Research On Estimation Method Of Harmonic Emission Level Of Decentralized Power Electronic Load Group

With the development of the national economy and the improvement of residents’ living standards,the electronic load of civilian power has been widely used,and distributed photovoltaic power generation has also developed rapidly in residential electricity consumption.These harmonic sources are distributed in the entire distribution network of the station area,and they are characterized by a large number of species,widespread distribution,and strong random emissions.Although the harmonic emissions of a single load are not significant,the harmonic effects of a large number of such loads on the grid cannot be ignored.If the harmonic pollution generated by these decentralized harmonic sources is not properly and effectively treated,it will further cause harmonic pollution to the entire power system.Accurate assessment of the group harmonic emission level in the station area can improve the economic operation of the system and provide a basis for harmonic pollution control.Therefore,it is of great significance to accurately estimate the dispersion and time-varying harmonic pollution caused by power electronic load in the power grid.This paper proposes a method for estimating the harmonic emission level based on the combination of mechanism modeling and data driving from the perspectives of harmonic source mechanism modeling and residential electricity consumption behavior.First,clustering and dividing loads of the same function according to Mahalanobis distance(MD)of multi-harmonic characteristic parameters between loads.The clustering can reflect the harmonic characteristics of load equipment of different brands and models of the same function.Differences,so as to separately establish the harmonic Norton model of the load under each divided category;and consider the access of distributed photovoltaic power sources,using distributed photovoltaic power sources as a type of harmonic source alone,and establishing the harmonic emissions of distributed photovoltaic power sources model.Then,according to the season and temperature of the target area,electricity price and per capita income and the number of permanent entrances,the electricity consumption data is obtained through sample migration,and then the electricity consumption data is carried out according to the non-intrusive load monitoring(NILM)technology.Decompose to get the breaking state of the user load equipment at each moment,and then use Markov Chain(MC)to simulate the changes of the power load and the number of grids connected to the grid.Finally,combining the change behavior of the number of simultaneous loads in the station area with the load harmonic source model can estimate the station’s load harmonic emission level,and then superimpose it with the distributed photovoltaic power harmonic emissions to finally obtain the station area harmonics.Emission estimation curve.the analysis of an example verifies the validity and accuracy of the method.