Artificial Neural Networks-Based HVDC System for Transient Stability Enhancement of Nigeria Power Grid

The increasing disturbances in power system networks present significant challenges to maintaining stability, especially in grid-tied generators, posing risks to synchronism and grid resilience. In this paper, an artificial neural networks (ANN) based high-voltage direct current system is applied as a FACTS device to improve the transient stability of a multi-generator power system. A comprehensive analysis was conducted on Nigeria 330kV 40-bus transmission network using the MATLAB-based Power System Analysis Toolbox (PSAT). An initial system assessment used the Newton-Raphson power flow method and eigenvalue analysis to establish base case stability metrics and to reveal critical stability issues. This analysis shows a significant voltage reduction of 0.70 per unit (pu) and synchronism loss under fault conditions on the test system. The application of proportional-integral (PI) controller-based HVDC systems improved the system to a minimum voltage magnitude of 0.80 pu, which is below the statutory transmission voltage limit of 0.95 to 1.05 pu. Therefore, an ANN-based HVDC system was along the lines and this shows significant improvement with a three-phase fault clearing time reduced to 2 seconds, compared to the 3 seconds obtained with the PI controller-based device. This method improved voltage profile to a minimum voltage magnitude of 0.98 pu, improving system stability and synchronism. The results highlight a 27.8% improvement in voltage magnitude, affirming the proposed method as a superior alternative for transient stability enhancement. This paper provides valuable insights into the integration of intelligent systems for sustainable power grid operation and improved fault resilience in complex power networks.