On-line self-learning PID controller design of SSSC using self-recurrent wavelet neural networks

Authors : Soheil GANJEFAR, Mojtaba ALIZADEH

Pages : 980-1001

Doi:10.3906/elk-1112-49

View : 35 | Download : 13

Publication Date : 0000-00-00

Article Type : Research Paper

Abstract :Conventionally, FACTS devices employ a proportional-integral insert ignore into journalissuearticles values(PI); controller as a supplementary controller. However, the conventional PI controller has many disadvantages. The present paper aims to propose an on-line self-learning PI-derivative insert ignore into journalissuearticles values(PID); controller design of a static synchronous series compensator for power system stability enhancement and to overcome the PI controller problems. Unlike the PI controllers, the proposed PID controller has a local nature because of its powerful adaption process, which is based on the back-propagation insert ignore into journalissuearticles values(BP); algorithm that is carried out through an adaptive self-recurrent wavelet neural network identifier insert ignore into journalissuearticles values(ASRWNNI);. In fact, the PID controller parameters are updated in on-line mode using the BP algorithm based on the information provided by the ASRWNNI, which is a powerful and fast-acting identifier thanks to its local nature, self-recurrent structure, and stable training algorithm with adaptive learning rates based on the discrete Lyapunov stability theorem. The proposed control scheme is applied to a 2-machine, 2-area power system under different operating conditions and disturbances to demonstrate its effectiveness and robustness. Later on, the design problem is extended to a 4-machine, 2-area benchmark system and the results show that the interarea modes of the oscillations are well damped with the proposed approach.
Keywords : Adaptive control, flexible AC transmission systems, power system control, power system stability, self recurrent wavelet neural networks