Skip to content

tx123bt_workflow.md

Latest commit

 

History

History
141 lines (95 loc) · 3.88 KB

tx123bt_workflow.md

File metadata and controls

141 lines (95 loc) · 3.88 KB

TX-123BT Reference Workflow

This document describes the optional TX-123BT reference workflow bundled with GridForge.

It is useful if you want a complete public-data example end to end, but it is not required for the core GridForge workflow.

What this workflow does

The TX-123BT reference pipeline:

  1. preprocesses the public TX-123BT raw data into one CSV per bus,
  2. optionally sanity-checks the generated bus CSV files,
  3. assigns those bus CSVs to a generated GridForge case,
  4. rescales load, solar, and wind to match the case configuration.

The Python entry points live in:

Dataset source

GridForge uses the open-source TX-123BT system and the paper A synthetic Texas power system with time-series weather-dependent spatiotemporal profiles.

Download the raw dataset from:

Place the downloaded zip file under:

  • data/raw_data.zip

Fastest path

The shortest way to run the full preprocessing pipeline is:

bash scripts/generate_tx123bt_bus_data.sh

This script will:

  • unzip data/raw_data.zip if needed,
  • run preprocess_tx123bt_raw_data(...),
  • run a small random sanity-check pass.

If you also want to remove the unneeded TX-123BT raw folders after a successful run:

KEEP_RAW_DATA=0 bash scripts/generate_tx123bt_bus_data.sh

Manual preprocessing path

If you want to run the steps manually, first unzip the raw data:

unzip -n ./data/raw_data.zip -d ./data/

Then run:

import numpy as np
from tqdm import tqdm
from gridforge.reference_data.tx123bt import (
    preprocess_tx123bt_raw_data,
    sanity_check_tx123bt_bus_csv,
)

preprocess_tx123bt_raw_data()

# Optional sanity check on a random subset of buses
no_day = 365
no_bus = 123
np.random.seed(42)
bus_idx_list = np.random.randint(1, no_bus + 1, size=20)
for bus_idx in tqdm(bus_idx_list, desc="Sanity checking per-bus files"):
    sanity_check_tx123bt_bus_csv(bus_idx=bus_idx, no_day=no_day)

Preprocessing outputs

The preprocessed bus CSV files are written under:

  • data/bus_data/

Each file is named:

  • bus_<idx>.csv

Each bus CSV contains:

  • calendar/context features
  • weather features
  • load
  • solar
  • wind

Each bus has at most one renewable plant (solar or wind) in this reference pipeline.

Prepare bus data for a generated case

After generating a GridForge Excel case workbook, assign and rescale the preprocessed TX-123BT bus data like this:

from gridforge.reference_data.tx123bt import construct_tx123bt_grid_data

data_dir = "14bus_data"
verbose = 0
construct_tx123bt_grid_data(config_path_xlsx, data_dir, random_seed, verbose=verbose)

This helper will:

  • assign preprocessed data from data/bus_data/ to each bus in the generated case,
  • rescale load, solar, and wind to match the case capacities,
  • save the resulting per-bus CSV files in the target output directory.

Use the same random_seed as construct_grid_config(...) if you want reproducible case/data alignment.

Notes

  • Preprocessing may take several minutes because the raw dataset is large.
  • After the first successful preprocessing run, you can reuse data/bus_data/ directly.
  • The TX-123BT pipeline is intentionally source-specific and keeps the fixed load / solar / wind convention.

Related files