The use of renewable energy sources is constantly increasing nowadays. The main reason for this tendency is the limit of the amount of economically exploitable fossil-based energy sources according the actual most advanced technologies. The burning of fossil fuels also raises serious environmental concerns such as global warming, air and water pollution. According to the principles of the sustainable development such economic methods should be used which do not cause permanent damage to the nature and do not limit the possibility of development at the same time. The use of renewable energy sources such as solar, wind, water, geothermal energy and biomass meet these requirements.
This thesis details the high efficiency utilization of solar energy to produce electricity. This work focuses on the development of a DC, low power, modular system with multiple solar panels, the required power and control electronics. The goal is to simulate and implement a system which maximizes power extraction under all conditions such as varying temperature, irradiance and partial shading. The neural network based MPPT (maximum power point tracking) algorithms are the most robust according to the literature. These methods are efficient even in the case of partial shading therefore this work shows the implementation of a neural network based algorithm.