> <tt>! includes only dem_Z check file for this output zone</tt>
<tt>Map Output Format</tt> <tt>==</tt> <tt>HRTIF</tt>
<tt>HRFLT Map Output Data Types</tt> <tt>==</tt> <tt>h d</tt>
<tt>HRFLT Map Output Interval</tt> <tt>==</tt> <tt>3600</tt>
<tt>HR Grid Output Cell Size</tt> <tt>==</tt> <tt>0.5</tt>
<tt>End Define</tt>

If HR output is defined in the Map Output Format, a high-resolution DEM_Z check file (DEM_Z_HR) is produced instead of the DEM_Zmin check file.

=Water Level Interpolation Method=
Similar to the standard output, the HR output needs to interpolate cell centre water levels to cell corners. However, the interpolation methods for the standard output (<tt>Map Output Corner Interpolation </tt> <tt>== </tt> <tt>Method C</tt>) can produce "bumpy" HR water level output in direct rainfall models with steep terrain, as the water level is linearly interpolated from the cell centres/corners, while the change of sub-grid elevations may not be linear. This often happens between fully wet cells and sheet flow cells, as illustrated below.

[[File:corner_h_intp_output.png|480px]]

In the HR water level output, these locations open have high water level with triangular shape.

[[File:HR_Intp_A.png|480px]] 
'''HR Interpolation Approach == Method A''' (Gray lines show the triangulation TINs)

Two additional methods have been added for the HR corner water level interpolation:

<tt>HR Interpolation Approach </tt> <tt>== </tt> <tt>Method A | Method B | {Method C}</tt>

* Method A applies the same water level interpolation method used for the standard output.
* Method B performs sheet flow checks at cell faces and ignores the water level from the upstream cell.
* Method C is the default option that applies the same sheet flow checks as the Method B. In addition, it also uses the number of wet SGS sampled points as a weighting that biases non-sheet flow cells that further improves the mapping to in-stream water levels.

The two images below present the high-resolution water level output at the same location, but with HR Interpolation Approach == Method B and Method C. As can be seen, the water level along the narrow stream is “smoother”.

[[File:HR_Intp_B.png|480px]] [[File:HR_Intp_C.png|480px]] 
'''HR Interpolation Approach == Method B (left) and Method C (right)'''

Note that when these two methods are applied, the interpolated corner water level is biased to the non-sheet flow cells, and consequently, sheet flow cells may appear as “dry” cells. The improved approach that takes into the account of the sheet flow water level is currently under development.

Whilst Methods B and C can substantially improve the water surface mapping of SGS models using direct rainfall (rain-on-grid), there will always be inaccuracies with mapping at a higher resolution than the 2D cell resolution due to interpolation and extrapolation. Regardless of the software, the greater the ratio of 2D cell size to the high-resolution DEM cell size, the greater the potential for mapping inaccuracies. Should better mapping accuracy be required, reducing the 2D cell size to compute the spatial variation in water surface and velocities more accurately is, by far, the best course of action.

Please also note that it is not necessary to use these options for non-rainfall on grid models.

=Interpolation near Thin Breakline =
Thin breaklines are often used to define hydraulic controls, such as levees and road embankments. These controls often experience upstream controlled weir flow, where the water levels on the two sides of the breakline can be considerably different. Retaining just one water level (upstream, downstream or average) at output points along the breakline may result in unsatisfactory water level map output as illustrated in the two figures below. This issue becomes more profound the larger the 2D cell size.

[[File:breakline_h_intp_output.png|480px]] 
[[File:HR_thin_brk_off.png|480px]] 
'''HR Thin Z Line Output Adjustment == OFF'''

A new HR output feature has been introduces to improve the water level interpolation along thin breakline.
Any further questions please email TUFLOW support: [mailto:support@tuflow.com?Subject=TUFLOW%201D%20gates%20help support@tuflow.com]

<tt>HR Thin Z Line Output Adjustment </tt> <tt>== </tt> <tt>OFF | {ON CELL SIDES} | ON ALIGNMENT</tt>

* “OFF”: does not apply any water level adjustment.
* “ON CELL SIDES” (default): two water levels, rather than one, are stored at cell corners selected by thin breaklines. These water levels are used to interpolate HR water levels on either side of thin breaklines. This option produces a sudden drop in water level across the breakline as illustrated in the figure below, noting the sudden drop follows the 2D cell faces where the breakline is applied.
* “ON ALIGNMENT”: As an extension to the “ON CELL SIDES” option, this method moves and aligns the 2D cell corner vertices to the original alignment of the breakline for producing HR water level and depth. The arrows in the figure below show how vertices of the TIN are moved to align with the breakline. This can produce substantially improved high-resolution mapping along thin breaklines with substantial drops in water levels, especially for larger 2D cell sizes. Note that the cell corners are NOT shifted for the hydraulic calculations.

