Mitigating Risks in Renewable Energy Power Systems: Predicting Transient Voltage Angles with Physics-Informed Neural Networks

As the penetration of renewable energy in global power systems continues to increase, new challenges to the stability and reliability of these systems have emerged. This study focuses on the dynamic impact of wind power generation's uncertainty and variability on power systems under gusty conditions characterized by significant short-term wind speed fluctuations. This study employs Physics-Informed Neural Networks (PINNs) to predict transient voltage angle variations in power systems subjected to wind power generation uncertainty. Using the IEEE 9 BUS system as a case study, voltage angle variations are generated through the swing equation and an ODE solver, serving as training data for the PINNs. The results demonstrate that PINNs effectively predicts transient voltage angles in unseen wind power datasets with only 20 training wind scenarios dataset. Overall, this paper illustrates the potential of PINNs in handling the complex dynamics of renewable energy-integrated power systems, particularly in forecasting and managing grid fluctuations caused by wind power generation.