Core+ DynamicFlood – Introduction

You can run hydrodynamic 2-dimensional surface flow simulations from SCALGO Live using Core+ DynamicFlood. You can initiate a simulation from a workspace, and the results are delivered as a modelspace where you can easily explore, download and share the output.

Note that we have not yet released the add-on module in all countries, please contact us if you have questions about availability.

Initiating a hydrodynamic simulation

When you want to run a hydrodynamic 2D simulation of surface water flows in a workspace, you go to the pane “More” in the workspace toolbox and press “Run hydrodynamic engine…”

This opens a new window with options regarding the rain input to the simulation. By clicking the eye icon over each rain event, information is shown in a side video, where the rain event can be edited and saved as a new rain event on the workspace.

You can add your own rain events through the “+” icon next to the list of rain events. In some countries, clicking the “+” icon will also allow you to switch the default rain events over to those belonging to another nearby region. Your organization administrator can configure a different set of default rain events that replace the ones provided by SCALGO.

Furthermore, by clicking on the cogwheel, you can specify the simulation's resolution as well as the length of the time step for saving results (this does not influence the simulation time step) and the total length of the simulation.

By default, the simulation will run in the same resolution as the workspace, but it is possible to select a coarser resolution. The full-resolution elevation data is still used to a certain degree in the model thanks to TUFLOW's Sub-Grid Sampling feature, but the simulation will happen on a coarser grid, which greatly speeds up the simulation. The TUFLOW manual contains more information about Sub-Grid Sampling (SGS) in Section 7.4.3. The modelspace contains the elevation layer and output layers in the selected resolution.

Progress information

While DynamicFlood computations are running, a single progress bar is shown to indicate how far the currently-running computation has progressed, and the progress bar can be expanded to view the queue of simulations currently waiting to run in your organization.

Note that your hydrodynamic computations take a great deal longer than regular workspace computations and that other users in your organization might be using all your available GPU/TUFLOW resources in SCALGO Live. In such case, your job will be automatically queued and started when the resources become available. If you want to cancel the running command, you can do that from this progress bar. Note that you can only cancel jobs started by yourself.

When you expand the progress bar, the status of individual rain events that are being simulated is shown. The queue shows both your own and your colleagues' simulations, each showing the name, resolution and area, and the number of rain events in the simulation. If your organization has a shared GPU license, your job may be queued while a user in another organization is running. In this case, the queue merely displays a placeholder indicating this situation and does not display any details of the other organization’s simulation.

For your own simulations, you can move the mouse cursor over each rain event to see the current progress in percent, the approximate remaining runtime and whether the current timestep of the model is considered slow or not. By clicking the pencil button, you can change the name and description of the modelspace that will store the results of the simulation.

Under normal circumstances, the progress bar is green for each rain event. If the model is slow, meaning that the timesteps are below an acceptable threshold, the progress bar becomes red, giving a quick indication that the model might take longer than expected.

The acceptable threshold to determine if a model is slow, is the model's resolution in metres divided by 20. This means that for a 1-meter resolution simulation, the model is considered to be slow if the timestep goes below 0.05 seconds.

You can click the rain event's progress bar to bring up the status window, which displays the simulation's current status within that event. The window shows the mass balance, as well as the number of repeated timesteps, which can indicate sudden changes in the model's behavior.

The graph, which shows where in the rain event the simulation is at the moment, can be enriched with additional charts, such as the model's timestep over time and control numbers. There are three different control numbers, Nu, Nc and Nd, that provide insight into various characteristics of the simulations. Expert modellers use these to better understand the factors influencing the model's speed and performance. We refer the mathematically inclined reader to the TUFLOW manual for more information on the three control numbers.

Looking at the result

As explained earlier, the result is delivered as a modelspace. The computation processes your rain events one by one, and they will appear in your modelspace as they are finished. Thus you will be able to see the result from some time steps before the full computation is finished.

The result is delivered in three different data types, namely Water Depth, Velocity and Flux. For each data type, a time-varying layer is available as well as a layer showing the maximum over the simulation. When the time-varying Water Depth or Flux layer is enabled on the map, you can use the Point Query and Profile tools to obtain graphs of depth or flow over time in the simulation.

The DynamicFlood info popup on the modelspace is accessed by the "i"-icon next to the event dropdown in the dock. In this info popup you can view the physical parameters (rainfall, infiltration, surface roughness), the output dimensions (model resolution and output interval), and information about the model's timestep, both in a graph that shows the model timestep over time similar to the status window available for running models (described above), and as a Model Speed layer showing the areas where the model has to use a small timestep to avoid becoming unstable.

During the computation, the model maintains information about the local complexity of the computation; in complicated areas where the model may become unstable, the model may need to use a smaller timestep in the simulation to ensure the output is well behaved. Understanding where this is happening will allow you to tweak the model in future runs, thereby making the run more stable and thus faster to run. The Model Speed layer, enabled via the info popup, shows this spatial distribution of the model timestep. The yellow and red colors on the map correspond to the "slow" and "very slow" designations used to describe the model speed while the model is running, as described above. That is, the yellow areas show where the timestep is considered "slow", which for a 1-meter resolution simulation means that the timestep is below 0.05 seconds, and the red areas show where the timestep is considered "very slow", which for a 1-meter resolution simulation means that the timestep is below 0.02 seconds. Note that these are the default thresholds for the yellow and red areas, and you are free to change the thresholds through the cogwheel menu on the Model Speed layer.

If you want a deeper understanding of how we setup the hydrodynamic model you can proceed to the next section of this manual.