Solar Panel Power Prediction Using Machine Learning Technique

One of the most significant renewable energy sources produced by photovoltaic panels (PV) is solar energy. Only during the day is solar energy available, and it is affected by meteorological elements such sun irradiation, cloud coverage, rainfall, and temperature. Due to these variables, the integration of solar power into the power grid is more challenging than it is for conventional energy sources. Changes in the solar power output result in increased electricity demand and higher operating costs. This leads to predict the power produced by solar PV panels accurately. There are numerous methods for estimating the power produced by PV systems, and the data sources used are either just historical data on PV power or historical data on both PV power and weather. In proposed system the data sources used are previous PV power output, weather data and weather forecasts in order to correctly get the effects of solar panel power prediction and actual operating condition data on the efficiency of power generation by using Support vector machine learning algorithm. The model has been trained and validated using previous year data from a input as year, month, date, time, temperature, humidity, direct radiation, diffuse radiation, direct normal irradiance, wind speed and outputs the power generated which gives higher prediction accuracy.