Research On Critical Technology For Harmonic Responsibility Quantification In Distribution System

While the vigorous development of industrial technology drives the continuous growth of economy,large-capacity and non-linear loads are inevitably connected into the distribution network,making the harmonic pollution increasingly prominent.In addition,with the implementation of China’s strategy of energy transition,the access of distributed generator and vehicle-network-linkagesystems has further aggravated the harmonic pollution of the distribution network.Harmonic pollution will threaten the normal operation of distribution system.Since the distribution network is the link between customer and utility,for customers,they always expect the utility to provide them with a steady and standard sine voltage with qualified power quality.For the utility,they hope that the customers can decrease their harmonic injection as more as possible.Therefore,in order to comprehensively consider the concerns and interests of both the utility and the consumer,the study of harmonic responsibility quantification in distribution network has important theoretical value and practical significance.This thesis focuses on the critical technology for harmonic responsibility quantification of the distribution system.Firstly,the methods of harmonic emission level estimation for single harmonic source are studied.According to the harmonic characteristics of different types of nonlinear equipment,the harmonic impedance calculation methods of each type of the harmonic source are studied respectively.After that,the harmonic contribution evaluation method is studied considering the interaction of each harmonic source.Finally,these methods are verified by computer simulation and engineering field data.This paper can provide a set of technical solutions for harmonic responsibility quantification of the distribution network,and the main research work about the aforementioned aspects are summarized as follows:For the traditional harmonic source in distribution network,utility harmonic impedance estimation is the critical step for its harmonic emission level estimation.Therefore,for the traditional harmonic sources,the following two research work has been carried out:(1)On the premise that a relatively stable period of background harmonic existing in the estimation interval,a utility harmonic impedance estimation method is proposed from the perspective of data screening.Firstly,obtain the harmonic impedance samples according to the binary linear regression,after that,the relevant characteristics of the harmonic voltage and current fluctuations at the PCC are studied,and the dynamic time warping distance is defined as the similarity measure to judge the effectiveness of the impedance samples.To enhancement the strength of data screening,according to whether the number of valid samples obtained of a whole calculation meets the set threshold,adaptively adjust the width of the binary linear regression window.Computer simulation and filed data are used to verify the effectiveness of the proposed method.(2)A utility harmonic impedance estimation method constrained by prior information of background harmonic is proposed.Considering that the sufficient condition of(1)is the existence of a stable period of background harmonic source,and it cannot be met in some scenes,which leads to the failure of the corresponding scene of(1).To overcome this problem,the prior information of background harmonic is introduced as a reference signal of constrained independent component analysis in the process of utility harmonic impedance estimation,which ensure the accuracy of the recovered utility harmonic source signal.Then,utility harmonic impedance is obtained through matrix transformation and comparison with the original hybrid matrix.The premise of the proposed method is that there are harmonic fluctuations on both sides,which is mutually supportive with the first research part,and is suitable for harmonic emission level evaluation of traditional harmonic sources in the distribution network.For the distributed harmonic sources,the following research work has been carried out:(3)Since the distributed harmonic source has the characteristics of severe background harmonic fluctuations and the utility harmonic impedance is close to that of customer side,which make the quantitative index only considering the utility harmonic impedance is invalid,and customer harmonic impedance become an essential parameter for harmonic emission level evaluation.However,there are few researches about customer harmonic impedance estimation.In this part,a harmonic emission level evaluation method for distributed harmonic source is proposed by introducing the inherent properties of the harmonic impedance matrix into the complex independent component analysis.The proposed method can calculation the customer and utility harmonic impedance at the same time while the impedance amplitude of both sides is close,which is suitable for the harmonic emission level evaluation of distributed harmonic sources.The scenario where the amplitude of harmonic impedance on both sides are close due to the influence of harmonic filters and compensation devices is also applicable.For the multi-harmonic source responsibility quantification under multicustomer interaction,the following research work has been carried out:(4)Multi-source harmonic responsibilities quantification problem is researched under the interaction of multiple loads.To overcome the negative impact of multiple linear correlation of harmonic currents and harmonic impedance changes during the responsibility estimation period,a multi-source harmonic responsibility estimation method is proposed based on Lasso regression and random matrix theory.Firstly,based on random matrix theory(RMT),a data-driven method to recognize the operated status of the distributed network is proposed to detect whether the harmonic impedance change,and evaluation interval is divided into different segments based on the change point.Then,a regression model of multisource harmonic responsibility determination is established and the Lasso algorithm is used to quantify the harmonic contribution of suspected harmonic sources for each responsibility sub-segment.Finally,the total harmonic responsibility of each harmonic source is obtained by weighting based on the number of samples in each sub-interval.The proposed method can consider effect caused by network operation state change and the multi-linear correlation between the measured current data,it has higher calculation efficiency and can realize the dynamic characterization of harmonic responsibility,which is suitable for the scenario that the network state changes frequently,and the harmonic sources are coupled closely.