Effects of Uncertainty Characterization of Energy Demand of a Neighborhood on Stochastic Day-ahead Scheduling

The paper presents the effects of different statistical representation of energy demand of a neighborhood on day-ahead scheduling. A stochastic energy hub model is developed to schedule the energy supply and storage components in day-ahead basis. The PV supply, electrical and thermal demand are considered as the uncertain parameters. In order to model them statistically, three different types of Probability Distribution Functions (PDFs) have been applied including uniform, normal distributions and Gaussian Mixture Model. The main objective is to minimize the amount of electrical energy purchased from the grid, where the stochastic outputs are compared with deterministic output. Two distinct parameters have been used to quantify the differences. Relative Mean Absolute Error (RMAE) represents the accuracy of the approach and bound deviation represents the reliability of the stochastic approach. Simulation analyses on the neighbourhood surrounding VU Medical Center and University campus in Amsterdam reflect that the GMM model representation is the most accurate and reliable.