[Feature] Link inverter and rectifier capacity

[remmyd]

Currently the rectifier and inverter capacity are optimized individually eventhough they should represent a bi-directional inverter. Linking them together could enhance offgridders. What would be a good way to start on this issue?

[smartie2076]

Hello @remmyd! Glad to hear from you. I agree that this is not an ideal representation (and related to #31).

I have already noted down some hints in the MVS issues (rl-institut/multi-vector-simulator#24 (comment)) already:

I might have found an inbuild-functionality to connect two inverter assets: https://oemof-solph.readthedocs.io/en/latest/usage.html#adding-additional-constraints

Coupling of two variables e.g. investment variables) with a factor -> equate_variables()

Also see here and here (Code).

So, I think you did not work with oemof yet, right? This will be a jump into cold water for you. This is how I would go about it:

• Read the documentation of equate_variables()
• Try and simulate the example code from above including the constraint, as well as an oemof-model without that constraint. I am not sure if you want to do that, as that would mean that requires you dive into oemof a bit (but in a managable way). Look at the lp-file of your code snippet and check how the constraint looks there.
• Create an abstract frunction that calls the equate_variables() in G2b_constraints_custom.py
• Figure out whether a bi-directional inverter is always the goal if inverter and rectifyier are activated, or if there are reasons to explicitly want two uni-directional transformers. Think about how / where you want future users to enable/disable the bi-directional asset constraint.
• Add function call to the G2b function after

  offgridders/src/G1_oemof_create_model.py

  Line 307 in a8fa6f1

  # ------------Rectifier AC DC------------#

  to

  offgridders/src/G1_oemof_create_model.py

  Line 355 in a8fa6f1

  )

  , or place it in a loop within those lines
• Add function call to the G2b function after

  offgridders/src/G1_oemof_create_model.py

  Line 181 in a8fa6f1

  # ------------ main grid bus and subsequent sources if necessary------------#

  to

  offgridders/src/G1_oemof_create_model.py

  Line 246 in a8fa6f1

  )

  , or place it in a loop within those lines
• Think of a pytest for the user input turning off/on the bi-directionality constraint
• Think of a pytest making sure that the bi-directionality constraint works as intended

... as you see, this is already a pretty large issue. You can learn basically everything you need to program a tool like Offgridders, but it requires a lot from you. Depending on your learning style, you might want to adress a smaller issue first, to get to know the github flow first.

[smartie2076]

I have heard you are not sure how to share code with me. There are actually different options:

1. Directly share the code-snippet in CODE markdown format in the post (three apostrophes around the code, see "Insert Code <ctrl +e>"). This will make it easily quotable.
2. If your code is too long for that, and you do not expect much feedback anyway, share it as an attachment to one of your posts (not sure if it can be a .py file, if not, change the format by renaming it to a .txt file)
3. If you need extensive feedback on your code, eg. as your have bugs in your code, AND you are planning to make a PR later on anyway, you can also already open a PR. Att the file in the root of the repository/project. Then you commit that file separately from any other changes and push. The goal then is to create a working code together, and then move snippets into the real src code folder and remove the working file later on. Make sure you create a draft PR, not a real PR!

[remmyd]

Hi @smartie2076, thanks for the reply. I used parts of test_connect_invest.py from the code example you shared above. I'll attach it to this post. The equal_variables() function does set the two transformers to the same capacity ('invest' variable for transformer 2 is set equal to transformer 1 invest variable eventhough transformer 2 is not used). If the function is commented out the transformer 2 variable is set to 0. Thus, it seem to do what we want. The constraint in the lp-file looks like the following:

c_e_equate_InvestmentFlow_invest(powerline_1_2_electricity2)_InvestmentFlow_invest(powerline_2_1_electricity1)_: +1 InvestmentFlow_invest(powerline_1_2_electricity2) -1 InvestmentFlow_invest(powerline_2_1_electricity1) = 0

test_oemof_equate_variables_constraint.txt

[smartie2076]

This is looking good then! I think we would implement

# define aliases for shorter calls below
line12 = energysystem.groups['powerline_1_2']
line21 = energysystem.groups['powerline_2_1']

# equate investment flows.
solph.constraints.equate_variables(om,om.InvestmentFlow.invest[line12, bel2], om.InvestmentFlow.invest[line21, bel1])

