Input Folder Structure

All input files must be placed inside the epm/input folder.

Each dataset lives in its own subfolder, which is passed to the model with --folder_input.

A config.csv file maps model parameters to the corresponding .csv files; the path of this file is passed with --config from --folder_input. See Configuration for details on how config.csv works.

Related Documentation:

• Configuration Flow - How config.csv, pSettings, and resources work together

• Input Overview - Summary of all input files

• Input Description - Detailed parameter descriptions

• Resources - Shared model constants (technologies, fuels, carbon content)

• Input Treatment - Automatic data validation and filling

Below is the current baseline structure for the data_test dataset.

To run the model with Python:

python epm.py --folder_input data_test --config config.csv

data_test/
│
├── config.csv                          # Main configuration file for baseline run
├── pHours.csv                          # Time-slice definitions
├── pSettings.csv                       # Default simulation settings
├── pSettings_NoTransmissionExpansion.csv
├── scenarios.csv                       # Scenario definitions (e.g., NoTransmissionExpansion, OptimalExpansion)
├── sensitivity.csv                     # Optional sensitivity inputs referenced by `sensitivity/`
├── y.csv                               # Year list (full horizon)
├── zcmap.csv                           # Zone-country mapping
│
├── constraint/                         # Policy and emissions constraints
│   ├── pCarbonPrice.csv
│   ├── pEmissionsCountry.csv
│   ├── pEmissionsTotal.csv
│   └── pMaxFuellimit.csv
│
├── h2/                                 # Hydrogen-specific parameters
│   ├── pAvailabilityH2.csv
│   ├── pCapexTrajectoryH2.csv
│   ├── pExternalH2.csv
│   ├── pFuelDataH2.csv
│   └── pH2DataExcel.csv
│
├── load/                               # Load and demand data
│   ├── pDemandData.csv
│   ├── pDemandForecast.csv
│   ├── pDemandProfile.csv
│   ├── pEnergyEfficiencyFactor.csv
│   └── sRelevant.csv
│
├── reserve/                            # Reserve requirements
│   ├── pPlanningReserveMargin.csv
│   ├── pSpinningReserveReqCountry.csv
│   └── pSpinningReserveReqSystem.csv
│
├── sensitivity/                        # Reduced input files listed in `sensitivity.csv`
│   └── y_reduced.csv
│
├── supply/                             # Generation and availability data
│   ├── pAvailabilityCustom.csv
│   ├── pAvailabilityDefault.csv
│   ├── pCSPData.csv
│   ├── pCapexTrajectoriesCustom.csv
│   ├── pCapexTrajectoriesDefault.csv
│   ├── pFuelPrice.csv
│   ├── pGenDataInput.csv
│   ├── pGenDataInputDefault.csv
│   ├── pStorageDataInput.csv
│   ├── pVREProfile.csv
│   ├── pVREgenProfile.csv
│   └── sensitivity/
│       └── pGenDataInput_linear.csv
│
└── trade/                              # Cross-border trade and transmission
    ├── pExtTransferLimit.csv
    ├── pLossFactorInternal.csv
    ├── pMaxAnnualExternalTradeShare.csv
    ├── pMaxPriceImportShare.csv
    ├── pMinImport.csv
    ├── pNewTransmission.csv
    ├── pNewTransmission_optimal.csv
    ├── pTradePrice.csv
    ├── pTransferLimit.csv
    └── zext.csv

Scenario and Sensitivity Files

• scenarios.csv: Lists the core scenario names and the specific files that each scenario should pull from the dataset.

• sensitivity.csv: Points to alternative data sources (e.g., reduced year sets) contained in the sensitivity/ directories at the dataset root or within specific subfolders such as supply/sensitivity/.

Data Package Schema (datapackage.json)

The epm/input/datapackage.json file defines the canonical schema for all EPM input datasets following the Frictionless Data Package specification. This file:

• Describes the structure and types of each input CSV file

• Defines foreign key relationships between resources (e.g., zones reference zcmap.z)

• Specifies whether data is stored in “wide” or “long” format

• Documents the expected columns, data types, and units for each parameter

This schema is useful for:

• Validation: Ensuring input data conforms to expected structure

• Documentation: Understanding relationships between input files

• Tooling: Building automated data pipelines that read EPM inputs

Example resource definition from datapackage.json:

{
  "name": "pDemandForecast",
  "path": "load/pDemandForecast.csv",
  "format": "csv",
  "schema": {
    "fields": [
      {"name": "z", "type": "string"},
      {"name": "type", "type": "string"},
      {"name": "year", "type": "integer"},
      {"name": "value", "type": "number"}
    ],
    "foreignKeys": [
      {"fields": ["z"], "reference": {"resource": "zcmap", "fields": ["z"]}}
    ]
  }
}