Artificial neural network based static VAR compensator for voltage regulation in a five bus system

This paper aims at voltage regulation at all buses in a five bus system for varying load conditions and power factors by using Static VAR Compensator(SVC). The load flow analysis by Newton Raphson method is done for the five bus system using ETAP software with and without compensation system. From the results, voltages at load buses, total active and reactive powers are taken and given as inputs to the Artificial Neural Network (ANN). The switching of Thyristor Switched Capacitors (TSC), Thyristor Controlled Reactor (TCR) and the firing angle for the TCR are given as reference output to the ANN. The ANN is trained with the load flow results which are given as inputs and target outputs. After testing, the entire ANN including weights are implemented in dsPIC30F4011. The switching of the TSC and TCR, which are the outputs of ANN are shown by the glowing of LEDs under varying load conditions and power factors. The hardware setup is made for voltage regulation in a single phase Induction motor for varying load conditions using TCR and Fixed Capacitor (FC). The firing pulses for the TCR is provided for varying load conditions from dsPIC. Index terms – Artificial neural Network, static var compensator, dsPIC, load flow analysis, ETAP.