Remote monitoring of photovoltaic panels and forecast of generation in photovoltaic panels using extreme learning machines

With the increasing energy consumption, the importance of energy resources increases in this direction. The importance of renewable energy sources in the production of electrical energy due to the depletion of reserves overuse of fossil fuels and environmental problems have increased. Photovoltaic panels (PV), which is one of the renewable energy sources, are used as an alternative production tool since they can convert solar energy directly to electrical energy. With the increase in the number of power plants based on photovoltaic solar energy, the monitoring of the power plants has became importance. In this study, Extreme Learning Machines (ELM) is used to predict future energy production of solar power stations by taking into consideration meteorological factors affecting panel efficiency. Current, voltage obtained from the panel where the system is located, as well as meteorological values such as solar radiation, wind, temperature and humidity affecting the panels will be measured periodically with appropriate sensors. These measured data will be recorded with the ESP8266 based Arduino board placed in the environment where the panels are located. Based on these data, the performance of the panels will be evaluated. Finally, it will be possible to instantly evaluate the production performance of the system. Extreme Learning Machine (ELM) will be used for this. Meteorological values, solar radiation, humidity, temperature and solar radiation will be entered as inputs to the ELM and, unlike the literature, the production performance of the system for a certain period will be estimated.