Probabilistic power flow, probability distribution, Monte Carlo simulation, unscented transform
Abstract
The power systems planning and operation face increasing challenges due to uncertainties associated with generation and consumption. In this context, this work explores the assessment of uncertainties using probabilistic power flow, applying the unscented transform (UT) technique. The UT can significantly reduce computational burden compared to traditional numerical methods, however, its inability to provide accurately probability distribution functions (pdfs) for random variables may be a limitation to its application. Therefore, complementary pdf reconstruction strategies become necessary. This study performs a comparison among different pdf reconstruction approaches via the method of moments: Cornish-Fisher expansion, Johnson system, and Pearson system. The results are compared with the solution via Monte Carlo simulation, in order to evaluate application limits and effectiveness in the approaches.