[[File:HR_thin_brk_on.png|480px]] [[File:HR_thin_brk_on_adj.png|480px]] 
'''HR Thin Z Line Output Adjustment == ON CELL SIDES (left) and ON ALIGNMENT (right)'''

Note the this feature (at present) only works for thin breaklines in Z Shape (2d_zsh) layers. Thick/wide breaklines raises more than 1 rows of cell centres and 2 rows of cell faces, and thus do not experience similar HR interpolation issue.

=Showing Face Elevation in HR_Z Check File =
When modelling breaklines in TUFLOW, thin breaklines modify the cell face elevations but do not modify the cell storages (see Section 7.2.6.2). When outputting the high-resolution outputs, the user can set whether the cell face elevations are included using the following .tcf command. The default is ON, to use face elevations to indicate whether water can flow through the face or not.

<tt>HR Grid Output Use Face Elevations </tt> <tt>== </tt> <tt>{ON} | OFF</tt>

[[File:HR_Z_Use_Face_Elevations_ON.PNG|480px]] [[File:HR_Z_Use_Face_Elevations_OFF.PNG|480px]] 
'''HR Grid Output Use Face Elevations == ON (left) and OFF (right)'''

{{Tips Navigation
|uplink=[[ TUFLOW 1D Channels and Hydraulic Structures | Back to 1D Channels and Hydraulic Structures]]
}}

File:HR Z Use Face Elevations OFF.PNG

2025-08-21T01:24:17Z

ShuangGao:

2025-08-11T05:55:57Z

ShuangGao: /* What are the limitations of explicitly modelling bridge piers in TUFLOW? */

= 2D Structure Modelling Theory =
These webinars by Bill Syme and Greg Collecutt (the TUFLOW Developers) discus the theory behind the energy losses and affluxes modelling associated with hydraulic structures.

*[https://www.tuflow.com/library/webinars/#structures Webinar Link: Modelling Energy Losses at Structures] 
*[https://www.tuflow.com/library/webinars/#nov2022_hydraulic_modelling_bridge Webinar Link: 1D, 2D & 3D Hydraulic Modelling of Bridges]
 

= 2D Bridge Modelling in TUFLOW - Overview =
The TUFLOW 2D solution explicitly predicts the majority of “macro” losses due to the expansion and contraction of water through a constriction, or around a bend, provided the resolution of the grid is sufficiently fine ([https://www.tuflow.com/Download/Publications/Flow%20Through%20an%20Abrupt%20Constriction%20-%202D%20Hydrodynamic%20Performance%20and%20Influence%20of%20Spatial%20Resolution,%20Barton,%202001.pdf Barton, 2001]; [https://www.tuflow.com/Download/Publications/Modelling%20of%20Bends%20and%20Hydraulic%20Structures%20in%20a%202D%20Scheme,%20Syme,%202001.pdf Syme, 2001]; [https://www.tuflow.com/Download/Technical_Memos/Modelling%20Bridge%20Piers%20in%202D%20using%20TUFLOW.pdf Ryan, 2013]). Where the 2D model is not of fine enough resolution to simulate the “micro” losses (e.g. from bridge piers, vena contracta, losses in the vertical (3rd) dimension), additional form loss coefficients and/or modifications to the cells widths and flow height need to be added.
==Contraction/Expansion Losses (“Macro” Losses)==
Loss of energy is caused by the flow contraction during the expansion of water after the vena-contracta inside a bridge section and the flow expansion downstream a bridge. As discussed above, this type of "macro" losses can be explicitly resolved by the TUFLOW 2D solver, provided that a proper turbulence model and mesh size are used (see the [https://www.tuflow.com/library/webinars/#structures webinar on Energy Losses at Structures] above). Below is an example of the 2D modelling of flow contraction/expansion at a pair of bridge abutments.
 
[[File:FC_Velocity_Example.PNG|600px]] 

==Pier Losses==
Piers are usually smaller than the 2D cell size in real-world flood models. Although flexible mesh solver or quadtree refinement can be applied to reduce the local cell size around the pier, it also comes with an expensive computational cost that could significantly increase the simulation time. More practically, the backwater effect of piers can be modelled as sub-grid form losses.

