Flow control model

[dlohmeier]

The flow controller is a branch component that controls the flow through itself. It means that the pressure drop is adapted to reflect the desired flow situation. This is especially helpful for heating grids, where otherwise the flow through heat exchangers needs to be controlled through the loss coefficient.
As the flow controller is a new component, it must be reflected in many parts of pandapipes:

• There is a new FlowControl component class (all other component models did not really change).
• There are new create-functions (single and bulk) in the create.py module.
• The flow controller was added to the plotting library (i.e. there is a new collection and a new patch function, and it was added to the simple_plot function call) as well as a "deafult" component to networkx graphs and toolbox functions.
• New tests were implemented to test the flow control behavior and other aspects like graph searches or plotting collections.

[codecov]

Codecov Report

Base: 87.90% // Head: 88.34% // Increases project coverage by +0.43% 🎉

Coverage data is based on head (68a2dd1) compared to base (888de82).
Patch coverage: 95.65% of modified lines in pull request are covered.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #413      +/-   ##
===========================================
+ Coverage    87.90%   88.34%   +0.43%     
===========================================
  Files           80       81       +1     
  Lines         4753     4863     +110     
===========================================
+ Hits          4178     4296     +118     
+ Misses         575      567       -8     

Impacted Files	Coverage Δ
pandapipes/component_models/component_toolbox.py	87.80% <ø> (ø)
pandapipes/io/file_io.py	82.14% <ø> (ø)
pandapipes/pf/pipeflow_setup.py	88.55% <ø> (ø)
pandapipes/plotting/plotting_toolbox.py	100.00% <ø> (ø)
pandapipes/toolbox.py	85.14% <ø> (ø)
pandapipes/topology/create_graph.py	85.29% <ø> (ø)
...pes/component_models/pressure_control_component.py	94.66% <50.00%> (ø)
pandapipes/plotting/simple_plot.py	82.79% <88.00%> (+9.32%)	⬆️
pandapipes/plotting/collections.py	83.33% <89.65%> (+1.21%)	⬆️
...dapipes/component_models/flow_control_component.py	98.24% <98.24%> (ø)
... and 11 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[jkisse]

This might solve #361

[jkisse]

I think this will be very helpful in heating networks. But does this actually work with gas grids? If the velocity is fixed, the mass flow still depends on the density (but the user will assume that it is enforced, since it is an input parameter). I'd propose that an Error or Warning should be raised if fluid.is_gas

[dlohmeier]

I think there is a misunderstanding here. We fix the velocity internally, that is true. But we have to keep in mind that this is the velocity at norm state, so in order to derive the mass flow from it, we will always multiply it with the norm density. The velocity itself will change indeed inside this component, depending on the pressure / compressibility.

[dlohmeier]

I added a test for the flow control model and also added it to the toolbox functions and nx-graph creation. What is still missing is plotting. I would like to draw a valve patch with a T above to show that the valve is regulated.

[dlohmeier]

Thanks for the great addition! Something that just crossed my mind is that we might need a keyword like "control_active" as in the pressure regulator component. If it is set to false, the flow control acts like an open valve. For plotting I thought of something fancy:

• if not "in_service": make it white
• if "in_service" but not "control_active": fill
• if "in_service" and "control_active": fill half (i.e. split the drawn triangles and make one filled and one white triangle out of them)

[dlohmeier]

Thanks for the hint, I did that.

[jkisse]

Heat exchanger and valve have a loss_coefficient - would it make sense to add a loss coefficent for the flow control as well?

[dlohmeier]

I am not too sure about that because the flow controller models an ideal control valve. In case of active control (which should be the normal behavior), the loss coefficient would have no effect on the model. It could only come into play when the flow controller acts as an open valve.

not enough reviewed

[jkisse]

I've added a chapter in the documentation. Please check & update it if required, @dlohmeier

[dlohmeier]

Looks good, thanks!