Regional and Seasonal Variation of Chaotic Features in Hourly Solar Radiation Based on Recurrence Quantification Analysis

Abstract

In this paper the solar radiation data available in some stations across Nigeria are analysed to understand the underlying dynamics of the natural time series for the purpose of good prediction and modelling of solar power generation in the country. The solar radiation data observed over a period of two years by National Space Research and Development Agency (NASRDA) from five different stations in the tropics are studied using recurrence based methods. The underlying dynamics of the hourly solar radiation data are investigated using recurrence plot (RP) and recurrence quantification analysis (RQA). From the RQA results for each month of the two years, it is observed for all the five stations that solar radiation exhibits high (low) chaoticity during the wet (dry) season due to nonlinear interaction of the solar radiation with high (low) atmospheric water vapour contents coupled with strong (weak) West Africa monsoonal effect during the wet (dry) season. Sudden transitions into or out of a ‘wet’ or ‘dry’ season which are caused mainly by external effect such as intertropical discontinuity (ITD) on solar radiation data are identified. The results show that recurrence techniques are able to identify areas and periods for which the harvest of solar energy for power generation is optimum (high predictability) and poor (low predictability) in the study areas.