|
|
| from dataclasses import dataclass |
| from typing import Optional |
| import pandas as pd |
| import numpy as np |
|
|
| @dataclass(frozen=True) |
| class DayInput: |
| """ |
| Data class representing the input data for a single day of optimization. |
| For this project, we used perfect foresight, meaning that |
| the prices are known in advance for the entire day. |
| In practice, one could input forecasted prices, but it's a topic on its own. |
| |
| index_ts : Timestamps for each time step in the day. |
| T : Number of time steps in the day. |
| dt : Duration of each time step in hours. |
| price_energy : Energy prices for each time step. |
| price_fcr : FCR prices for each time step. |
| price_afrr_up : aFRR up prices for each time step. |
| price_afrr_down : aFRR down prices for each time step. |
| soc0 : Initial state of charge as a fraction of the battery capacity. |
| """ |
| index_ts: pd.DatetimeIndex |
| T: int |
| dt: float |
|
|
| price_energy: np.ndarray[float] |
| price_fcr: np.ndarray[float] |
| price_afrr_up: np.ndarray[float] |
| price_afrr_down: np.ndarray[float] |
|
|
| soc0: Optional[float] = 10.0 |
|
|
| @classmethod |
| def from_df(cls, |
| day_df: pd.DataFrame, |
| config: dict, |
| soc0: Optional[float] = 10.0) -> "DayInput": |
| """ |
| Infers DayInput from a DataFrame for a single day. |
| |
| day_df : DataFrame containing the data for a single day. The index should be Timestamps. |
| config : Configuration dictionary. |
| soc0 : Initial state of charge as a fraction of the battery capacity |
| |
| """ |
| return cls( |
| index_ts=day_df.index, |
| T=len(day_df), |
| dt=(day_df.index[1] - day_df.index[0]).total_seconds() / 3600.0, |
| price_energy=day_df[config["columns"]["energy"]].to_numpy(), |
| price_fcr=day_df[config["columns"]["fcr"]].to_numpy(), |
| price_afrr_up=day_df[config["columns"]["afrr_up"]].to_numpy(), |
| price_afrr_down=day_df[config["columns"]["afrr_down"]].to_numpy(), |
| soc0=soc0, |
| ) |
|
|
