Research Of The Harmonic Currents Detecting Approach Based On The Neural Network

As lots of power electronics devices have been applied in many fields, theproblem of harmonic pollution in the power system is more serious. Active powerfilter is the effective solution to this problem .The harmonic detection is the key factoraffecting the compensation performance of APF, but also the extension application. Asthe development of the harmonic research, the harmonic detection is applied not onlyin the APF, but also in the other sides of the power system such as the harmonicstandard. To satisfy the harmonic measure requirement of the modern power system,this paper studies the new approach and realization of the harmonic detection. Based on the analysis and compare of these approaches to the harmonic detectionin the power system, an adaptive detecting approach of the harmonic currents basedon Hopfield neural network optimum theory for active power filter is proposed in thispaper. The adaptive neural network-based signal processing technique is applied todetermine the magnitude and phase of the fundamental and each of the harmoniccurrent components in real time, and this approach can be applied in single-phasesystem and three-phase system. Compared with other approaches, this one can satisfythe requirement of the APF, but also can be applied in the other sides of the powersystem. Because the modern power system needs high accuracy and real time to theharmonic detection, the harmonic detection realization technology will develop fromthe analog circuit technology and unprogrammable digital circuit technology to theprogrammable digital devices.This paper applies the Application Specific IntegratedCircuit technology on the design of the harmonic detection based on the FieldProgrammable Gate Array technology, the behavior simulation of the experimentverify this approach is feasibility.