Hanane Dagdougui – Professeure agrégée, Département de mathématiques et de génie industriel, Polytechnique Montréal, Canada
Présentation sur YouTube.
La présentation sera donnée en français.
Distributed demand-side and energy resources remain less exploited resources for power system since single prosumers are too small to provide enough flexibility, thus necessitating an aggregator to create a portfolio of many smaller participants to act as a single entity. Hierarchical architectures have been increasingly used to tackle energy management challenges at the scale of aggregators. To date, various strategies have been developed which can be generally categorized into centralized and distributed approaches. Despite the advantages of centralized methods, there exist several challenges associated with their intractability, security, and vulnerability issues. Distributed energy management strategies are more immune to single point failure, while preserving users’ privacy. Among various decomposition methods, considerable attention has been recently given to the alternating direction method of multipliers due to its fast convergence and reasonable precision. In this presentation, we will address various versions of hierarchical distributed frameworks involving multi-layers control of aggregators and local energy systems. We will also investigate how the hierarchical designs can be modeled following a distributed model predictive control, which allow the real-time operation of the network while capturing the uncertainties and achieving the fairness in energy trading mechanisms among various players.