1D Urban Links Tab

The 1D Urban Links tab of the 2D Simulation window is accessed by checking the box Include link to 1D urban model provided on the General Tab.

This tab is used to provide details about 1D urban networks connected to the 2D domain to create an integrated simulation.

Urban network

This field is to provide the 1D urban network file (extension .inp). A button to the right provides access to a standard Windows file explorer window to navigate to the file location. A button below the field enables viewing of the .inp file selected within a text editor.

Override simulation times checkbox

A checkbox to override the start/end times of the 1D urban simulation, checked off by default. By ticking this box, the 1D urban component will assume the simulation start and end times from the 2D simulation.

This is equivalent to setting the value of the overridetimes attribute in the <ief> element to “true” within the control file, e.g.

<ief overridetimes="true">..\SWMM\LONG_Op1200_Linked1_15.inp</ief>

Use ponded volume checkbox

A checkbox to use the ponded volume as the linked flow. This instructs the solver to automatically handle the ponded areas specified in the urban model, and can make the setup of integrated models significantly more efficient, as it negates the need to manually edit individual ponded areas for all linked nodes. However, it should be noted that this is only appropriate under certain conditions, which are discussed below.

Linking methodology

To be able to advise which method is better, we first need to understand how the respective methods work: in general, the 2D solver passes the “above-ground” volume to the 1D urban solver. The 1D urban solver then computes the volume exchange as a discharge rate to or from the 2D model. This rate is then applied to the 2D solver as a sink or source term. The differences in the two methods are in how that volume is calculated, which is as follows:

Up to and including v5.0, the volume was always calculated by summing the depths above the link shape in the 2D Solver, whether this is a point, polyline, or polygon, and multiplying by the grid cell surface area. This is then applied to the 1D urban solver, effectively as an average depth, by dividing this volume by the ponded area.

The alternative option introduced from v5.1 is to apply the average water surface elevation over the 2D link shape – this is calculated by summing the elevations over only the wet cells on that link shape and dividing by the number of wet cells – and then passing this elevation to the 1D urban solver; the 1D urban solver then assumes a ponded volume, based on this average elevation, the 1D urban ground level and the ponded area.

Furthermore, there are essentially three modes in which water can be exchanged between the 2D and 1D urban solvers (in addition to dry mode – no water exchanged). These are as follows:

1. Water freely discharges into the 1D urban model (i.e. if the Urban model is not running “full” at that junction); in this case, the inflow rate is determined by the 2D, using a weir-type equation, and added effectively as an External Inflow to the 1D urban model

2. Water discharges into the 1D urban model as “drowned” inflow (i.e. if the Urban model is running “full” at that junction); in this case, the inflow rate is determined by the 1D urban model, and removed from the 2D model as a sink term

3. Water discharges out of the 1D urban model (this can happen whether the 2D model is wet or dry at the linked cell location[s]); in this case, the outflow rate is determined by the 1D urban model, and added to the 2D model as a source term

Exchange volume calculation

Returning to the volume calculation, the advantages of the second method are that one does not have to ensure that the respective areas match, and that this is independent of the linking area chosen. However, there may still be results sensitive to the choice of areas, as illustrated in the following cases:

• 2D link area is “too small” (this could occur for instance with a point link file, combined with relatively high discharges from 1D to 2D and/or a small grid size). This is due to the discharge being applied as a source term to a single or small number of cells, resulting in a high [additional] depth, which may not disperse before being passed back to the 1D urban solver at the next timestep

• Large ponded area – this has beneficial effects when the flow is predominantly in mode 2 (flow direction from 2D to [submerged] 1D urban), although can pose inconsistencies when the flow is in mode 3 (flow from 1D urban to 2D). As described below, for flow ingressing into the 1D urban model [mode 2], the iterative behaviour of the latter ensures that the recalculated water level does not oscillate with a relatively large ponded area; however for flow emanating from the 1D urban model, this can lead to similar issues as described above – the flooded depth assumed by the 1D urban model will be lower than it is in the (more physically realistic) 2D model. This difference can result in more water attempted to be returned to the 1D urban model than is present in the 2D model, and lead to mass conservation errors.

• Uneven 2D ground level – in either case, this can generate multiple warnings of the type “Warning: Q-link is trying to extract flux (..) from dry area …”, which occurs when there is not enough volume immediately above the link shape to provide the inflow requested by the 1D model component, indicative of a link area or depth which is too small. Options to reduce these warnings are:

  • Locate 2D link shape elements where ground elevations over the extent of the link shape do not vary significantly.

