Profile Generator

The engine includes a profile generator that creates synthetic solar (PV) and wind production timeseries from weather data via the atlite library. The underlying data combines SARAH-3 satellite irradiance with ERA5 reanalysis (wind, temperature). Generated profiles are stored in SQLite and can be referenced in simulations by UUID.

Overview

Profiles can also be imported from CSV files instead of generated from weather data.

Profile Types

PV (Solar)

Solar profiles model photovoltaic generation using temperature-dependent panel efficiency (Huld 2010 model). The generator supports three panel technologies and multiple mounting options.

Parameter	Type	Default	Description
`latitude`	float	required	Site latitude (-90 to 90)
`longitude`	float	required	Site longitude (-180 to 180)
`capacity_mw`	float	required	DC peak power (MW)
`panel_type`	string	`"CSi"`	Panel technology: `CSi`, `CdTe`, `KANENA`
`slope`	float	`35.0`	Panel tilt angle (0-90 degrees)
`azimuth`	float	`180.0`	Panel orientation (degrees, 180 = South)
`tracking`	string	`"fixed"`	Mounting type (see below)
`latitude_optimal`	bool	`false`	Auto-optimize tilt for latitude
`system_loss`	float	`0.0`	System losses as fraction (0-1)
`inverter_limit`	float	—	AC clipping limit (MW)
`inverter_efficiency`	float	`0.9`	Inverter efficiency (0-1)
`clearsky_model`	string	`"simple"`	Clearsky model: `simple` or `enhanced`

Tracking types:

Value	Description
`fixed`	Fixed tilt at specified slope and azimuth
`horizontal`	Single-axis horizontal tracking
`tilted_horizontal`	Single-axis tilted horizontal tracking
`vertical`	Single-axis vertical tracking
`dual`	Dual-axis tracking (follows sun in both axes)

Panel technologies:

Type	Description
`CSi`	Crystalline silicon — most common, good all-round performance
`CdTe`	Cadmium telluride — thin film, better in high temperatures
`KANENA`	Kaneka thin film — lightweight, lower efficiency

Example: Generate a solar profile

bash

curl -X POST http://localhost:8000/profiles \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Solar NL 10MW",
    "profile_type": "pv",
    "description": "10 MW fixed-tilt CSi in Netherlands",
    "preload": true,
    "cutout_year": 2024,
    "config": {
      "latitude": 52.0,
      "longitude": 4.9,
      "capacity_mw": 10.0,
      "panel_type": "CSi",
      "slope": 35.0,
      "azimuth": 180.0,
      "tracking": "fixed",
      "system_loss": 0.1,
      "inverter_limit": 9.0
    }
  }'

Wind

Wind profiles model turbine output using manufacturer power curves. The generator supports 21 pre-configured turbine models with hub height interpolation.

Parameter	Type	Default	Description
`latitude`	float	required	Site latitude (-90 to 90)
`longitude`	float	required	Site longitude (-180 to 180)
`capacity_mw`	float	required	Total farm capacity (MW)
`turbine_model`	string	`"Vestas_V112_3MW"`	Turbine model (see below)
`hub_height`	float	—	Hub height in meters (uses turbine default if omitted)
`interpolation`	string	`"logarithmic"`	Wind speed height interpolation: `logarithmic` or `power`
`smooth`	bool	`false`	Apply Gaussian smoothing to power curve
`system_loss`	float	`0.0`	System losses as fraction (0-1)

Available turbine models:

Onshore turbines

Model	Capacity
`Vestas_V112_3MW`	3 MW
`Vestas_V90_3MW`	3 MW
`Vestas_V80_2MW_gridstreamer`	2 MW
`Vestas_V66_1750kW`	1.75 MW
`Vestas_V47_660kW`	660 kW
`Vestas_V25_200kW`	200 kW
`Enercon_E126_7500kW`	7.5 MW
`Enercon_E101_3000kW`	3 MW
`Enercon_E82_3000kW`	3 MW
`Siemens_SWT_107_3600kW`	3.6 MW
`Siemens_SWT_2300kW`	2.3 MW
`Suzlon_S82_1.5_MW`	1.5 MW
`Bonus_B1000_1000kW`	1 MW
`NREL_ReferenceTurbine_2020ATB_4MW`	4 MW
`NREL_ReferenceTurbine_2020ATB_5.5MW`	5.5 MW
`NREL_ReferenceTurbine_2020ATB_7MW`	7 MW

Offshore turbines

Model	Capacity
`Vestas_V112_3MW_offshore`	3 MW
`Vestas_V164_7MW_offshore`	7 MW
`NREL_ReferenceTurbine_5MW_offshore`	5 MW
`NREL_ReferenceTurbine_2020ATB_12MW_offshore`	12 MW
`NREL_ReferenceTurbine_2020ATB_15MW_offshore`	15 MW
`NREL_ReferenceTurbine_2020ATB_18MW_offshore`	18 MW

Example: Generate a wind profile

bash

curl -X POST http://localhost:8000/profiles \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Wind NL 5MW",
    "profile_type": "wind",
    "cutout_year": 2024,
    "config": {
      "latitude": 52.0,
      "longitude": 4.9,
      "capacity_mw": 5.0,
      "turbine_model": "Vestas_V90_3MW",
      "hub_height": 80,
      "interpolation": "logarithmic",
      "system_loss": 0.05
    }
  }'

Generation Process

Generation takes 30-60 seconds per profile. The mini-cutout optimization (3x3 grid around the target location) significantly reduces processing time compared to loading the full European cutout.

