After the MVS simulation is finished (with the inputs discussed in #605), the EPA is to be supplied with its results. We have to make sure that the EPA can read the provided output file.

ToDos:

• Identifiy which results should be provided to the EPA
• Provide the output files of tests/inputs (same example as in [User interface] MVS/EPA data exchange: Inputs of the MVS #605)
• Propose a clean structure for the file provided to the EPA with the results (@vicalk)
• Decide if we have to write a parser to convert the initial json output of the MVS to the file that EPA wants to receive
• Decide and implement a compression method

Summary of discussion points

1. There are several parameters (as well as assets) names ending with “_excess”. Are these assets provided in the input of MVS or are they generated by the simulation process?

• @smartie2076 - These parameters are auto-generated. Every bus needs an excess sink, and the parameters (and assets) with "_excess" relate and stem from that.

2. After closely inspecting the JSON file (~29MB), we observed timestamps are the same for all timeseries. Is this true (i.e. all timeseries refer to the same time period or is it possible for certain asset to have different timestamps)? If yes, we can significantly reduce the size and redundancy of the file by removing all timestamps vectors from the file. The period can be deduced by simulation settings.

• @smartie2076 - This is true! @Bachibouzouk what do you think of reducing this?

• @Bachibouzouk - will do

3. We also spotted attributes which do not exist in the EPA side (e.g. costs_investment_over_lifetime). Are such attributes generated after MVS completes a scenario simulation?

• @smartie2076 - Those attributes are resulting from the pre-processing, ie. before optimizing the model. Do you think this is information that you would like to have for the EPA?

• @Gr3at - I think it depends on whether the end user needs such information.

• @smartie2076 - Then I would leave it out for now and add it upon end-user request. When discussing with them, we can prepare a list of possible amendmends.

4. We noticed that there is an “EnergyBusses” section. Is this the same as the one provided as input by EPA or is this generated by a preprocess function of MVS?

• @smartie2076 - Currently, this is the result of a pre-process step of the MVS. However, it is better if the EPA provides the MVS with this information, and we change the MVS so that it understands this. The way it is now is sometime implicit and prone to errors.

5. Data which are NOT modified by MVS (e.g. economic_data, EnergyConsumption{development_costs, age_installed, optimizedCap, input timeSeries, etc.}) shall not be send back to EPA to prevent redundancy.

• @smartie2076 - Ok. But you want the pre-processed data, eg. CRF, annuity factor, replacement costs per unit over lifetime...?

• @Gr3at - If they are to be provided in the results visualization page, then yes.

• @smartie2076 - Same as point 3, lets add it to the list of possible amendments.

6. Could you explain what fixed_costs dictionary represent and whether we need to present it as a result to the end-user?

• @smartie2076 - This is actually not used / included in the MVS currently. It would be fixed project costs that do not depend neither on assets nor on capacities to be installed. They might differ from scenario to scenario (eg. if suddenly an energy manager has to be employed), but are of no importance for the optimization.

7. “path_to_plots” seem to be referencing plotly dash generated figures. We can omit it.

• @smartie2076 - I agree

8. cost_matrix, scalar_matrix

• @smartie2076 - Do you want to omit this or do you want to know what this is? This summarizes data that is availablie for the assets individually, eg. capacities, peak flow, costs...

9. optimizedFlows – how do they differ from the flows presented in the previous sections?

• @smartie2076 - They do not differ from the flows in the assets. However, they list the flows with timeseries on each bus, making it easier to plot data then when always looping though all the energyAssets

• @Gr3at - If the information is duplicate in the Json file it might be better to keep one instance.
  In any case we will parse the Json and save required data in the database.

• @smartie2076 - I think @Bachibouzouk will think about how to make the json output more efficient.

• @Bachibouzouk - I will remove it as you already save things in the database. data processing clientside might not be very long anyway. Otherwise we can discuss in which form we should pack the results to send them back to you

10. Is the mapping table provided in Appendix correct? Please fill in the missing parts and highlight any errors.

• @smartie2076 - I will answer this part later, hopefully tomorrow or on Monday.