On the Use of the Coefficient of Variation to measure temporal correlation

Rudy Calif (1) , Ted Soubdhan (1)

1 LARGE - Laboratoire de Recherche en Géosciences et Energies [UR2_1]

In this work we perform a statistical analysis of global solar radiation measured at two sites, Petit Canal and Pointe a Pitre, 38 km distant from, Guadeloupe, French West Indies. We have established a correlation model based on the coefficient of variation assuming a time scale separation. The coefficient of variation is calculated on 10 min interval with data measured at 1 Hz. This analysis highlights the dynamic correlation that can occur between measurements from two different sites with a time step of 1 s. From these results, knowing the coefficient of variation on a site, we have established a new correlation model on this parameter for another site. A diagram linking the standard deviation for the studied sites, for a given coefficient of variation is proposed for correlated and non-correlated cases. Moreover this analysis evidences the existence of a threshold time under which there is no significant correlation. The methodology and the model can be applied to any other sites to establish diagrams of the coefficient of variation.This model can be useful in choosing new sites of PV production in establishing correlation from on site to another

Mots clés: Solar radiation; Cross-correlation; Time scale separation; Coefficient of variation

Domaines: Sciences de l'ingénieur [physics]

Rudy Calif, Ted Soubdhan. On the Use of the Coefficient of Variation to measure temporal correlation. Renewable Energy, 2015, 88, pp.192-198. ⟨10.1016/j.renene.2015.10.049⟩. ⟨hal-01528078⟩