  • Alternatively, use the “Auto-adjust DEM…” [2D ground level] option in the 2D simulation settings, if appropriate.

Summary

The linking methodology between 1D urban and 2D models can be summarised as follows:

• The main resultant differences in the “new” method occurs in modes 2 and 3 above, whereby the ponded volume is treated as the product of depth above surface and ponded area. This temporarily creates / removes artificial volume in the 1D urban component, but is only used in order to calculate the discharge rate, as opposed to any “real” mass balance error. The real above-ground volume remains in the 2D component, and the volumes reported in the continuity section in the report file account for this.

• When the “new” method iterates towards a solution, the urban component calculates an interim water level (after the ponded inflow/outflow is applied) based on what it sees as the ponded volume; therefore, if the ponded area is relatively small, this interim water level can be significant [jumping between the true level and the interim level during the iterative procedure] and therefore influence the results.

• Since mode 1 (free discharge from 2D to 1D urban) involves no calculation of ponded area/volume, and is independent of the 1D urban model (other than ascertaining that it is not surcharged), then this method is identical irrespective of which option is chosen.

Link-element details

In this field details of the link-element shapefile(s) connecting the urban network and the 2D domain should be provided. Buttons to the right enable the addition and removal of shapefiles. 

Advanced Parameters

The 1D Urban Links tab also displays some additional parameters that affect linking. Generally, these can be left at their default settings.

Weir discharge coefficient and weir modular limit

The weir discharge coefficient refers to the coefficient k in the weir equation Q=kbh^1.5and weir modular limit refers to the depth ratio above which the weir is considered drowned. These are global parameters that may be applied to any weir type links in your model. However, each individual weir link can have its own discharge coefficient specified as a feature attribute in the link shapefile. This value would override the global figure set in the 2D model interface (set the individual attribute value to “-9999”, the default value, to utilise the global setting).

Ground elevation auto-adjustment

The 2D model data will include a ground grid that defines a ground elevation at each 1D urban link location. However, there may also be elevation* values set inside the urban network (.inp) file (possibly based on more accurate local survey data). This setting in the 2D interface instructs the 2D solver to auto-adjust ground elevations if they are significantly different to the urban model elevation. This is to ensure that the water levels exchanged between the two solvers are consistent with the respective ground/crest elevations in each component solver (otherwise the depths seen by one may be inconsistent with that perceived by the other). If you activate this option, you are also required to set the following:

• a minimum elevation difference, beyond which the solver will make a level adjustment.

• whether to adjust if the 2D ground level is lower than the 1D (i.e. adjusts the 2D ground level upwards, the default), higher than the 1D (i.e. adjusts the 2D ground level downwards), or in either direction .

*The 1D urban ground elevation is defined as invert level + maximum depth

Further Technical Details

This is equivalent to setting in the xml simulation control file via the adjustz element, using the attribute direction (valid values are “upwards” [if 2D elevation is lower] – the default, “downwards” [if 2D elevation is higher], or “both” [for either]). The (numerical) value of the element is the threshold, beyond which any differences in the 2D ground elevation are applied, e.g.

<adjustz direction="upwards">0.1</adjustz>

NB The 2D ground elevation is offset so that it is, as a minimum, equal to the 1D urban elevation level plus the requested offset.

Information / warnings are written to the 2D simulation log file if:

• An adjustment to the 2D ground elevation has been made

• If the automatic adjustment setting is not selected, but the ground elevations (1D Urban and 2D) differ by more than 0.1m (0.328ft).

Hint: it is advisable to check surrounding ground elevations around the link area too – any adjustments made to the ground grid will only be at the link location, which could lead to abrupt changes in the ground grid. After processing, one can view the zmod (topographical modifications only) and zcen (entire ground grid used) check files to observe this, and use topographical modification features to further amend as necessary.

Toolbar functionality

The toolbar provides functionality as follows:

GPU Solver / CPU Solver 

Radio buttons to select the solver used for the simulation. Please see the page The GPU solver for further details.

Threads to use in run

A drop-down to provide the threads to use in CPU-based simulation runs. This field is disabled when the GPU solver radio button is selected.

Run

Runs the simulation. A new window will open providing live diagnostics and information about the simulation run. 

Save

Saves the details of the simulation.

Close

Closes the simulation window.