Using a Markov Chain Monte Carlo Technique to Simulate Synthetic Natural Inflow Energy Scenarios

摘要

In recent years, studies on Natural Inflow Energy (NIE) synthetic scenario simulations have resulted in new methodological proposals. Such developments often assume Gaussianity in the residues; thus, making it possible to transform the data into a parametric distribution. It was noticed that, in most real cases in the Brazilian Electric Sector, the noise cannot be treated thus, since it presents intrinsically skewed tail behaviors that are challenging for the National Interconnected System's operational planning to reproduce. Thus, this work proposes a nonparametric approach to simulate and sample the NIE-series residues using the Markov chain Monte Carlo technique and the Kernel Density Estimation; hence, it is possible to simulate synthetic NIE scenarios. The presented results show that the proposed methodology is a good alternative to the current model.