Distinguish between energy transfer and specific energy transformations

[l-emele]

Description of the issue

From OEKG:

• distinguish between transfer and transformation via energy input, e.g. electricity generation has energy input not electrical energy
• energy transfer: axiomatise that the type of energy input is the same as the energy output (and make equivalent class)

Ideas of solution

Adapt electricity generation process and heat generation process to:

• 'energy transformation' and ('has energy output' some 'electrical energy') and ('has energy input' not 'electrical energy')
• 'energy transformation' and ('has energy output' some 'thermal energy') and ('has energy input' not 'thermal energy')

Axiomatise somehow that the type of energy input and energy output are the same for energy transfer.

Workflow checklist

• [ ] I discussed the issue with someone else than me before working on a solution
• [ ] I already read the latest version of the workflow for this repository
• [ ] The goal of this ontology is clear to me

I am aware that

• [ ] every entry in the ontology should have a definition
• [ ] classes should arise from concepts rather than from words

[l-emele]

@stap-m : I think, the first part (adapting electricity generation process and heat generation process) can be implemented immediately, right?

For the second part (adapting energy transfer) we probably need some input from @OpenEnergyPlatform/oeo-general-expert-formal-ontology .

[stap-m]

I think the 'energy transfer' also depends on the potential axioms of #1530 , e.g. whether we need to express not conditions.

[stap-m]

Now that #1530 is implemented, we should adjust the proposal for electricity generation process and heat generation process to:

• ~'energy transformation' and ('has main energy output' some 'electrical energy') and not ('has main energy input' some 'electrical energy')~ 'energy transformation' and ('has main energy output' some 'electrical energy') and ('has main energy input' only ('chemical energy' or 'thermal energy' or 'radiative energy' or 'potential energy' or 'kinetic energy'))
• 'energy transformation' and ('has main energy output' some 'thermal energy') and not ('has main energy input' some 'thermal energy')

For energy transfer subclasses we can as well find axioms now:

• electrical energy transfer:
  • 'energy transformation' and ('has main energy output' some 'electrical energy') and ('has main energy input' some 'electrical energy')
  • maybe "electricity transfer" as alternative term?
• heat transfer: 'energy transformation' and ('has main energy output' some 'thermal energy') and ('has main energy input' some 'thermal energy')

[stap-m]

We have also chemical energy transfer, which has no subclasses yet. Probably, this should imply things like the transfer or natural gas through a pipeline? I would argue to depict this in a class material transfer or whatever called, transferring energy carriers instead of chemical energy.

[l-emele]

The Sabatier reaction is an example of a chemical energy transfer that is relevant for our domain: It is both a chemical reaction and a chemical energy transfer: Chemical energy that is stored in hydrogen is tranfered to methane. For our domain it is much more important that a chemical energy transfer happens here than that it is a chemical reaction.

[l-emele]

I agree to your proposed axioms for electricity generation process, heat generation process and energy transfer. However, we have to check whether existing inferred subclasses need then additional or refined `has main energy input/output' axioms.

[stap-m]

For the record, these are the current inferred subclasses:

[stap-m]

The classes are not inffered correctly. There seems to be a problem with the not operation.

[stap-m]

However, we have to check whether existing inferred subclasses need then additional or refined `has main energy input/output' axioms.

Would it make sense to add them to all energy transformations anyway, where being determinable?

[stap-m]

The classes are not inffered correctly. There seems to be a problem with the not operation.

I slightly changed the equivalence definition, see above, for electricity generation process. The axioms of its inferred subclasses have to be changed in the following pattern:

• wind energy transformation:
  • input: 'has main energy input' **only** 'wind energy'
  • output: 'has main energy output' **some** 'electrical energy'

[stap-m]

To be discussed:

• Not inferred (correctly) as electricity generation process:
  • combined heat and power generation because it has no main energy input defined
  • electrical energy transfer: the new definition does't allow the inference any more, I guess thats ok.
• Not inferred as heat generation process:
  • steam reforming: thats correct I guess. I'd change the axiom from 'has energy output' some 'waste thermal energy' to EITHER 'has waste energy output' some 'thermal energy' OR 'has waste energy output' some 'waste thermal energy'
• I am hesitating with proposed equivalence definition of heat generation process: The increase of thermal energy performed by a heat pump would be excluded. Is that what we want? We could instead just define it like that: 'energy transformation' and ('has main energy output' some 'thermal energy')

[l-emele]

• I am hesitating with proposed equivalence definition of heat generation process: The increase of thermal energy performed by a heat pump would be excluded. Is that what we want? We could instead just define it like that: 'energy transformation' and ('has main energy output' some 'thermal energy')

A heat pump does not generate thermal energy, it transfers thermal energy from on temperature level to another.

[l-emele]

I am very unhappy about the only solution, as 'has main energy output' only 'electrical energy' can also mean it has no energy output, see http://protegeproject.github.io/protege/class-expression-syntax/

I played a bit around with the equivalent axiom. If we axiomatise electricity generation process as

'energy transformation'
  and ('has main energy output' some 'electrical energy')
  and ('has main energy input' some energy)
  and ('has main energy input' only (energy or not 'electrical energy'))

it works without having use only everywhere.

This is adapted from the solution @jannahastings had for diesel vehicle where we had a similar problem, see #1021

[stap-m]

I thought about adding the combined some and only operators to the subprocesses itself, but that would have been ugly. But this solution is neat! I'll try it out. However, this does not fix the problem that combined heat and power is not inferred as electricity generation process, see my comment above. Any idea?