Pier form loss coefficients can be derived from information in publications such as [https://www.fhwa.dot.gov/engineering/hydraulics/library_arc.cfm?pub_number=1&id=5 ''Hydraulics of Bridge Waterways'' (Bradly, 1978)] or [https://austroads.com.au/publications/bridges/agbt08 ''Guide to Bridge Technology Part 8: Hydraulic Design of Waterway Structures'' (AUSTROADS, 2019)]. Energy loss estimated from bridge piers or other obstructions, vertical or horizontal, that do not cause upstream controlled flow regimes like pressure flow, are dependent on the ratio of the obstruction's area perpendicular to the flow direction to the gross flow area of the bridge opening, the shape of the piers or obstruction, and the angularity of the piers/obstruction to the flow direction. For example, using Hydraulics of Bridge Waterways (Bradly, 1978) the approach is:
<ol>
<li>Calculate the ratio of the water area occupied by piers to the gross water area of the constriction (both based on the normal water surface) and the angularity of the piers. These inputs are used to calculate "J" in the FHA documentation.</li>
<li>Use the Figure 7 ''Incremental Backwater Coefficient for Piers'' data to calculate Kp. 
[[File:FHA_Kp_arrow_crop.png|400px]]
 
'''NOTE''': the pier form loss coefficients in Hydraulics of Bridge Waterways are derived based on the cross-sectional averaged velocity through the bridge opening in the absence of piers. It's not necessary to specify a blockage value if a pier form loss coefficient estimated from this method is used.
</li>
</ol>

==Bridge Deck and Rail (Super Structure)==
When a bridge deck become partially or completely submerged, the deck could generate extra afflux resulting in increased water levels and flood extents upstream of the structure. The flow around the deck is highly 3-dimentional and complexed due to the different deck designs/profiles and/or the occurrence of pressure flow. In 2D SWE solver, depth-varying form loss values are often needed to reproduce the afflux caused by such structure. Due to the complexity of the flow, guidelines on how to set the form loss coefficient for the bridge deck are rare. We have carried out a joint research with QLD TMR (Queensland Department of Transport and Main Roads) regarding how to choose a proper form loss value for the bridge deck ([https://www.tuflow.com/media/7554/2022-bridge-deck-afflux-modelling-benchmarking-of-cfd-and-swe-codes-to-real-world-data-collecutt-et-al-hwrs.pdf Collecutt et al, 2022]). In the research, CFD modelling was conducted to investigate the characteristics of energy loss of a simple bridge with a flat bottomed deck and guardrails.
 
[[File:CFD_study.png|600px]]

Below are the key findings from the study:
*The results displayed a characteristic shape for head loss coefficient as a function of downstream water level over the deck thickness (TW/T).
*The head loss (afflux) peaks when the water level is approximately 1.6*T above the bridge soffit, and decays slowly as the bridge becomes progressively drowned out.
[[File:FormLoss_vs_TWT.png|600px]]
 
*The peak loss coefficient value is a function of the ratio of the depth underneath the deck (hB) and the thickness of the deck (T)
<ol>
{| style="text-align: center;" class="wikitable" width="35%"
! style="background-color:#005581; font-weight:bold; color:white;" width=55%| Deck Height to Thickness Ratio
! style="background-color:#005581; font-weight:bold; color:white;" width=45%| Peak Form Loss Coefficient
|-
| Scenario A (hB/T) = 2 || 0.42
|-
| Scenario B (hB/T) = 4 || 0.28
|-
| Scenario C (hB/T) = 6 || 0.20
|}
</ol>

This table can be used to estimate the deck form loss coefficient based on the bridge design (hB/T). The solid portion of the guard rails (blockage * rail depth) can be added to T in addition to the deck thickness to calculate hB/T. For bridge with more complicated designs (e.g. girders), higher form loss might be required due to the higher surface roughness of the bridge.

