SOLMATE

• 

  Main Page

SOLMATE is an acronym for Solar and MESONET Analysis for the Tracking of Energy. We have developed a solar modeling system based on MESONET weather data. MESONET data consists of observed weather patterns including, the measure of solar irradiance, precipitation, barometric pressure, wind speed, and temperature. Solar irradiance is the output of light energy from the sun. This data, combined with real-life solar power production data from sites in Oklahoma, allows us to track solar energy by date. By inputting the month with your highest power consumptions and the county you live in, our application will calculate the money you would save per month, the price of the solar panel system you would need, your carbon savings, and your return on investment.

Using the scikit-learn library, a Python script trains a deep perceptron neural network using a stochastic gradient descent algorithm. The input data was gathered from the nearest of Oklahoma’s MESONET trackers and compared to the kilowatt hours per day of a given solar array. The neural network is then pickled for later use. Using R Markdown, the user inputs the county location, desired month, and the average amount of their electricity bill. The algorithm then predicts the average kilowatt hours per day. We use this information to inform the user about their possible ROI, carbon savings, money saved, and price of a system to negate any costs. The R Markdown HTML file is combined with an information page and hosted on Google Cloud Platform using a virtual machine. The hosted page uses the Domain.com domain “solmate.tech” to get accessed.

Built With

• apache
• bash
• css
• google-cloud
• html
• javascript
• numpy
• python
• r
• scikit-learn
• shiny
• ubuntu