A novel multi-agent DDQN-AD method-based distributed strategy for automatic generation control of integrated energy systems

Lei Xi,Lu Yu,Yanchun Xu,Shouxiang Wang,Xi Chen
DOI: https://doi.org/10.1109/TSTE.2019.2958361
IF: 8.31
2020-01-01
IEEE Transactions on Sustainable Energy
Abstract:The widely adoption of distributed renewable energy sources (DREs) effectively reduces carbon emission and beat atmospheric haze in developing countries. However, random disturbance issues emerge in power grids with DREs when applying traditional centralized automatic generation control (AGC) strategies. Therefore, a multi-agent distributed control strategy is proposed for AGC in this article, which is mainly based on the concept of deep reinforcement learning, and developed by the strategy of action discovery. Moreover, area control error and the amount of carbon emission are employed in reward functions to obtain optimal solutions in the implementing process of the proposed strategy. Simulations are provided in the work to show the effectiveness of the strategy, while comparisons are also offered, where the simulating results obtained by two other intelligent AGC algorithms are used as references, according to which, the superiority of the proposed strategy is confirmed.
What problem does this paper attempt to address?