Adaptive variable structure series compensation for voltage stability improvement using internal recurrence neural network controller

The paper presents a control technique for variable structure series compensation (VSSrC) using internal recurrence adaptive neural network, IRANN controller for voltage stability enhancement in power systems. The present IRANN controller response is dependent on the power system response but independent on it's parameters. The IRANN implements a nonlinear adaptive functions which tracks the weights and bias matrices of the constructed internal recurrence neural network according to the power system response. The present controller implements speed deviation signal, Delta omega and terminal voltage deviation signal Delta V-t added to feedback signals from the hidden layer as input signals. The output signal of the proposed controller is related to the power system response. The studied power system is modeled by a set of nonlinear algebraic and differential equations and solved by MATLAB software. The proposed scheme stabilize the studied system voltage in case of severe disturbance. A three phase short circuit fault at the main bus is considered for a period of 200 m.sec. To judge the present controller a comparative study is made with the conventional PI controller. The time response shows the superiority of the proposed IRANN controller over the PI controller in stabilizing the system voltage very fast.