The role of spatial correlations in Monte Carlo studies on power system

This paper presents a method to generate new samples of the stochastic phenomena affecting the power system, notably the consumption, uncontrollable renewable generation and hydro inflows. The method aims at reproducing characteristics of the time series observed in historical data: their seasonality, temporal correlation, marginal distribution and spatial correlations. It has been implemented into a Monte Carlo adequacy simulator. The sensitivity of some reliability indicators to spatial correlations is analysed for a French test case at the horizon 2030. It has been shown that the reliability of the system can be significantly overestimated and the assessment of the inter-zonal exchanges can be biased if spatial correlations are neglected.