Neuro-fuzzy forecasting of reliability of autonomous solar power generation
Abstract
Neuro-fuzzy forecasting of reliability of autonomous solar power generation
Incoming article date: 05.08.2025The work is devoted to the problems of assessing and predicting the reliability of photovoltaic generation devices. The purpose of the work is to identify factors affecting the volume of electricity generation, as well as to build models and procedures for predicting the reliability of the panels during their use depending on these factors. An overview of the types of solar power plants and the photovoltaic panels used is given. An analysis of the factors affecting their reliability is performed, on the basis of which a hierarchy of fuzzy factors related to each other by fuzzy production rules is built. It is proposed to use a statistical two-parameter Weibull model to predict the reliability of the panels. An algorithm for neuro-fuzzy tuning of the reliability forecasting model depending on the factors considered is developed and software implemented, which can be used to create information and analytical systems for decision support in the design and operation of solar power plants.
Keywords: solar energy, photovoltaic panel, reliability prediction, statistical model, neuro-fuzzy network