Weather Data

Profiles are generated from cutout files that combine two data sources:

SARAH-3 (EUMETSAT CM SAF) — satellite-derived surface solar irradiance (direct and diffuse). Higher spatial resolution and more accurate for PV modeling than reanalysis products alone.
ERA5 (ECMWF) — reanalysis data providing wind speed (10m and 100m), temperature, roughness length, and other atmospheric variables. Used for wind profile generation and PV temperature correction.

Each cutout merges both datasets into a single NetCDF file covering one full year of hourly data across a European domain. The atlite library handles the data fusion and variable selection depending on the profile type being generated.

Year	File	Data Sources
2023	`europe-2023-sarah3-era5.nc`	SARAH-3 irradiance + ERA5 wind/temperature
2024	`europe-2024-sarah3-era5.nc`	SARAH-3 irradiance + ERA5 wind/temperature
2025	`europe-2025-sarah3-era5.nc`	SARAH-3 irradiance + ERA5 wind/temperature

Which data is used for which profile type:

Variable	Source	Used by
Surface solar irradiance (direct/diffuse)	SARAH-3	PV profiles
2m temperature	ERA5	PV profiles (temperature correction)
Wind speed (10m, 100m)	ERA5	Wind profiles
Surface roughness length	ERA5	Wind profiles (logarithmic interpolation)

List available cutouts via the API:

bash

curl http://localhost:8000/profiles/cutouts

Get metadata for a specific year (coverage area, resolution, time range):

bash

curl http://localhost:8000/profiles/cutouts/2024/info

Using Profiles in Simulations

Reference a profile by its UUID in the simulation run configuration. The profile data replaces any scenario-level solar or wind data.

json

{
  "params": {
    "solarInstalledCapacity": 10,
    "windInstalledCapacity": 5
  },
  "run": {
    "strategy": "rolling_lp",
    "profile_id": "a1b2c3d4-uuid",
    "wind_profile_id": "e5f6g7h8-uuid",
    "enabled_markets": ["day_ahead", "intraday"]
  }
}

Parameter	Description
`profile_id`	Solar production profile UUID (overrides scenario solar data)
`wind_profile_id`	Wind production profile UUID (overrides scenario wind data)

The simulation loads profile data from cache (or database on first access), converts it to numpy arrays, and uses it to model co-located renewable generation alongside battery trading.

Importing from CSV

Instead of generating from weather data, profiles can be imported from CSV files.

bash

curl -X POST http://localhost:8000/profiles/upload \
  -F "file=@solar_data.csv" \
  -F "name=Imported Solar" \
  -F "profile_type=pv" \
  -F "latitude=52.0" \
  -F "longitude=4.9" \
  -F "capacity_mw=10.0"

Accepted CSV columns:

Timestamp: timestamp, time, datetime, or date
Output: solar_output_mw, wind_output_mw, output_mw, power, or output

The importer auto-detects delimiter (, or ;) and resamples to 15-minute intervals if the input is at hourly or 30-minute resolution.

Profile Statistics

Every generated or imported profile includes computed statistics:

Metric	Description
`max_mw`	Peak output power
`mean_mw`	Average output power
`total_mwh`	Total annual energy production
`capacity_factor`	Average output / installed capacity
`full_load_hours`	Equivalent hours at full capacity
`peak_hours`	Hours with output > 50% capacity
`zero_hours`	Hours with zero output

Monthly breakdowns are also available, providing seasonal production patterns.

bash

# View profile metadata and statistics
curl http://localhost:8000/profiles/{profile_id}

Caching

Profiles are cached in memory for fast access during simulations. The cache uses an LRU (Least Recently Used) eviction policy with a configurable memory limit (default: 500 MB).

Preloading: Mark profiles with preload: true to load them into cache on API startup. Preloaded profiles are never evicted.

bash

# Enable preloading for a profile
curl -X POST http://localhost:8000/profiles/{profile_id}/preload \
  -H "Content-Type: application/json" \
  -d '{ "preload": true }'

# View cache statistics
curl http://localhost:8000/profiles/cache/stats

Managing Profiles

Operation	Endpoint	Method
Generate	`/profiles`	POST
Import CSV	`/profiles/upload`	POST
List	`/profiles`	GET
Get metadata	`/profiles/{id}`	GET
Update	`/profiles/{id}`	PUT
Delete	`/profiles/{id}`	DELETE
Get timeseries	`/profiles/{id}/data`	GET
Export CSV	`/profiles/{id}/export`	GET

See the API Reference for full endpoint details.

Profile Generator ​

Overview ​

Profile Types ​

PV (Solar) ​

Wind ​

Generation Process ​

Weather Data ​

Using Profiles in Simulations ​

Importing from CSV ​

Profile Statistics ​

Caching ​

Managing Profiles ​

Profile Generator

Overview

Profile Types

PV (Solar)

Wind

Generation Process

Weather Data

Using Profiles in Simulations

Importing from CSV

Profile Statistics

Caching

Managing Profiles