[l-emele]

I see the problem with combined heat and power -- and it is not an implementation, but a problem in our logic.

1. We want to define electricity generation process as some process that has only electrical energy as main energy output.
2. We want to define heat generation process as some process that has only thermal as main energy output.
3. We want to define combined heat and power generation process as some process that has not only electrical energy nor only thermal energy as main energy output, but has both energies as main energy output.

As combined heat and power generation process has two main outputs it cannot be a subclass of processes that can have only one main output.

If we want to have combined heat and power generation process as subclasses of electricity generation process and heat generation process we have to circumnavigate in this case the only condition. We might do that by having those two axioms for combined heat and power generation process:

'has energy output' some 
    ('electrical energy' and 'thermal energy')
'has main energy output' some
    ('electrical energy' or 'thermal energy')

Those two axioms combined basically say: The combined heat and power generation process has two energy outputs which are electrical and thermal energy (first axiom) and one of them is the main energy output but we do not decide which of the two energy outputs is main (second axiom). With this it should work.

To be honest, it feels a bit like a trick for this class. But having a trick at one class still feels to much better than having the only axioms at a lot of classes which all say, this process might have this output or might have it not.

[stap-m]

Ok, now I am mixing up stuff, sorry! 🙈 But at least we don't need any tricks...

  * `combined heat and power generation` because it has no main energy input defined

For CHP, output is not the problen, but the missing input definition. We didn't specify them yet. Probably, there other main energy inputs than the classic combustion fuels. Then we might need subprocesses.

[l-emele]

Okay, if the missing main energy input is the problem, then we can simply add to combined heat and power generation the axiom: 'has main energy input' some (energy and not 'electrical energy' and not 'thermal energy') which basically says: It has a main energy input, but that is neither electrical energy nor thermal energy.

[stap-m]

I replaced the only-axioms as proposed. The hierarchy is inferred correctly. Yet, heat transfer is still inferred as heat generation process and electrical energy transfer as electricity generation process, as before. That contradicts the original idea of the issue.

[stap-m]

Obviously, we are kind of stuck here.

I would like to rethink the two concepts of "specific" energy transformation and energy transfer. Maybe it is not the right approach to define them exclusively, such that you have either a tranfer or a transformation, In many use cases we have both, see the example of charging in #1622.

For specific energy transformations, the focus on the transformation of one kind of energy into another might be suitable. However, for energy transfer, it might be more senible to look at the sources and sinks instead, from which to which the energy is spatially transmitted.

I'd like to discuss this in the next oeo-dev-meeting.

[chrwm]

@stap-m : I think, the first part (adapting electricity generation process and heat generation process) can be implemented immediately, right?

Just stumbled upon this issue now and it seems you've implemented this already. I just want to point out that in order to run the eletricity generation process one needs eletrical energy as an input. Eventhough it's not the main input.

[stap-m]

@stap-m : I think, the first part (adapting electricity generation process and heat generation process) can be implemented immediately, right?

Just stumbled upon this issue now and it seems you've implemented this already. I just want to point out that in order to run the eletricity generation process one needs eletrical energy as an input. Eventhough it's not the main input.

Sure, this is why we introduced the distinction between main and auxillay energy input. Anyway, the implementation is only in the PR and not merged, and still not satisfying, see comment history.

Any comments on my latest proposal @chrwm @l-emele @areleu and @OpenEnergyPlatform/oeo-domain-expert-energy-modelling?

[l-emele]

I would like to rethink the two concepts of "specific" energy transformation and energy transfer. Maybe it is not the right approach to define them exclusively, such that you have either a tranfer or a transformation, In many use cases we have both, see the example of charging in #1622.

For specific energy transformations, the focus on the transformation of one kind of energy into another might be suitable. However, for energy transfer, it might be more senible to look at the sources and sinks instead, from which to which the energy is spatially transmitted.

Sounds reasonable.

[l-emele]

From OEO dev meeting 65: Distinction of energy transformation and energy transfer makes no sense to us. Speak to OEKG developpers, whether the distinction between energy transformation and energy transfer is still relevant as we now have the technology subclasses.

[stap-m]

We should still try to axiomatize the spatial difference between input and output for energy transfer.

[stap-m]

We may consider out energy transfer function and energy transformation function.

[UStucky]

I think I can remember this discussion took place in our meeting a few months ago, when I had other ideas ;-) Indeed, there is a significant physical distinction (and also technical btw), this is approved by Wikipedia definitions right in their first sentences (in German for Energieübertragung and Energieumwandlung redirected to Energiewandler; in English energy transfer -only Simple English for some odd reason concerning a special physical usage of the term - and energy transformation, respectively). So, energy transfer is not only distinct from energy transformation, it is also not a subclass of it, as it is currently the case in the ontology. Moreover, the definition of transformation is very sparse, in my opinion it also applies to all processes, I think this should be thoroughly discussed.

[l-emele]

Moreover, the definition of transformation is very sparse, in my opinion it also applies to all processes, I think this should be thoroughly discussed.

See also issue #1654 where I proposed to make transformation an equivalent class.

[UStucky]

Thanks Lukas vor the tip with issue #1654. A few weeks ago I wrote down my arguments on the subject transformation and its sublasses in case there might be another round of discussion. I had a similar idea but meanwhile developed some doubts, which may be due to already open questions. Maybe the discussion on both, this issue here and #1654, should be continued a little bit more.