Open AccessArtificial emotional reinforcement learning for automatic generation control of large‐scale interconnected power grids
Author(s) -
Yin Linfei,
Yu Tao,
Zhou Lv,
Huang Linni,
Zhang Xiaoshun,
Zheng Baomin
Publication year - 2017
Publication title -
iet generation, transmission and distribution
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.92
H-Index - 110
eISSN - 1751-8695
pISSN - 1751-8687
DOI - 10.1049/iet-gtd.2016.1734
Subject(s) - reinforcement learning , automatic generation control , controller (irrigation) , computer science , grid , function (biology) , electric power system , control theory (sociology) , control engineering , power (physics) , artificial intelligence , control (management) , engineering , mathematics , quantum mechanics , geometry , evolutionary biology , agronomy , biology , physics
This study proposes an artificial emotional reinforcement learning (ERL) controller for automatic generation control (AGC) of large‐scale interconnected power grids. In the scheme of ERL, the agent consists of two parts, a mechanical logical part and a humanistic emotional part, which essentially develop the control strategies of the agent. These two parts in proposed controller are introduced by reinforcement learning (RL) and artificial emotion (AE), respectively. The ERL controller can generate different control strategies depending on the operating scenarios, by highly integrating AE functions, which are quadratic function, exponential function, and linear function, respectively, with the elements of RL, such as action, learning rate, and reward function. The effectiveness of ERL controller with nine control strategies has been demonstrated considering AGC on a two‐area load frequency control power system and China Southern Power Grid (CSG) power system. Results of simulation show the superior performance of ERL over that of proportional–integral control and four RL techniques.