A bit differently - I think I called the investmentFlows differently in Offgridders:

offgridders/src/G2b_constraints_custom.py

Line 891 in a8fa6f1

CAP_inverter += model.InvestmentFlow.invest[el_bus, inverter]

(This is just one example for a call for an inverter).

probably, your function would look more like this:

def constraint_equate_bi-directional_transformer_capacities(model, case_dict, el_bus, dc_bus, transformer_in, transformer_out, name_transformer_in, name_transformer_out):
        CAP_inverter = 0 
        CAP_rectifyer = 0
        if case_dict[name_transformer_in] != None and case_dict[name_transformer_out] != None:
            if case_dict[name_transformer_in] is False and case_dict[name_transformer_out] is False:
                  CAP_inverter += model.InvestmentFlow.invest[el_bus, inverter]
                  CAP_rectifyer += model.InvestmentFlow.invest[dc_bus, rectifyer]
       solph.constraints.equate_variables(model,CAP_inverter, CAP_recitfyer)

Notes:

• I am using a generalized function with name_transformer_in and transformer_in etc. here because then you can apply the function to both the PCC and the bi-directional inverter by just calling it twice).
• I think it is called the dc_bus for the rectifyer, but I am not sure. It is the bus connected to the PV panel/battery.

[smartie2076]

Ah, and of couse you need an if-loop somehow that would make sure that you only apply the constraint when appropriate.

[smartie2076]

Do you have an answer to the following question yet?

• Figure out whether a bi-directional inverter is always the goal if inverter and rectifyier are activated, or if there are reasons to explicitly want two uni-directional transformers. Think about how / where you want future users to enable/disable the bi-directional asset constraint.

[remmyd]

Thanks for the hints. I think letting the user decide if they want to link the inverter and the rectifier together is a good idea. I also saw in this publication that the inverter and rectifier can be linked by a capacity ratio factor, and if I am not mistaken Homer also defines the rectifier as a % of the inverter. However, I am unsure on how we would implement this.

For the if-loop you mentioned would you call it outside of the function then in G1?

[smartie2076]

Yeah, Homer also does that: https://www.homerenergy.com/products/pro/docs/3.10/converter.html

Basically, you would also add a parameter like inverter_rectifyer_capactiy_ratio_factor, and then the user can define if they assume it to be 1 or <1. Then, use

CAP_inverter_with_capacity_ratio = CAP_inverter * inverter_rectifyer_capactiy_ratio_factor
solph.constraints.equate_variables(model,CAP_inverter_with_capacity_ratio , CAP_recitfyer)

This results in 0=CAP_rectifyer-CAP_inverter*inverter_rectifyer_capactiy_ratio_factor or CAP_rectifyer=inverter_rectifyer_capactiy_ratio_factor*CAP_inverter

[smartie2076]

I did call the if-loops for the other constraints in G1, see:

offgridders/src/G1_oemof_create_model.py

Line 432 in a8fa6f1

# ------------Renewable share constraint------------#

You can follow that approach for now.

In the end, we will still have to figure out how to easiest add pytests for the new constraints, and then you might need to move the if-loop (but that is no big thing!).

[smartie2076]

CAP_inverter_with_capacity_ratio = CAP_inverter * inverter_rectifyer_capactiy_ratio_factor
solph.constraints.equate_variables(model,CAP_inverter_with_capacity_ratio , CAP_recitfyer)

This results in 0=CAP_rectifyer-CAP_inverter*inverter_rectifyer_capactiy_ratio_factor or CAP_rectifyer=inverter_rectifyer_capactiy_ratio_factor*CAP_inverter

I ment in the LP-file, you could also try that out with your code snippet! That would even be the best approach, actually :)

[remmyd]

Yes, that makes, sense. equal_variables() actually has the option to use a factor:

equate_variables(model, CAP_rectifier , CAP_inverter, inverter_rectifier_capactiy_ratio_factor)

which result in the same constraint: CAP_rectifier==inverter_rectifier_capactiy_ratio_factor*CAP_inverter (tested with the code snippet)

I guess this would be useful if the user gives only an inverter cost. The options now would be:

1. let the user give inverter costs and use a inverter_rectifier_capactiy_ratio_factor
2. let the user give bidirectional costs and set inverter and rectifier capacity to the same value
3. let the user give inverter and rectifier costs and treat them as two uni-directional transformers

I am usure if they all make sense or if we should decide on which one to use.