'''NOTE''': This form loss value should not be confused with the value of 1.56 used in the pressure flow approached adopted in [[1D_Bridges | TUFLOW 1D "B" and "BB" bridge]]. TUFLOW 1D bridge pressure flow approach is based on the section 4.13.2 "All Girders in Contact with Flow (Case II)" of ''Guide to Bridge Technology Part 8: Hydraulic Design of Waterway Structures'' (AUSTROADS, 2019). The original hydraulic experiment conducted by [https://hdl.handle.net/10217/39009 Liu et al (1957)] in a laboratory flume with a pair of bridge abutments and a deck. The flow conditions were similar to orifice flow due to the high blockage ratio caused by the abutments and the deck. When modelling bridges in 2D, the contraction/expansion losses caused by the abutments would be handled explicitly by the 2D solver, so a value 1.56 can lead to duplication of the contraction/expansion losses caused by the bridge abutments. 
 

=TUFLOW 2D Bridge Setup=
Traditionally, 2D Layered Flow Constriction (2d_lfcsh) has been used in TUFLOW 2D modelling to specify the depth varying form loss of a bridge structure. As of 2022 release a new GIS layer called 2D BG Shape (2d_bg) has been implemented in order to simplify the model input based on the findings from the joint TMR study ([https://www.tuflow.com/media/7554/2022-bridge-deck-afflux-modelling-benchmarking-of-cfd-and-swe-codes-to-real-world-data-collecutt-et-al-hwrs.pdf Collecutt et al, 2022]).
Both methods provide options for representing flow surcharging, the pressure flow of bridge decks and eventually submerged bridge flow at higher water levels. During the surcharging of bridge decks, higher energy losses can be specified to simulate the pressure flow. Four flow constriction layers are represented. The lower three layers represents the pier, the bridge deck and the rails. Each layer has its own attributes to specify the blockage and the form loss coefficient. The top (fourth) layer assumes the flow is unimpeded, representative of flow over the top of a bridge. Within the same shape, the invert of the bed, and thickness of each layer can vary in 3D.
==2D Layered Flow Constriction (2d_lfcsh)==
Four layers are used for 2d_lfsch: 
*'''Layer 1''': Beneath the bridge deck. If bridge piers exist a small form loss is usually expected due to the energy losses associated with the piers. [https://www.fhwa.dot.gov/engineering/hydraulics/library_arc.cfm?pub_number=1&id=5 ''Hydraulics of Bridge Waterways'' (Bradly, 1978)] can be used to estimate the pier form loss coefficient.
*'''Layer 2''': The bridge deck. This would be 100% blocked and the form loss coefficient would increase due to the additional energy losses associated with flow surcharging the deck. The hB/T vs FLC table from the joint research with TMR can be used to estimate the deck form loss coefficient.
*'''Layer 3''': The bridge rails. These might be anything from 100% blocked (solid concrete rails) to 10% blocked (very open rails). Sensitivity testing with 100% blockage is recommended as often debris during a flood can be substantial (see images from the Q&A section below). Some form losses would be specified depending on the type of rails and blockage. The solid portion of the rails (pBlockage*L3_Depth) can be added to L2_Depth to calculate hB/T in the table above to estimate the combined form loss coefficient of the L2 and L3.
*'''Layer 4''': Flow over the top of the rails - flow is assumed to be unimpeded.
<ol>
[[File:2d_lfcsh_attributes.png | 500px ]]
</ol>
 

The 2d_lfcsh functions by adjusting the flow width and the form loss of 2D cell faces. The combined blockage across the 4 layers is calculated at each simulation timesteps:
<ol>
 
[[File: Blockage_total_equation_01.png|600px]]
 
where 
'''''yi''''' is the actual depth of water in layer '''''i''''' 
'''''ytotal''''' is the total water depth 
</ol>

