data_format.md

Data format description

• study_list
• dataset
• dataset_revenues
• investment_metrics
• data_outputs
• decisions

Let's have a look across all the different data structure, the tree structure is easily accessible with a "$" symbol:

• study_list is obtained by the the mother function build_study_list() which creates the list to easily store each study_id, output of each simulation launched by the investment model. This automatically generated based either on api_study_list or local_study_list depending on your use_api parameter. The list is organised as follows :

  • year_simulated
    • reference_id study_id of the reference simulation. This is crucial for the investment to work properly with antares web. Locally study_id = year_simulated
    • variant_id study_id of the variant created. This is the simulation which will be modified.

• power_plants is obtained by the function make_powerplants_from_antares() which creates the list of the powerfleet and all its description and the hypothesis costs by clusters. The list is organised as follows :

  • year_simulated
    • area

• dataset is obtained by the the mother function build_dataset() which reads and format data after the Antares runs on the years simulated. The list is organised as follows :

  • year_simulated
    • area
      • clusters_filters
        • not_supported vector of the clusters not supported, those who get their revenues only from market mecanisms
        • supported vector of the cluster supported, and get their revenues from other mecanisms
        • candidates_investments vector of the clusters candidates to the investment decision
        • candidates_retirement vector of the clusters candidates to the retirement decision
      • data_antares a list of the outputs obtained from the runs of antares simulation, for example you could find :
        • thermal_availability
        • areas
        • cluster_description
        • links
        • etc ...

• dataset_revenues is obtained after the computation of the revenues from the energy only market and capacity market simulation. It is organised as follows. When it is not precised, the revenues are expressed in euros/MW :

  • year_simulated
    • area
      • market_info_supply a dataframe with all the market information of the supply side (revenues, bids on capacity market, capacity ..)
      • auction_info a dataframe with the capacity market auction information (clearing price, marginal bid ..)
      • revenues a list of dataframe, keys are the clusters.
        • revenues_annual available for heat, energy and system service revenues
        • revenues_hourly available for heat and energy revenues
        • production a dataframe with the hourly production as of the different montecarlo years

• investment_metrics is obtained after reformatting and extracting data from the dataset_revenues variable.

  • area
    • metrics the dataframe with the investment metric computed on simulated years for each cluster
    • revenues_total_net dataframe with the total net revenues for each cluster on each years (energy only + capacity market operational costs). Used for decommissioning
    • policy_retired vector of clusters names that are already decommissionned by the antares simulation dataset.

• data_outputs data outputs from the model. This a record of outputs variable after each loop, for each year.

  • area
    • year_simulated
      • loop number the loop number on the year_simulated
        • spot the averaged annual spot price
        • power_plants the fleet of power_plants
        • LOLD loss of load duration
        • revenues the annual revenues for each cluster.
        • ...

• decisions retirement/investment decisions taken by the model. If continue = 1, the model continues to loop on the same actual year.power_plants_moves is the record of every moves of the model. With useful information related to it.