Impacto das mudanças climáticas nos atuais sistemas fotovoltaicos e projeções de longo prazo

Abstract

This thesis presents some proposals related to climate change that can influence photovoltaic systems and, possibly, infer the influence of solar cycles on their performance. Projections of these climate changes were carried out in relation to ambient temperature, location, precipitation and irradiance for all states in Brazil. As solar irradiation data only began to be collected from the year 2000 in some locations in the country, it was necessary to convert solar insolation data into solar irradiation using the Ångström-Prèscott search. To ensure greater reliability in the prediction results, this search was adjusted to the specific characteristics of each location using coefficients for all Brazilian states, as such information is not yet available in the literature. These simulations are based on time series using regressions and two software programs. This work used data from all existing residential photovoltaic systems in Brazil, which have already surpassed 818 thousand installations, and proposes a simple and accurate method that can be easily expanded worldwide. 658 meteorological stations were analyzed, with at least one from each state, for the period from 1961 to 2020, totaling 59 years of information, with which future climate forecasts were projected for up to 100 years. With the results obtained, it was found that all Brazilian states already suffer large efficiency losses due to high environmental temperatures, which are gradually increasing even further, reaching a maximum average annual increase of more than 5 ºC in the north-northeast states. of the country and two predominant cycles were observed for all Brazilian states in relation to solar irradiance, lasting approximately 22 and 65 years. Even with these prospects for a reduction in the yield of PV electrical generation, it is unlikely that such climate changes will threaten this sector in Brazil, as with all this there remains a strong potential for use.