Voltage quality enhancement in distribution system using artificial neural network (ANN) based dynamic voltage restorer
Voltage quality issue and invariably power quality has become an important issue in distribution power system operation due to presence of increased use of nonlinear loads (computers, microcontrollers and power electronics systems). Voltage sags and swells as well as harmonics are problems for industrial, commercial and residential customers with sensitive loads, which need urgent attention for their compensation. In this paper, the modeling and simulations of a dynamic voltage restorer (DVR) was achieved u sing MATLAB/Simulink. The aim was to employ artificial intelligence to provide smart triggering pulses for the DVR to mitigate and to provide compensation against voltage sags and swells. The Artificial Neural Network (ANN) was trained online by data gener ated via a 3 - phase programmable voltage generator and these were used as inputs to the ANN, fault conditions were simulated to create voltage sags and swells in the source supply, while faultless condition of the system was simulated and the data obtained from it was used as targets of the ANN. A net fitting, feed forward back propagation, Lavenberg - Marquardt training algorithm and mean square error performance were used. ANN Simulink block was used as control for the gate of the full wave 3 - phase Insulated Gate Bipolar Transistor (IGBT) inverter employed in constructing the DVR. Three single phase injection transformers were employed to regulate the output amplitude voltage from the DVR, while filters were used to reduce the harmonics from 11.09% to 3.5%. A t the end, voltage sags and swells were effectively mitigated and harmonics in the system reduced to 3.5%, which is within the maximum acceptable IEEE standard 519 of 1992 for harmonic distortion.
Key words : Voltage, Distribution System, ANN, Dynamic Volt age Restorer, Voltage quality enhancement, non - linear loads