Restructure grid, node and link

[solar-c]

Since graph,node,link are more abstract terms I would be careful to create any relations. A graph can exist of a single node, so a node can be a graph itself. link could be defined as part_of some graph, while node can be part_of some graph but also a subclass of graph. The order-zero graph (empty set graph) has no nodes and links, the singleton graph has no links. Therefore, a graph does not necessarily have a link or node as parts.

[stap-m]

There are no definitions provided yet. We should discuss the terms from the energy modelling point of view and add definitions.

• node = electricalGridNode --> Are there other kinds of nodes in this domaine which should be considered in the OEO, e.g. a mathematical understanding?
• graph in which context does this term occur in the energy modelling context?
• link = powerLines like SuedLink --> are there other kinds of links in this domaine to be considered in the OEO?

[l-emele]

Nodes and links in energy modeling not only occur in modeling of electricity grids, but also in other grids (district heating grids, gas grids, etc.). So it an electricalGridNode is a node but is not equal to a node.

Likewise relation between link and powerLines.

[stap-m]

Thanks for the clarification @l-emele. What about:

grid: BFO: object with subclasses 'electricityGrid', 'gasGrid', etc definition: assembly of systematically connected elements that can work as a supply system

gridComponent: BFO: fiat object part, 'part_of grid' definition: discrete element of a grid which serves a certain purpose

gridNode: BFO: fiat object part, 'is_a gridComponent', 'part_of grid' definition: part of a grid where two or more links meet

link: same as gridNode definition: a grid component that serves as a connection between two other grid components

electricalGridNode: BFO: fiat object part, 'part_of electricityGrid'

@gnn: Please check BFO parent classes. Relation to #41.

[akleinau]

[updated version down below]

[stap-m]

For grid I agree.

[stap-m]

ElectricityGridComponent:

• class: ArificialObject. Def: ElectricityGridComponent is an Arificial Object that is a discrete element of an Electricity Grid, (serving a certain purpose). ?
• I don't think we need GridComponent aswell.

[stap-m]

Node and link:

• I agree on including the relations 'part_of_grid'. Then we don't need to reclassify them in my opinion. If we added GasGrid this would make sense.

[akleinau]

I agree with ElectricityGridComponent as an Arificial Object. But if we implement a class Gasgrid (we need a definition) mybe we should replace ElectricityGridComponent with the general GridComponent to also include gasgrid components?

[akleinau]

for the ElectricityGridComponent def I think we can leave the purpose part out because a component of a grid is also a component of a grid without a purpose.

[akleinau]

so as bullet points current situation, as tasks a bit modified changes proposed in this issue

grid

• classification: BFO: object aggregate
• definition: A grid is an object aggregate of systematically connected artificial objects that can work as a supply system
• subclass 'electricityGrid'
• [x] add: subclass 'gasGrid'

ElectricityGridcomponent

• classification: ArtificialObject, part of some ElectricityGrid
• definition: A grid component is a part of an electrical grid
• [x] change definition: "Arificial Object that is a discrete element of an Electricity Grid."

Node

• classification: BFO: fiat object part
• [ ] definition: "a fiat object part of a grid where two or more links meet"
• [ ] name change: GridNode/ ElectricalGridNode
• [ ] add relations 'part_of grid'
• [ ] reclassify as ElectricityGridComponent?

link

• classification: BFO: fiat object part
• [ ] definition: a fiat object part that serves as a connection between two other grid components
• [ ] add relations 'part_of grid'
• [ ] reclassify as ElectricityGridComponent?

[akleinau]

hmm thats a bit complicated, are graph, link and node fiat object parts or objects?

[jannahastings]

I suggest we try to find primary labels for these that are less generic and more specific to the electricity context, to avoid ambiguity with other types of grid, node and link? (I see some are already suggested in the proposal above.)

e.g. (just using the proposed definitions)

supply grid: A supply grid is an object aggregate of systematically connected artificial objects that can work as a supply system. alternative term: grid

electrical grid node: "a fiat object part of an electrical grid where two or more links meet" alternative term: node

grid component link: "a fiat object part that serves as a connection between two other grid components" alternative term: link

[stap-m]

Grid/ supply grid: I guess all the grids concerning energy system modelling are supply grids. @l-emele what do you think? Along with electricity grid, we'd need gas grid and heating grid. Anything else?

• Electricity grid has the def: An electricity grid is a grid that distributes electrical energy / electricity.
• For gas grid it could be something like: A gas grid is a grid that distributes gaseous fuel, mostly methane.
• For heating grid it could be: A heating grid is a grid that distributes thermal energy via a thermal transfer fluid from a heat souce to a heat sink.

I could go with both terms, grid or supply grid.

grid component link: The def of @jannahastings sounds good to me. The technology specific types of links (power line #18, gas pipeline, heat line/heating pipeline) would then be subclasses of this class.

Grid node: all kinds of grids do have grid nodes, which are at least the systematically connected objects. So, an upper class grid node would make sense I think. Apart from link endings, an electricity grid node unusally constists of other components, too, like trafo/swithyard,... And I think this holds true for other grids, too (e.g. compressor stations for gas). Maybe we should reflect that in the def.

Do we also need/want a class grid component as parent of electricity grid component and the other grid components?

[akleinau]

I think it would be more consistent if we make grid component classes for all grids or just for the superclass grid instead of just electricity grid. Making grid component classes for all grids would be nice to see e.g. all the electricity grid classes together

[stap-m]

Ok, I'll start implementing.

[stap-m]

We still have the class graph left, without definition. I am going to delete it.

[stap-m]

It makes sense, that link and node are fiat object parts of a grid. On the other hand, its current subclasses, e.g. powerline or transformer, should rather be defined as artificial objects/grid components in my opinion. Could we use the has_function property, or something like that, to describe that a powerline is a link (although link is not defined as function)?

[akleinau]

if an object like a powerline is a link than it can be clearly differentiated from the rest of the grid and is therefore no fiat object part. So I think it makes more sense to classify link and node as sth else. Maybe make them roles? A powerline has the role "link" in a grid?

[akleinau]

@jannahastings what do you think?

[jannahastings]

if an object like a powerline is a link than it can be clearly differentiated from the rest of the grid and is therefore no fiat object part. So I think it makes more sense to classify link and node as sth else. Maybe make them roles? A powerline has the role "link" in a grid?

I don't think we should move link and node to the dependent entities branch (i.e. as roles or functions), as they do not seem to me to be attributes (that inhere in other entities) but entities in their own right. And looking at the current definition of powerline: "A powerline is a grid component that is an uninterrupted transmission system between two nodes in an electricity grid." -- that also suggests that this class is always used to pick out a part of the larger grid. So if it is not a fiat object part, it is nevertheless still an object part. So I think we can just go with classifying power line as a link, and classifying link as a grid component. The parent of grid component can be "object and part_of some grid" (fully defined)

[stap-m]

Ok, I implemented @jannahastings suggestions. This is how the structure looks right now:

Since powerline is definitly an electricity grid component I'd suggest

• to make all subclasses of electricity grid component direct subclasses of grid component
• to make electricity grid component, heating grid component and gas grid component equivalent classes with "grid component and part of some electricity grid" etc.

[akleinau]

yes sounds good