Wind Energy Analytics
Summary
Explored support vector machines for hind casting wind speed data for wind farms and created an open-source library for wind energy analysis.
The story
What started as my master’s thesis on wind resource assessment turned into first open source python library for wind energy analytics. Working at Brightwind Analysis we were looking at ways to predict the wind energy at a site using machine learning. It is necessary to predict it accurately for the next twenty years before setting up a wind farm to evaluate financial viability of setting up a farm. We ended up writing enough code to feel the need to open source it so other analysts can save time using it. It ended up being a popular project with many adopting the tool for their everyday wind energy assessment.
The documentation of the library can be found here.
Some cool things the library can do:
Advanced plots for wind analysts
Fig. 1: Wind rose for turbulence calculation.
Analyzing different types of wind distribution
Fig. 3: Wind distribution analysis
Shear calculation
Fig. 2: Shear profile calculation
A suite of wind speed forecasting methods, including ones proposed in my thesis
Fig. 4: Simple linear regression model for wind speed prediction