ALGORITHM FOR ONLINE COEFFICIENT CORRECTION ARTIFICIAL NEURAL NETWORK IN MPPT CONTROLLERS FOR SOLAR BATTERIES

Abstract

The main element of solar power plants is usually a power cascade (DC/DC converter, inverter). Converters in such generation systems should have high efficiency (at least 90%), high output signal quality and ensure operation of the power plant with maximum selection of power from the solar battery. The characteristics of solar panels depend heavily on weather conditions such as light and temperature. During the day, the temperature and power of the solar generator are constantly changing. These changes result in a shift in the maximum power point and a partial loss of power. In order to obtain the maximum possible power from the solar battery, it is necessary to use the appropriate maximum power point tracking algorithm (MPPT). For MPPT, specialized controllers are used, which use one of the algorithms to optimize the working point of the photomodules. The most commonly used methods are perturbation and observation, increasing conductivity, constant voltage. The maximum power point tracking method used will largely determine the efficiency of the photovoltaic generation system. Maximum power recovery from solar panels is possible only when the battery voltage is continuously regulated at an optimal operating point. Thus, the design and development of modern efficient photovoltaic generation systems should address not only the improvement of high efficiency solar cell technology, but also a number of issues of designing photovoltaic converters and their control systems to significantly improve their energy efficiency.