A DRL-enabled Adaptive Sparsity Promoting Wide-area Damping Controller for Power System with Time-delay and Packet Dropout
- Resource Type
- Conference
- Authors
- Cao, Di; Zhang, Guozhou; Hu, Jiaxiang; Hu, Weihao; Huang, Yuehui; Chen, Zhe
- Source
- 2023 Panda Forum on Power and Energy (PandaFPE) Power and Energy (PandaFPE), 2023 Panda Forum on. :1854-1858 Apr, 2023
- Subject
- Power, Energy and Industry Applications
Signal Processing and Analysis
Damping
Deep learning
Adaptation models
Adaptive systems
Reinforcement learning
Benchmark testing
Control systems
Power system control
sparsity
time-delay
packet drop
deep reinforcement learning
- Language
This paper proposes a deep reinforcement learning (DRL) based adaptive sparsity promoting wide-area damping controller for power system. Different from traditional fixed wide-area control structure, the proposed method employs DRL algorithm to identify the optimal sparsity promoting control structure and obtain corresponding parameter settings under different operating conditions considering the time-delay and packet drop. Simulation are carried out on 10-machine 39-bus system to demonstrate the effectiveness of the proposed approach.