Physics- Informed Neural Network Based Dynamic Estimation for Power Grid Using a Kuramoto-Like Model
- Resource Type
- Conference
- Authors
- Tian, Jiangpeng; Chi, Ming; Liu, Zhi-Wei
- Source
- 2023 China Automation Congress (CAC) Automation Congress (CAC), 2023 China. :9437-9441 Nov, 2023
- Subject
- Aerospace
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Transportation
Training
Heuristic algorithms
Power system dynamics
Neural networks
Training data
Rotors
Mathematical models
power system
dynamic estimation
physics-informed neural network
complex network
- Language
- ISSN
- 2688-0938
Dynamic state estimation for power grid is an important task for power system monitoring and operation. In this paper, we build the power network dynamic model with the form of the Kuramoto model, and use the physics-informed neural network (PINN) to capture the dynamics of the power system. Physics-informed neural network is a new deep learning method that can incorporate the mathematical models describing power grid dynamics into network training to achieve high accurate estimation, while using less training data. We also verify the effectiveness by a simulation on a two generators-one machine power grid system.