Markov chains estimation of the optimal periodicity for cleaning photovoltaic panels installed in the dehesa

Autor: Joaquín García Sanz, Justo García-Sanz-Calcedo, Gonzalo Sánchez-Barroso, Jaime González-Domínguez
Rok vydání: 2021
Předmět:
Zdroj: Renewable Energy. 179:537-549
ISSN: 0960-1481
DOI: 10.1016/j.renene.2021.07.075
Popis: The European Dehesa has a very high potential for the production of clean energy due to the solar irradiation it receives. Its arid climate, however, means that airborne dust particles accumulate on the photovoltaic panels, with the resulting reduction in transmittance of the glass top-sheets. Cleaning the module surfaces involves an economic investment that, to be profitable, has to be offset by sufficient increased energy production. The objective of the present study was to determine the optimal periodicity for cleaning photovoltaic panels installed in the Dehesa, and thus subject to its specific climate. To this end, an experimental installation was set up, and three cleaning plans (monthly, quarterly, and semi-annually) were tested against equivalent not-cleaned controls. The results showed monthly cleaning to increase a year's worth of energy generation by 11.15%. From weekly inspections and continuous monitoring of the panels' output power, a Markov-chains based mathematical model of the degradation of energy production was developed. The conclusion drawn from it was that the cleaning frequency should be monthly from July to October (with the optimal frequency being every three weeks), but that from November to June cleaning is unprofitable since it provides no significant improvement in output. Modelling the degradation of energy output constitutes a powerful tool with which to increase the bankability of photovoltaic plants.
Databáze: OpenAIRE