The combined form loss coefficient can be estimated using different approaches, which can be individually specified by the 2d_lfcsh Shape_Options attribute, or globally specified by command: 
Layered FLC Default Approach == [ METHOD A| {METHOD B} | METHOD C ] 
*METHOD A (previously named CUMULATE): the losses are accumulated as the water level rises through the layers according to the following equation.
<ol>
[[File:Eq_flc_cumulate.png |450px]]
</ol>
*METHOD B (previously named PORTION): the losses are applied pro-rata according to the depth of water in each layer using the equation below.
<ol>
[[File:Eq_flc_portion.png |430px]]
</ol>
*METHOD C (Recommended): this approach combines the CUMULATE and PORTION approaches by utilising CUMULATE through to the top of Layer 3 and PORTION above Layer 3.
<ol>
[[File:Eq_flc_methodC.png |520px]]
</ol>
All three methods apply a constant form loss value of L1_FLC when the water level is below Layer 2. Above Layer 2, the total form loss coefficient is increased gradually based on the thickness of water in Layer 2 and 3. Due to the depth proportioning approach used in the PORTION approach, larger L2_FLC/L3_FLC values are needed to achieve the same peak form loss coefficient as the other 2 methods. Above Layer 3, the PORTION and METHOD C approaches gradually reduce the total FLC with the increase of the water level, while the CUMULATE method continues to applies the cumulated form loss value. An example, taken from a calibration of a bridge structure from the Iowa River Flood Study is shown below. With water slightly overtopping a bridge deck, a combined form loss coefficient of 0.35 was used to match the observed head loss.
<ol>
{| style="text-align: center;" class="wikitable" width="48%"
!colspan="1" rowspan="2" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Layer
!colspan="1" rowspan="2" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Depth (m)
!colspan="1" rowspan="2" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Blockage (%)
!colspan="3" style="background-color:#005581; font-weight:bold; color:white;"|Form Loss Approach
|-
! style="background-color:#005581; font-weight:bold; color:white;" width=8%| METHOD A
! style="background-color:#005581; font-weight:bold; color:white;" width=8%| METHOD B
! style="background-color:#005581; font-weight:bold; color:white;" width=8%| METHOD C
|-
| 1 || 5.0 || 5 || 0.07 || 0.07 || 0.07
|-
| 2 || 1.5 || 100 || 0.15 || 1.05 || 0.15
|-
| 3 || 1.0 || 50 || 0.13 || 0.70 || 0.13
|}
</ol>
The figure below compares how the form loss value varies with height for the 3 methods.
 
<ol>
[[File:FLC_vs_height.png | 600px ]]
</ol>

==2D BG Shape (2d_bg)==
2D BG Shape is similar to the Layered Flow Constriction, but has several update to simplify the input based on the findings from the joint study with TMR. The lower three layers have been renamed for clarity.
*'''Pier layer''': Similar to Layer 1 in Layered Flow Constriction.
*'''Deck layer''': The bridge deck.
*'''Rail layer''': The bridge rails. The deck layer and the rail layer are treated as one '''Super Structure''' layer in 2d_bg. A combined form loss coefficient is specified using the 'SuperS_FLC' attribute. The solid portion of the rails (Rail_pBlockage*Rail_Depth) is added to Deck_Depth to calculate hB/T in the table above to estimate the combined form loss coefficient of the Super Structure.
<ol>
[[File:2d_bg_attributes.png | 700px ]]
</ol>
*'''Layer 4''': Above the top of the rails, flow is assumed to be unimpeded.

'''Inflection Point''': As shown in the joint study above, the head loss peaks when the water level is approximately 1.6*T above the bridge soffit, and decays slowly as the bridge becomes progressively drowned out. The 'SuperS_IPf' attribute (inflection point factor, default = 1.6) can be used to define the height of the inflection point. The solid portion of the rail layer is also added to the deck thickness to calculate the depth to the inflection point (DIP), i.e.:
<ol>
[[File:eq_flc_bg_infection_point.png | 520px ]]
</ol>
The form loss approach is similar to the FLC approach METHOD C, with L2/L3 replaced by a single super structure layer:
<ol>
[[File:eq_flc_bg.png | 480px ]]
</ol>

Using the same bridge example with SuperS_FLC of 0.28 and SuperS_IPf of 1.6, DIP would be set as 3.2m above the bridge soffit, and the figure further below shows how the form loss value varies with height.
<ol>
{| style="text-align: center;" class="wikitable" width="32%"
!colspan="1" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Layer
!colspan="1" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Depth (m)
!colspan="1" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Blockage (%)
!colspan="1" style="background-color:#005581; font-weight:bold; color:white;" width=8%| Form Loss
|-
| Pier || 5.0 || 5 || 0.07
|-
| Deck || 1.5 || 100 || 0.28
|-
| Rail || 1.0 || 50 || -
|}

[[File:FLC_vs_height_bg.png | 600px ]]
</ol>

== 2D Bridges Line vs Polygon Layer ==
The form loss coefficient (FLC) is applied differently when using a line compared to a polygon.

