Solar energy is a clean and renewable source of energy that helps reduce carbon footprint. Installing solar panels in the most optimal locations minimizes their cost and maximizes energy production and energy efficiency, contributing to a more sustainable future. Machine learning helps predict the solar irradiation metrics at various locations which we can use to select optimal locations.

What it does

SolarSense is a machine learning model that predicts various solar indicators to optimize solar production by finding the most efficient places to utilize solar energy. We trained a model to predict the global horizontal irradiance (GHI), the diffuse horizontal irradiance (DHI), and the direct normal irradiance (DNI) based on various environmental factors such as time, humidity, temperature, cloud covering, and the solar zenith angle, in order to determine the optimal locations for solar panel placement.

How we built it

We created and trained a backend model for DHI, DNI, and GHI prediction, and then wrote a data processing script to format location/weather data to fit the model correctly. We then implemented several visualizations of the data through graphs and interactive maps.

Challenges we ran into

The most challenging part of the project was cleaning and formatting the data, as we had to deal with data in many different formats that had to be formalized to run predictions.

Accomplishments that we're proud of

Several things we are proud of: Successfully modeling multiple solar energy indicators in a short time frame, successfully solving the issue of inconsistent data formats, and developing an easy-to-use and intuitive solution to a renewable energy problem.

What we learned

We learned about how the metrics of DNI, DHI, and GHI are related to solar panel/power efficiency. We also learned the importance of data cleaning by designing an external program that allowed us to properly fit data into the model.

What's next for SolarSense

Training on more data, analyzing more parameters/predicting more indicators, and even branching out to optimize other renewable energy sources.

Built With

Share this project: