Fuzzy Modelling of Diffuse Solar Radiation

Modelling the diffuse solar radiation is a topic in many scientific and research papers, but most of the presented models are conventional, approximate and constrained to a certain geographic region, lacking the universality. Furthermore, many of these mathematical models are polynomials of appropriate degree, and/or consist of a set of expressions derived for specific intervals of their parameters. This paper presents two unconventional models of the diffuse solar radiation built on fuzzy logic. First, a position fuzzy model is discussed, and then a position-gradient fuzzy model is given. To identify these models, a modification of the algorithm presented by Sugeno and Yasukawa is applied. The diffuse solar radiation is modelled with sufficient climatic data including total solar radiation, clearness index, total sky cover, precipitation water, relative humidity, etc. The presented models offer an alternative unconventional approach to modelling of diffuse solar radiation and they have shown very good performance and certain advantage compared to other existing models through simulation. There are, of course, other unconventional techniques for modelling the diffuse solar radiation, but the most remarkable about these models is that they successfully process many different input parameters and choose the ones that most significantly influence the model output.