Recurrent neural network based design of fractional order power system stabilizer for effective damping of power oscillations in multimachine system

The complexity of modern power systems has risen due to the growing demand for electricity. To mitigate this problem, the role of the electric utility in supplying safe and reliable electricity has become increasingly crucial. Low-frequency oscillations (LFOs) in a power system are an inevitable aspect. Therefore, the reliability of the system requires appropriate damping to overcome these LFOs. This paper proposes a novel design methodology for a fractional-order power system stabilizer (FO-PSS) to enhance the stability of the power system network. FO-PSS is designed to effectively dampen LFOs in both single-machine infinite bus systems (SMIB) and multi-machine power systems (MMPS). Recurrent neural network (RNN) is used to predict the parameters of proposed PSSs. To evaluate the efficacy of the proposed PSS, various test cases are considered. The results are compared to those obtained from traditional PSSs and optimization based PSSs. The effectiveness of the proposed RNN-based FO-PSS is demonstrated under diverse loading conditions.