search

OpenEnergyPlatform / ontology

Repository for the Open Energy Ontology (OEO)
Creative Commons Zero v1.0 Universal
106 stars 23 forks source link

Include relations from emission coefficient to the processes which cause emissions #582

Closed Vera-IER closed 3 years ago

Vera-IER commented 3 years ago

Description of the issue

We need to include some relations from emission coefficient to the processes which cause emissions.

Ideas of solution

primary energy production has_process_attribute emission factor combustion has_process_attribute emission factor industrial process has_process_attribute emission factor (when industrial process is implemented #534 ) energy transformation has_process_attribute emission factor (maybe we need to include subclasses of energy transformation or wait until #372 is solved and include relations like that: coal power unit participates in energy transformation has_process_attribute emission coefficient)

Workflow checklist

I am aware that

l-emele commented 3 years ago

Not every energy transformation has emissions, e.g. transformation of electricity to heat. So has process attribute cannot be related to the general energy transformation, but maybe to subclasses.

For a coal power unit, there are a couple of (sub)processes:

The partial energy transformations can be abstracted to an overall energy transformation from chemical energy to electrical energy. The emissions are solely associated to the combustion process. Concluding from that I would argue that we need the following statements in the OEO:

l-emele commented 3 years ago

Similar for industrial processes and energy production: Only a subset of these have emissions and hence an emission factor.

l-emele commented 3 years ago

I think, a better solution would be a subprocess relation as emission is a process and we want to related this emission process to other processes. The emission factor is only a process attribute, so I think we should deal with the relations between the processes itself first.

Maybe someone of @OpenEnergyPlatform/oeo-general-expert-formal-ontology could give some input here?

sfluegel05 commented 3 years ago

I think it would help to use has part relations here for structuring. For the example above, we could use a "overall chemical to electrical energy process" as a starting point. This process has the relations

The combustion could have the relations

And the emission has the relation has process attribute some emission factor

This means that we have several different processes which are either proper parts of other processes or temporally coextensive to another process.

Vera-IER commented 3 years ago

I would try to keep it as simple as possible. We don't have the subprocess of a coal power unit in the ontology yet and I think they are not neccesary to describe an energy system model. We consider the coal power plant in the model more like a black box, which has some inputs, some outputs, some ratios between them, some costs etc.. One output of the process are the emissions and the emission factor describes how much emission per unit activity of that process is produced. So the emission factor is related to the (power generating/combustion) process, not to the emission itself.

Only a subset of these have emissions and hence an emission factor.

Can we solve that problem by changing the relation name to can have process attribute. Or explaining that the emission factor could also be zero?

l-emele commented 3 years ago

One output of the process are the emissions and the emission factor describes how much emission per unit activity of that process is produced. So the emission factor is related to the (power generating/combustion) process, not to the emission itself.

According to our definition, emission is not an output but a process. The output you are talking about would probably be something like an emission value that has a mass unit as unit. We currently do not have this but can define such a value class if needed.

Only a subset of these have emissions and hence an emission factor.

Can we solve that problem by changing the relation name to can have process attribute. Or explaining that the emission factor could also be zero?

I am not sure whether it is useful to have statements in an ontology that are not clearly true or false. What do you think, @sfluegel05?

Vera-IER commented 3 years ago

The output you are talking about would probably be something like an emission value that has a mass unit as unit. We currently do not have this but can define such a value class if needed.

Thank you for pointing that out. I think a class emission value is definitely helpful for describing model results or datasets.

sfluegel05 commented 3 years ago

When I think about it, I am not sure if we even need a concept of emission. emission is currently defined as a process releasing byproducts from human activity (e.g. production, distribution or consumption) into the environment. So it always depends on another process and is just special because of who participates in the process (humans) and where some of the outputs go.

That means we could treat emissions just like any other output: We have a process (e.g. the power generation in our coal plant). This process has several inputs and outputs (e.g. coal, electrical energy, ...) and one of these outputs happens to be a carbon dioxide (or another emission gas). The carbon dioxide is then related to a quantity value.

In that case we wouldn't need an emission-process or emission value. If we still want to explicitly state that something is an emission we could make that a quality of the carbon dioxide (similar to origin, maybe called destination). Also, an emission factor as a process attribute could be used if we don't want to say what kind of emission we have but just how much. So an emission factor would be directly related to a quantity value.

Only a subset of these have emissions and hence an emission factor.

Can we solve that problem by changing the relation name to can have process attribute. Or explaining that the emission factor could also be zero?

I am not sure whether it is useful to have statements in an ontology that are not clearly true or false. What do you think, @sfluegel05?

I think it would be best to leave out this relation if there are processes which don't have it. That way, anybody who wants to use an emission factor for his process can do this but no one has to.

l-emele commented 3 years ago

Regarding:

We don't have the subprocess of a coal power unit in the ontology yet and I think they are not neccesary to describe an energy system model.

