Application of echo state networks for estimating voltage harmonic waveforms in power systems considering a photovoltaic system

摘要

An estimator based on echo state networks (ESNs) is presented in this study to estimate voltage harmonic distortion waveforms at non-monitored sensitive loads using current and voltage at a monitored location. Since distributed generations such as photovoltaic systems have played a special role in distribution networks so they are considered in this study and their effects on a harmonic voltage waveform estimator are evaluated. Voltage harmonics are considered as the main type of waveform distortion in a power quality approach. Voltage and current harmonics cause an increase in the malfunction of electrical equipment in power systems. To detect and analyse the voltage harmonics, installation of power quality monitors (PQMs) at all buses is not economical. Reducing the cost associated with the monitoring procedure can be achieved by optimising the number of PQMs to be installed. The presented technique is examined on an IEEE 37-bus network and results from the studies indicate an acceptably high accuracy of ESN estimator.