Live Demo/GitHub Repository/PPT Link
Hexolar is an AI Solar Tool built upon an MPPT (Maximum Power Point Tracker) using Smart Electronics and Smart Softwares with IoT to maximise the efficiency of solar panels. It helps to check the feasibility of installing solar panels in a particular location and also monitors and displays past, live and future performance of the panels using informative graphs. It has a dedicated Battery management system and highly customizable settings. It will also notify the user when there is a problem or issue in the system by performing periodic checks. It helps in switching between supplying power to the utility grid and your house when needed based on upcoming weather conditions.
Usage of solar energy has been seeing an increasing trend in India & is on track to become the most important source of energy. But we are not utilizing it to its maximum capacity. There are some parameters which the existing technology are missing which when carefully monitored can increase the efficiency of Solar based energy sources.
Hexolar's smart features help solar based energy generation to take place in an efficient manner. It helps in detecting any failures of components very easily in the web application so that the user can fix it immediately. It also analyses the power generated with the NASA solar irradiance so that the user can easily notice when the panel is performing at a lower efficiency than expected.
As the first step, we decided upon the idea we are going to work with and split up tasks between ourselves. We already had most of the components with us for the Hardware part. We started working with the Hardware, Frontend, Backend and Machine Learning parts simultaneously. By 7:30 am we were able to successfully complete the project and we started working on the PPT and Video.
- Calculations required for the prediction of power were difficult to find. We spent a lot of time finding the formula.
- Multiple cursor error kept on occurring with SQL because we were requesting data via multiple APIs at the same time. We later fixed this by making a single API request to get all data.
- Large scale testing can yield better results
- To support multiple types of batteries (Lithium-ion, Nickel-cadmium, Lead Acid, etc.)
- Implement water sprinklers for the solar panels to clean dust
- More parameters can be added for improving the accuracy of the ML model (Wind Speed, Dust Particulates, etc.)