I don't think that we need coal power unit process in detail, but we could define a more abstract process: Process XYZ is a transformation that contains at least two sub process: An energy transformation of chemical energy from fuel into electrical energy and an emission process. This process XYZ could then be applied to the fuel power unit class via participates in. Then all subclasses including the coal power unit would inherit this property.

l-emele commented 3 years ago

That means we could treat emissions just like any other output: We have a process (e.g. the power generation in our coal plant). This process has several inputs and outputs (e.g. coal, electrical energy, ...) and one of these outputs happens to be a carbon dioxide (or another emission gas). The carbon dioxide is then related to a quantity value.

In that case we wouldn't need an emission-process or emission value. If we still want to explicitly state that something is an emission we could make that a quality of the carbon dioxide (similar to origin, maybe called destination). Also, an emission factor as a process attribute could be used if we don't want to say what kind of emission we have but just how much. So an emission factor would be directly related to a quantity value.

Please be aware that we already had a lot of discussion on emissions, e.g. in #56, #234 and several meetings. It was a difficult compromise as a lot of aspects had to be reflected, as this is also closely related to the whole portion of matter topic. So please let's find a solution that builds upon what we already have in the ontology without restructuring to much. My last comment describe such a potential solution.

@stap-m: What is your opinion?

Vera-IER commented 3 years ago

Process XYZ is a transformation that contains at least two sub process: An energy transformation of chemical energy from fuel into electrical energy and an emission process. This process XYZ could then be applied to the fuel power unit class via participates in.

I like the proposal.

And I think we don't need emission value necessarily if we have all the emissions individually described in the OEO (thanks @sfluegel05 for reminding me). The emission factor should be always specific about which greenhouse gas or pollutant is released.

I think it would be best to leave out this relation if there are processes which don't have it. That way, anybody who wants to use an emission factor for his process can do this but no one has to.

Seems like I had a wrong understanding of relations. I thougt if there is no relation implemented, I cannot use the "attributes"/in this case the emission factor to describe a specific process.

l-emele commented 3 years ago

Process XYZ is a transformation that contains at least two sub process: An energy transformation of chemical energy from fuel into electrical energy and an emission process. This process XYZ could then be applied to the fuel power unit class via participates in.

I like the proposal.

Great! The most straight forward label for such a process would be fueled power unit process as it is the process all fuel power units apply. This label is a bit awkward, however, and we might find a better label.

Vera-IER commented 3 years ago

I think the process can be applied to fuel power plants and fuel power units. Are they supposed to be the parent class for the individual fuel power plant/units? Because I see them all on the same level in the OEO.

A label for the XYZ process could be fuel-powered electricity generation (process)?

l-emele commented 3 years ago

To sum up what I understand we agreed so far:

Additionally we should also add the relation between emission and emission factor via:

I think that would solve this issue, wouldn't it?

Vera-IER commented 3 years ago

I agree with the first 3 bullet points, but I have another view on the relation between emission and emission factor. I think only the fuel-powered electricity generation has the process attribute emission factor. One would multiply the activity of the fuel-powered electricity generation process with the emission factor and gain one or several of the portion of matter emissions. So I think the relations between the emission process and emission factor are not necessary or could be even confusing.

l-emele commented 3 years ago

Sorry, I don't understand this. In #243 #581 we just defined the emission factor as: An emission factor is a process attribute that quantifies the emissions or removals of a gas per unit activity. This clearly gives the link to the emissions.

EDIT: The fuel-powered electricity generation has several subprocesses, one of it being emission. The fuel-powered electricity generation inherits the property has process attribute emission factor from the emission. If emission was no part of fuel-powered electricity generation then this process could not have an emission factor.

Maybe @stap-m could provide a third opinion.

Vera-IER commented 3 years ago

I don't agree to the has process attribute between emission and emission factor. The def ofhas process attribute is: A relation between a process attribute and the process it existentially depends on. How much emissions are produced depends on the activity of the fuel-powered electricity generation process. If there is no activity (electricity generation) in a certain time period, there are no emissions. So I would say that the emission factor existentially depends on the fuel-powered electricity generation process and not on the emission process.

l-emele commented 3 years ago

Any opinions by @OpenEnergyPlatform/oeo-domain-expert-energy-modelling or @OpenEnergyPlatform/oeo-general-expert-formal-ontology ?

jannahastings commented 3 years ago

From my side I agree with what you have proposed @l-emele

Vera-IER commented 3 years ago

Okay, seems like my understanding of how I would built an energy system model cannot be fully translated to how we built the ontology. Shall I implement it according to the summary above?

l-emele commented 3 years ago

Go for it!

stap-m commented 3 years ago

Looks fine to me, too.

stap-m commented 3 years ago

Since it's already implemented, I'm closing this issue.