Project inspiration
Wind power dominates growth of world's energy sector. As well, wind is one of the fastest growing renewable energy sources of electricity in Canada.
Due to the natural and anthropogenic (caused by climate change) variability of wind, the integration of wind energy into electrical power systems is challenging due its intermittency. One approach to deal with wind intermittency is forecasting future values of wind power production.
Our inspiration was develop an artificial neural network to improve wind speed and wind power production forecasting in order to provide better information to improve energy supply for a market for which demand is increasing.
What it does
Our model predicts short-term wind power production from wind speed data at 5 minute intervals for a 6.5 hour period.
How we built it
We build with an artificial neuronal network with two layers using Python. Our steps:
- Data ingestion
- EDA (data visualization, NaN)
- Data preparation (interpolation for raw data, calm winds)
- Train and model testing
Challenges we ran into
a) Data cleaning b) Deal with missing data c) Attend different activities at the same time d) A partner experienced a tremor during the development of the challenge (5.5 degrees in Chile)
Accomplishments that we're proud of
a) Build our database b) Create new variables from existing variables c) Follow up on a previous project d) Detect errors in databases e) We get acquainted with other algorithms f) Encourage teamwork!
What we learned
a) Socialize with different platforms (devpost, miro, github) b) About wind energy, sustainability b) Teamwork
What's next for Wind Energy
Test our model, improve it (tuning) and apply it to a forecasting and power supply system.
Log in or sign up for Devpost to join the conversation.