Skip to Main Content
Voltage swell/sag along with extended undervoltage conditions and voltage harmonics are of vital concern to the majority of industries today. A neural network control method for the problems of voltage sag/swell and harmonics is proposed. A novel topology of the AC-AC boost converter solves the problems of voltage swell/sag, while a voltage controlled-voltage source inverter working as a series isolator compensates the harmonics contained in the voltage at the point of common coupling (PCC). The composite converter, named a solid-state voltage restorer, uses an injection transformer with three windings. The primary winding injects a voltage in series with the line to regulate the load voltage to a desired value with a sinusoidal shape. The centre tap secondary winding is connected to an AC-AC boost converter. The tertiary winding is connected to a DC-AC inverter to compensate harmonics in the voltage at the PCC. The neural network based control technique of the AC-AC converter helps to correct the wave shape and amplitude of the voltage to feed sensitive load at rated voltage conditions. A detailed design of the neural network methodology is given and a mathematical model of the voltage restorer is used to simulate its performance. Simulation results are discussed in detail.