Bivariate Copula Modeling of Electricity Load, Case Study of Kwame Nkrumah University of Science and Technology

Abstract

Dependence between random variables is a phenomenon that cannot be over emphasized. This study considered the dependence between consumption and maximum demand for a bulk electricity customer when it comes to its electricity load. The study applied the Clayton, Frank, Gumbel, Joe and Tawn Type 1 copulas to the realizations of the random variables (consumption and maximum demand). Considering the AICs and BICs of the models under study, the Tawn Type 1 copula model best represented the dependence between consumption and maximum demand. Using the inverse marginal distributions of both consumption and maximum demand, actual values were obtained from the pseudoobservations provided by the Tawn Type 1 copula. All the models revealed the presence of lower tail dependence between the two variables, with the exception of the Frank copula. The selected model was used to forecast load consumption and maximum demand. � 2019 IEEE.