On-line identification of series capacitive reactance compensator in a multi-machine power system using a radial basis function neural network

With a properly designed external controller, the series capacitive reactance compensator (SCRC) can be used to damp low frequency power oscillations in a power network. Conventionally, linear control techniques are used to design the external controller for a SCRC around a specific operating point where the nonlinear system equations are linearized, However, at other operating points its performance degrades. The indirect adaptive neuro-control scheme offers an attractive approach to overcome this SCRC control problem. As air essential part of this control scheme, air adaptive neuro-identifter has to be firstly designed in order to provide art accurate dynamic plant model for the design of the external neuro-controller. In this paper, all adaptive neuro-identifier rising a radial basis Junction neural network (RBFNN) is proposed for on-line identification of an SCRC connected to a multi-machine power system. Results are included to show that this RBF neuro-identifier continuously tracks the plant dynamics with good precision.