A New Probabilistic Load Flow Method Using MCMC in Consideration of Nodal Load Correlation

This paper proposes a new method for the probabilistic load flow calculation. In this paper, a hybrid method of deterministic annealing expectation maximization (DAEM) algorithm, Markov chain Monte Carlo (MCMC) and the AC load flow is presented to evaluate the effect of uncertainties of input variables on the output ones. DAEM is effective for estimating the maximum likelihood estimate (MLE) of probability density function (PDF) while maintaining the non-Gaussianity and the nonlinear correlation of the variables. DAEM is an extended algorithm of EM that calculates estimates for incomplete data. MCMC is used to generate the samples from arbitrary distribution while reflecting the non-Gaussianity and the nonlinear correlation of PDF. The proposed method is successfully applied to a sample system with real data.