The FLC is applied at cell sides (u and v faces) as this is where velocities are calculated. When using a polyline, the FLC attribute depends on the type of the polyline: 
*Thin line (width attribute of zero) - The FLC attribute in the GIS object should reflect the total form loss value for the bridge. A thin 2d_lfcsh line will apply the FLC to a single row of cell sides. As such, this approach is cell size independent. Thin line lfcsh are the easiest setup and the preferred / recommended approach.
* Thick line (width attribute between zero and 1.5 times the cell size) - The FLC attribute is half of the total loss as the form loss is applied on each cell side of the selected cells. A cell is selected if the polyline intersects the cell crosshair. Caution should be taken when using a "thick" line, due to the fact changes in cell size can trigger a "thick" line to become a "wide" line. If this were to occur the FLC attribute would need to be recalculated to not overestimate losses.
* Wide line (only supported for 2d_lfcsh, width attribute larger than 1.5 times the cell size) - The FLC attribute is a portion of the total loss based on number of cell sides in the predominant direction of flow. Caution should be taken when using a "wide" line due to the fact changes in cell size can trigger the need to recalculate and define losses.
The number of cell sides and the assigned FLC value can be checked in the [[Check_Files_2d_lfcsh_uvpt | lfcsh_uvpt_check]] and [[Check Files 2d bg uvpt check | bg_uvpt_check]] files.
 
<ol>
[[File:2d_lfcsh_thinline.png|400px]] [[File:2d_lfcsh_thickline.png|400px]]
</ol>

For larger bridges that spread across multiple cells, it is recommended to use a polygon layer, which selects all u and v faces falling within the polygon.

Caution should be taken when specifying the FLC values for the two different 2d bridge features:
*2d_lfcsh: FLC attribute is the total loss '''per unit length''' (meters or feet) in the direction of flow. The FLC is applied to each face as 'FLC * cell size'
*2d_bg: FLC attribute is still the '''total form loss'''. Instead of converting it to "form loss per meter", the "Deck_Width" attribute is used to automatically distribute the total FLC to the selected faces, i.e. FLCface = FLC / Deck_Width * cell size. 

[[File:2d_lfcsh_polygon.png|450px]] [[File:2d_bg_polygon.png|450px]]

 
It is a good modelling practice to check the [[Check_Files_2d_lfcsh_uvpt | lfcsh_uvpt_check]] and [[Check Files 2d bg uvpt check | bg_uvpt_check]] files to confirm the number of faces selected and the FLC values assigned. It is also strongly recommended to undertake a sensitivity analysis on the applied form losses in the model to check if it makes any difference to the results and/or double check against other methods (hand calculations, other software, CFD modelling), especially if the bridge is anywhere near the area of interest. If calibration data is available, this should be used to guide the form loss value specification. 
 

= Common Questions Answered (FAQ)=
== What blockage values should I use for bridge guard rails? ==
The blockage of bridge guard rails can be anything from 100% blocked (solid concrete rails) to 10% blocked (very open rails). In addition, the accumulation of debris during a flood can be substantial as shown in the image below. Sensitivity testing with 100% blockage is recommended.
 
[[File:Bridge rail debris.jpg | 500px]]

== How to conduct sensitivity test for 2D bridges? ==
General recommendations to cross-check the results are:
* Compare computed affluxes against desktop methods (e.g. Hydraulics of Bridge Waterways, 1978) and/or other software including CFD, especially for unusual bridge designs.
* Use any recorded flood marks or general observations from past events to check and calibrate FLC values.
* Conduct sensitivity testing by assessing the impact and influence of FLC values on your modelling objectives. The afflux resulting from the FLC values will be proportional to the velocity head, i.e. ∆h=FLC*(v^2/2g). As such, if velocities are low (e.g. 1 m/s), the results may not be overly sensitive to uncertainties in the FLC values. If completing a check using this equation for a long skew bridge it is best to calculate the total structure velocity from a PO line digitised in the same location as the bridge.

Finally, after completing sensitivity testing and understanding the range of uncertainty due to unknowns like the degree of blockage and influence of FLC values (e.g. +/-20%), you are in a position to discuss with your client how best to proceed. For example, if the modelling is to set planning levels for a development upstream then it may be appropriate to choose values on the higher side (higher FLC values and/or blockage assumptions), noting that the uncertainty may be amply covered by a regulatory freeboard. Conversely, if the development is on the downstream side the conservative approach would be to use the results at the lower end of your FLC/blockage values. 
[[File:Bridge Flood Debris Loading.jpg | 500px]] 

