Probabilistic Computational Model for Correlated Wind Farms Using Copula Theory

Abstract

This paper proposed a probabilistic load flow analysis of correlated wind farms based on Copula theory. This method addresses the linear and non-linear dependence between random variables more efficiently and accurately than other methods. The proposed method is nearly unconstrained to the marginal probability distribution types of the input random variables. The dependency between the input random variables is established using Copula theory in this paper. An improved Latin hypercube sampling is adopted due to the real discrete data. Uncertainty and dependence factors are considered to access the load flow of the power system accurately and comprehensively. The validity of the probability distribution between the correlated random variables is evaluated by adopting the power output of wind farms located in New Jersey. The effectiveness and accuracy of the proposed model are investigated using the comparative test in modified IEEE 14-bus and IEEE 118-bus test systems.

Keywords

• Correlation
• improved Latin hypercube sampling
• Monte Carlo simulation
• probabilistic load flow
• random variable
• uncertainty