Harmonic real time estimation by artificial neural network based on network and GPS technology

A new harmonic real time identification approach by adaptive neural network based on GPS technology and distributed Ethernet network was proposed in this paper. The method uses an adaptive neural network to estimate the amplitudes and angles of the distorted current in power system. In this method, only half cycle harmonic current signal was used as the input of the neural network. In order to improve the accuracy of harmonic source identification, GPS (global positioning system) is used as the synchronized signal for the embedded measurement system based on digital signal processor (DSP). The samples selecting and training methods of artificial neural network are explained and the hardware structure of the embedded harmonic identification system is given. RTDS (real-time digital simulator) simulation results illustrate the effectiveness of the proposed approach