== Should I use both FLC and blockage for layer one in 2D bridge layered flow constriction? ==
When applying FLC and blockage values to model obstructions such as piers, the following considerations need to be taken into account:
* The FLC value applies an energy loss along 1D channels or across 2D cell faces equivalent to FLC*V2/2g where V is the 1D channel velocity or the 2D cell face velocity.
* FLC values are often sourced from publications such as Hydraulics of Bridge Waterways or AustRoads (e.g. Kp chart for piers).
* If possible, establish whether the source of the FLC value is based on the approach velocity (the velocity in the absence of piers) or structure velocity (the velocity with area blocked out by the piers) noting that it often isn’t clear or stated.
** If it is the structure velocity, this is usually the velocity at the vena-contracta (point of greatest contraction within the entrance to the structure and therefore highest velocity) - see image below. Bluff or sharp-edged obstructions will have a much more pronounced vena-contracta, and therefore higher velocity compared with a round-edged obstruction.
** FLC values based on the approach velocity will be higher than those based on the structure velocity to achieve the same energy loss.
* Applying a blockage equivalent to the obstruction width will increase, usually very slightly, the velocity of the 1D channel or 2D cell face. This won’t be the vena-contracta velocity, but a velocity between the approach velocity and the vena-contracta velocity. A greater blockage will need to be applied to emulate the vena-contracta velocity.
* If the FLC source value is based on:
** The approach velocity then there is no need to apply a blockage value.
** The structure velocity then the blockage value should be applied noting that it may be appropriate to apply a larger blockage to take into account the vena-contracta.
* If it is not clear or unknown whether the FLC source value is based on the approach or structure velocity, the recommendation would be to apply the blockage in the interests of being slightly conservative on the upstream flood level calculation.
* For most minor obstructions such as bridge piers, the blockage is usually relatively small and whether included or not has a negligible or minor affect on flood levels compared with other factors such as the approach embankments and the bridge deck.
* Blockage from debris wrapped around piers can have a greater influence on the results than the effect of applying or not applying a blockage. Debris wrapped around piers can be accounted for in the FLC value calculated for the pier layer.
* As always, sensitivity testing with and without blockage and +/- the FLC value is highly recommended to understand their importance in regard to the broader modelling objectives and the effects of uncertainties in the input data, boundaries, other parameters such as Manning’s n values, and the accuracy of the numerical solution scheme (see [https://www.tuflow.com/library/webinars/#maximise_accuracy Maximising the Accuracy of Hydraulic Models webinar]).
[[File: Vena_contracta.png]] 
''Image showing the formation of the vena-contracta.'' 
 

==I don't see results that I expect when using 2d_lfcsh layer==
The 2d_lfcsh layer is a versatile feature that was designed to model bridges in 2D, but can also be used for other applications like fences, buildings raised on pillars and so on.
Some of the unexpected results could be:
* Water level going through the bridge deck in 2D map output.
* Water transiting through 100% blocked Layer 1, e.g. fences with solid base.
* SHMax.csv reporting values above the bridge deck when 2D map output reports water level lower than the top of the bridge deck.

TUFLOW is a 2D solution (not 3D), in the 2d_lfcsh layer the percent blockage and form loss coefficient applied to the cell faces is depth averaged across the entire cell face (across Layer 1, 2 and 3): 
*For bridges, where Layer 2 has a 100% blockage applied, the minimum flow width of 0.001m is used and is averaged with the Layer 1 blockage (based on the depth of the water). This may result in a water level being reported within or above the bridge deck, which would represent the pressure head.
*Layered flow constriction works by adjusting the flow area of the cell faces by any blockages to generate the correct depth averaged velocity at each face at which the form losses are applied as a fraction of the V2/2g kinetic energy. Calculating the correct velocity is critical for determining the losses as the losses are proportional to the velocity squared. 
*For a layered flow constriction cell face the flow area cannot be zero above the invert of Layer 1 to avoid a divide by zero in the computations, therefore a minimum average flow width after applying blockages of 0.001 m is applied. if Layer 1 is 100% blocked, a very small amount of water will flow through Layer 1. If this is unacceptable, instead of applying 100% blockage of Layer 1, the preferred approach is to start the layered flow constriction at the top of Layer 1 or raise the ground elevation to the top of Layer 1 using one of the Z Shape modification functions (e.g. a breakline). 
<ol>
[[File:100% Blockage Diagram.png | 500px]]
</ol>

== Can I model bridge piers explicitly in 2D using very small cells? ==
It isn't recommended to explicitly model bridge piers by blocking out the pier faces in TUFLOW, or any hydraulic modelling software based on solving Shallow Water Equations(SWE), due to the mesh resolution required and the 3-dimentiality of the flow and turbulence around the piers.

