Support Vector Machines Analysis of Photovoltaic Power Dataset Utilizing Internet of Things

The recent years have been evidencing a huge leap forward in the application of internet of things in assessment of PV systems. It altered the way we store the real - time monitoring data from the sensors on the Internet Cloud by utilizing sensor friendly Raspberry Pi computers. The detection of the unforeseeable conditions of the PV plants that effect on the reliability of the power generation being one of the contemporary problems of our times has found the solution by monitoring the weather condition in real time applying Internet of things. Relying on the weather data acquired from sensors connected to Raspberry Pi statistical model that predict the solar power generation is built. This paper proposes the 2kW photovoltaic station power performance and implements predictions by means of support vector machines (SVM) and analyses the results derived from applying different kernels. In order to assess and forecast the dataset, the models were built in MATLAB software. The most actual result was achieved in Medium Gaussian SVM model.