Rotor angle stability enhancement using DDPG reinforcement agent with Gorilla troops optimized input scaling factors

Area of Science:

• Electrical Engineering
• Control Systems
• Artificial Intelligence

Background:

• Rotor angle stability is crucial for power system reliability.
• Traditional Power System Stabilizers (PSS) have limitations in adapting to complex grid dynamics.
• Reinforcement Learning (RL) offers a promising approach for adaptive control.

Purpose of the Study:

• To develop and evaluate a Reinforcement Learning (RL)-based Power System Stabilizer (PSS) for improving rotor angle stability.
• To optimize the RL agent's performance using the Gorilla Troops Optimization (GTO) algorithm.
• To validate the proposed PSS across diverse power system test cases.

Main Methods:

• A Deep Deterministic Policy Gradient (DDPG) algorithm was employed to train the RL agent for the PSS.
• Input features included scaled generator accelerating power, its derivative and integral, and real power.
• The Gorilla Troops Optimization (GTO) algorithm was used to optimize scaling factors for input observations.
• A discrete reward function focused on generator accelerating power below a threshold.

Main Results:

• The proposed RL-based PSS demonstrated superior performance in simulations compared to Multiband dw speed-based PSS (MB-PSS), lead-lag dw speed-based PSS (dw-PSS), and lead-lag dPa accelerating power-based PSS (dPa-PSS).
• The stabilizer exhibited enhanced transient stability capabilities, even under long-duration fault conditions.
• Simulations were conducted on Single Machine Infinite Bus (SMIB), Kundur's four-machine, and IEEE 39 bus ten-machine systems.

Conclusions:

• The developed RL-based PSS with DDPG and GTO optimization effectively enhances rotor angle stability.
• The proposed method offers a robust and adaptive solution for power system stabilization.
• This approach presents a significant improvement over conventional PSS designs in terms of transient stability.