Widi Aribowo, Supari Muslim, Munoto, Bambang Suprianto, Unit Three Kartini, I.G.P. Asto Buditjahjanto
The Overshoot and time settling of the electromechanical are a serious problem for tuning PSS. Cascade Forward Backpropagation (CFBNN) has a topology similar to Feed Forward Back Propagation. It is using Backpropagation to updating weights. The Network has the advantage that it is a bypass. That connects the input layer that passes through the hidden layer. The networks are dynamic. The research was implemented to oppose conventional PSS (CPSS) and Cascade Forward Backpropagation Neural Networks (CFBNN-PSS). The focus of the research was on rotor angle and angular frequency. The result of proposed CFBNN has better performance to reduce of overshoot angular frequency and rotor angle. The CFBNN PSS can reduce overshoot of angular frequency until 90.7% with faster time-settling © 2020 IEEE.
Universitas Negeri Surabaya, Department of Electrical Engineering, Surabaya, Indonesia
