Neural-network-based adaptive UPFC for improving transient stability performance of power system

Area of Science:

• Electrical Engineering
• Control Systems
• Artificial Intelligence

Background:

• Power systems face challenges in maintaining transient stability during disturbances.
• Unified Power Flow Controllers (UPFC) are crucial for regulating power flow and improving stability.
• Existing control methods, like PID and EKF-based RBFNN, have limitations in performance and computational efficiency.

Purpose of the Study:

• To develop and evaluate a new control scheme for UPFC using a single-neuron Radial Basis Function Neural Network (RBFNN).
• To improve the transient stability and damping performance of a multimachine power system.
• To investigate the effectiveness of H(infinity)-learning for updating RBFNN parameters and a genetic algorithm for optimizing control coefficients.

Main Methods:

• A single-neuron RBFNN architecture is employed for the UPFC control scheme.
• H(infinity)-learning is utilized for updating the RBFNN parameters.
• A genetic algorithm is used to determine coefficients for error difference, error, and auxiliary signals.
• The proposed controller is tested on a four-machine power system under various disturbances.
• Coordinated effects with conventional Power System Stabilizers (PSS) are also analyzed.

Main Results:

• The novel RBFNN-based UPFC controller significantly enhances damping performance.
• The proposed scheme outperforms conventional PID and EKF-based RBFNN controllers in stabilizing power systems.
• The controller effectively stabilizes previously unstable operating conditions.
• The simple architecture leads to reduced computational load, suitable for real-time applications.

Conclusions:

• The H(infinity)-learning-based single-neuron RBFNN controller offers a superior and computationally efficient solution for UPFC control.
• This approach effectively improves the transient stability and damping of multimachine power systems.
• The controller's simplicity and performance make it a practical choice for real-time implementation in power system stabilization.