Small scale obstructions to the flow, such as trees, poles, piers, etc. cause additional head losses along a flow path due to their drag characteristics. Historically, form loss (or drag) coefficients for various profile shapes have been determined as a function of Reynold’s number through experimental testing. 

[[File:Flow round a cylinder.png]]

''The point of flow separation around an object has a major bearing on the drag coefficient and is not reliably reproduced by 2D or 3D software.''

Modelling 2D flow around profiles with the 2D or layered 3D form of the shallow water equations (SWE) as used by TUFLOW and other free-surface water flow solvers is no different in this regard. While mesh-resolved wakes behind the piers using a fine mesh can be seen in the results, the predictions for head losses show the same sensitivities (mesh size, mesh design, choice of turbulence model) as seen in 3D CFD. 

The safest and strongly recommended approach with regard to establishing head losses and consequently flood levels, is to not resolve the obstructions in the mesh but instead model the effects of such obstructions with form loss (drag) coefficients (applied to selected mesh cells) that have been derived from physical testing. This approach has been shown to provide the most consistent results across various mesh resolutions. It also has the added benefit that, by avoiding small cells in the mesh, it will provide much more efficient run times for flow solvers.

(the possible exception being CFD software)

More recently, computational fluid dynamics (CFD) has been used to attempt to reproduce the velocity field in the wake of such objects. Although providing better results than 2D modelling, the results have not always agreed well with physical tests. In particular, the drag of a given profile depends on the exact location of flow separation points, which in turn depends on the ability of the CFD code to predict the laminar to turbulent transition in the boundary layer, which is many times smaller than the profile shape itself. In general, the form loss results from CFD models show significant sensitivity to mesh size, mesh design, and choice of turbulence model. Considerable caution needs to be exercised even for CFD modelling.

== How to best convert flow constriction data (2d_fc or 2d_fcsh) into newer formats (2d_lfcsh or 2d_bg)? ==
The form loss parameters can be transferred from the flow constriction (2d_fc or 2d_fcsh) to the first layer of the layered flow constriction (2d_lfcsh) or pier layer of the 2d_bg. Definition of the remaining form loss and blockage layer inputs should follow the guidance outlined in [[TUFLOW_2D_Hydraulic_Structures#2D_Layered_Flow_Constriction_.282d_lfcsh.29 | 2D Layered Flow Constriction]] and [[TUFLOW_2D_Hydraulic_Structures#2D_BG_Shape_.282d_bg.29 | 2D BG Shape]] paragraphs. 
When using floating pontoon (type FD in the 2d_fc or 2d_fcsh) different setup might need to be used for different events. For large events when floating pontoon becomes fixed at the top of the supporting piles, standard 2d_lfcsh setup can be used. Smaller events when the pontoon is floating at different heights might require more sensitivity testing of the structure parameters to find out a setup the matches the reality as close as possible. 

== Should I model bridges in 1D or 2D Domain? ==
The recommended approach typically depends on the study objectives and if the channel upstream and downstream of the bridge is modelled in 1D or 2D. To preserve the momentum as accurately as possible the bridge should be modelled in the same dimension as the channel, e.g. 1d_nwk bridge if the channels is in 1D and 2d_bg or 2d_lfcsh if the channel is modelled in 2D. 
In 2D, the expansion/contraction losses are modelled based on the topography and don't need to be estimated as attributes as for 1D modelling. Also, for higher flows where the bridge is overtopped, 2D is preferable approach.

== What is the difference between downstream and upstream controlled flow? ==
Downstream control means a change in downstream water level will cause a change in upstream water level. Upstream control means the upstream water level is insensitive to the downstream water level and usually indicates the occurrence of supercritical flow.

== What FLC values should be used for 2d_bg bridge if hB/T is below 2 or above 6? ==
Unfortunately, we have no data beyond the ends.
* hB/T ratios of less than 2 represent a very unusual bridge sitting low to the ground, and the peak FLC may increase above the end value (FLC of 0.42) in a way that doesn't follow the research trend or extrapolation. For these cases we would recommend using CFD modelling to obtain a more informed value. Alternatively, computing an FLC based on pressure flow might be considered.
* For hB/T ratios of greater than 6, the FLC is likely to continue to decrease, but probably not significantly. Clamping to the end value (FLC of 0.20) might be considered the more conservative approach (if the primary concern is flood levels upstream of the bridge). 
 

{{Tips Navigation
|uplink=[[ TUFLOW_Modelling_Guidance | Back to TUFLOW Modelling Guidance]]
}}