A Novel Spatial-Temporal Agent for Transmission Section Limit Calculation Considering Multiple Stability Constraints

The advancement of new power systems has heightened challenges in ensuring power system stability, with transmission section limits now impacted by a wide array of factors, including thermal limits, transient stability, renewable energy integration, and HVDC constraints. To address these complexities, this paper proposes a novel spatial-temporal agent leveraging Graph Convolutional Networks (GCN), Recurrent Neural Networks (RNN), and Deep Deterministic Policy Gradient (DDPG) to calculate transmission limits by integrating diverse stability constraints. The agent dynamically monitors key power system measurements, such as voltage, current, rotor speed, and fire extinction angle, analyzing these state variables to determine the maximum safe transmission power. Meanwhile, multiple stability constraints are considered with the deployment of the safety layer. Testing demonstrates that this innovative approach achieves transmission limit calculations within 10 milliseconds, with a minimal error margin of 1.72%.