Advancements in Power System Modeling and Control

The field of power systems is undergoing significant transformations with the integration of renewable energy sources and advancements in power electronics. Recent developments are focused on improving the accuracy of power system modeling and control, enabling the efficient and reliable operation of modern power grids. Researchers are exploring innovative approaches, such as topology-aware graph neural networks and reinforcement learning, to predict power system states and optimize control strategies. These advances have the potential to enhance the stability and resilience of power systems, particularly in the presence of high renewable penetration. Noteworthy papers in this area include: PowerGNN, which proposes a topology-aware graph neural network framework for predicting power system states, achieving substantial improvements in predictive accuracy. Nuclear Microreactor Control with Deep Reinforcement Learning, which demonstrates the effectiveness of deep reinforcement learning for real-time drum control in microreactors, showcasing its ability to generalize and extrapolate to longer, more complex transients.

Sources

Generalised Harmonic Domain Analysis for Transformer Core Hysteresis Modelling

PowerGNN: A Topology-Aware Graph Neural Network for Electricity Grids

RL2Grid: Benchmarking Reinforcement Learning in Power Grid Operations

Quantifying Grid-Forming Behavior: Bridging Device-level Dynamics and System-Level Stability

Traffic Engineering in Large-scale Networks with Generalizable Graph Neural Networks

Analysis of the French system imbalance paving the way for a novel operating reserve sizing approach

Nuclear Microreactor Control with Deep Reinforcement Learning

An Assessment of the CO2 Emission Reduction Potential of Residential Load Management in Developing and Developed Countries

Built with on top of