Generation of Multi-Resolution Scenarios of Stochastic Variables for Operation Planning Studies

As the penetrations of variable renewable energy sources (VRES) deepen in power systems, the net demand features greater degrees of variability and uncertainty. Several works focus on modeling the uncertainty in various decision-making processes (e.g., planning and operation). However, variability, particularly that occurring within operating periods (e.g., every hour), has received less attention. Modeling such variability in detail is of paramount importance when determining operating reserves (e.g., regulation, flexibility), and when exploring different futures in planning exercises. Only a few works have focused on this and in a partial manner. This work presents a statistical approach to produce any desired number of multi-resolution (i.e., at any desired sub-hourly level) scenarios, that follow all the statistical properties of the historical or projected data sets. Numerical results on the RTS-GLMC data set demonstrate the accuracy of the proposed method and validate the results.