Impedance Parameter Measurement Of AC Microgrid Based On BP Neural Network

With the increasing demand for energy and the gradual improvement of environmental awareness in human society.Microgrid has become a hotspot of international and domestic research.The microgrid composed of distributed small capacity distributed generation system,various energy storage systems and new energy generation power generation.The total capacity of the microgrid is small,but it contains a large proportion of new energy and large amount of power electronic equipments.Therefore,the stability of the microgrid is a key consideration during actual designing and operation process.It is a more widely used method to analyze the stability of microgrid by judging whether the impedance parameters of the power supply and the admittance parameters of the load part match,it is easy to apply in practical engineering.Therefore,the measurement of impedance parameters of microgrid is very important.The purpose of this paper is to measure the impedance parameters of microgrid quickly and accurately.First,in this paper,some improvements have been made to the shortcomings of the current impedance parameter measurement methods of the AC system.Two different frequency sinusoidal amplitude modulation is carried out by the square wave,and the harmonic current of the injection system is superimposed by the superposition of the two modulated signals.This method combined with the characteristics of square wave current and sinusoidal current to improve the harmonic current of injection system.This method not only solves the problem of missing spectrum of injected pulse current,but also solves the problem of slow measurement of sinusoidal current injected by sweep frequency method.The improved method can effectively reduce the measurement speed and ensure the accuracy of the measurement.Secondly,the simulation model of the microgrid is built in Matlab,the impedance parameters of the power supply and the admittance parameters of the load part are measured by injecting the newly obtained harmonic current.In order to verify the accuracy of the simulation results,the harmonic impedance calculation value of the main components of the microgrid isderived by theory.By contrast,it can be proved that the measurement accuracy and speed are obviously improved when the impedance parameters are measured by the injection of the harmonic current proposed in this paper.Finally,using the characteristics of BP neural network,by injecting an effective input signal of the impedance parameter,after a large number of experimental training,a network model which can accurately simulate the dynamic behavior of the measured system is obtained.The prediction ability of impedance parameters by BP neural network ensures the measurement accuracy and reduces the on-line measurement time in impedance measurement,reduces the influence of measurement time on the accuracy of the measurement results,and reduces the interference of impedance measurement to the system.