EPW Wind Rose Generator

Generate wind rose charts from EnergyPlus Weather (EPW) files, styled with a CFD-contour-like colour range.

The script scans a folder of .epw files, plots a wind rose for each one using the hourly wind direction and wind speed records, and saves the charts as PNG images named after the year of the weather data.

Features

• Batch-processes every .epw file in the input folder.
• Reads wind direction and wind speed directly from the EPW hourly data rows.
• Automatically filters out EPW missing values (e.g. wind speed flagged as 999).
• Uses the jet colormap (blue → cyan → green → yellow → red) for a CFD contour look.
• Names each chart Wind rose <year>.png, where the year is read from the EPW data rows.

Requirements

• Python 3.8+
• windrose
• matplotlib
• pandas
• numpy

Install the dependencies with:

pip install windrose matplotlib pandas numpy

Folder structure

wind-rose-gen-test/
├── generate_wind_roses.py
├── README.md
├── input_files/        # put your .epw files here
└── output_files/        # generated wind roses are saved here (created automatically)

Usage

1. Place your .epw weather files inside the input_files/ folder.

2. Run the script:

   python generate_wind_roses.py

3. Find the generated wind roses in the output_files/ folder, named Wind rose <year>.png (for example, Wind rose 2023.png).

How it works

EPW files contain 8 header lines followed by 8,760 hourly data rows. The script:

1. Skips the 8 header lines and reads the comma-separated data rows.
2. Extracts the year (column 1), wind direction (column 21), and wind speed (column 22).
3. Removes any rows with missing/invalid wind values.
4. Determines the chart year using the most frequent year in the data (robust for both single-year actual-year files and mixed-year TMY files).
5. Plots the wind rose with frequency rings shown as percentages and a wind-speed legend (m/s).

Configuration

A few settings can be adjusted near the top of generate_wind_roses.py:

Setting	Description	Default
NUM_SPEED_BINS	Number of wind-speed bands (more bands = smoother gradient)	8
DPI	Output image resolution	200
INPUT_DIR	Input folder for .epw files	input_files
OUTPUT_DIR	Output folder for charts	output_files

License

MIT