Neural network based conductance estimation control algorithm for shunt compensation

Abstract

For mitigation of power quality problems in a distribution system, it is important to estimate effecting factors which are responsible for their origin. Main objectives of neural network application in Distribution Static Compensator (DSTATCOM) are to enhance the efficiency, robustness, tracking capability according to requirements. A control algorithm based on load conductance estimation using the neural network is implemented for DSTATCOM in a four wire distribution system. The proposed control algorithm is used for extraction of load fundamental conductance and susceptance components of distorted load currents. It is implementated for mitigation of power quality problems such as reactive power compensation, harmonics elimination, load balancing and reduction of neutral current under linear/nonlinear loads. Test results on a developed DSTATCOM have shown the acceptable level of performance under balanced and unbalanced loads.