Abstract:
With the emergence of fluctuating renewable electricity sources, the need for flexibilities both on the generation side and on the demand side becomes crucial to ensure power systems reliability. In this context, flexible loads such as thermostatic loads (water-heaters, fridges, heating, cooling, etc) or electric vehicle charging can be viewed as virtual storage opportunities that could contribute to the balance of the sytem. One major technical issue is then to design a process allowing to control, in a real time basis, a large population of flexible loads. We consider the distributed control framework proposed in [BM2016, CHBM2018] allowing to track a given reference consumption profile in order to provide ancillary services to the system. In a real time basis, e.g. every 15 sec. a common control signal is sent to a population of water-heaters that return randomized consumption responses. To design the control signal, we propose a new approach based on mean-field inversion, which differs from the original Proportional Integral approach. The performances of both approaches are empirically compared in terms of accuracy and robustness.
(a joint work with Pascale Bendotti and Cheng Wan)
Références:
[BM2016] Busic A. and Meyn S. Distributed randomized control for demand dispatch In IEEE 55th Conference on Decision and Control
(CDC), Dec. 2016, pp. 6964-6971.
[CHBM2018] Chen Y., Hashmi MU., Mathias J., Busic a. and Meyn S. Distributed control design for balancing the grid using flexible loads
Energy Markets and Responsive Grids, 383-411, 2018