Research Of Quantifying The Harmonic Impacts Of Multiple Harmonic Sources In Power System

With distributed harmonic sources springing up in a large amount, the harmonic problem of the power system has attracted increasing attention. All kinds of harmonic sources inject higher harmonic into the system, which affect the whole electrical environment and produce many hazards. With the rapid development of power market, switching out the customer by force to limit excessive harmonic injection does not work. To control the harmonic in the grid more effectively, the "rewards and punishment scheme" was put forward internationally. After that, the scientific and technical workers have made a lot of beneficial exploration in the field of harmonic source identification. The research findings mainly focused on distinguishing the harmonic impacts of the system side and user side on the Point of Common Coupling (PCC), and evaluating the harmonic emission level of single harmonic source. However, in the actual system, the harmonic problem on the PCC is usually produced together by multiple harmonic sources. In order to avoid the dispute of power quality and control harmonic effectively, it is necessary to distinguish and quantify the harmonic impact of each harmonic source. At present, the research in this respect is less. Quantifying the harmonic impacts of multiple harmonic sources is an important direction in harmonic sources identification field. It will provide the theoretical foundation for establishing a fair and effective mechanism of rewards and punishments to limit and control harmonic pollution. Also, it is expected to make a contribution to the intelligent management of power quality on user side in smart power grid.This thesis did research work on the problem of distinguishing the harmonic impacts of multiple harmonic sources. Based on the measured data, the paper tried to quantify the individual impact of each harmonic source in the system. For a certain harmonic frequency, the harmonic impact of a single harmonic source can be expressed as the projection of the harmonic voltage generated by the harmonic source on the total harmonic voltage of the observation bus. On this basis, the paper provided the theoretical foundation for distinguishing harmonic impacts, and defined the harmonic impact index factor. By means of harmonic impedance, this index factor can be expressed as the combination of the harmonic voltage on the observation bus and the harmonic currents of suspicious harmonic loads. According to the measured data, the research extracted the key information, defined the statistical correlation relationship between the harmonic voltage and current data, and then established the mapping relationship between them. Based on M-estimation robust regression and partial least square regression, two methods were proposed to estimate harmonic impacts. Details are summarized as follows:(1) The M-estimation robust regression method, according to the relationships among all harmonic phasors, through mathematical derivation and transformation, based on data selection, solved the harmonic impact index utilizing the robust regression parameters. This method estimated the regression coefficients by adopting an iteratively least squares algotithm, which overcome the shortages of the traditional least square method that the parameter was sentive to singular data and susceptible to damage. Both simulation and case study results verified the validity of the proposed method and its superiority over the traditional least square method.(2) The partial least square regression method, based on the analysis of the phasor relationships between harmonic voltage and current, from the perspective of statistical properties, established the partial least square regression equation with the harmonic voltage of the observation bus and the harmonic currents of the suspicious harmonic loads. It evaluated harmonic impact utilizing the regression coefficients. Without data selection, this method can be used widely. Especially when the number of harmonic source is larger, the performance is excellent.(3) The paper also discussed the problem of distinguishing harmonic impacts in the system with multiple harmonic feerders. About how to quantify the contribution of the harmonic load connecting with each feeder to the harmonic voltage on the PCC, the paper build a model. For a certain harmonic frequency, the model can be equivalent to the thevenin equivalent circuit. The harmonic voltage generated by single harmonic feeder can be expressed with the product of its harmonic current and corresponding impedance. On this basis, the theory of quantifying harmonic impacts was provided. The harmonic "theory currenf’and "actual currrent" were defined. The paper analyzed their correlation relationship and different accuary of harmonic assesment. The validity of the proposed methods was verified respectively in this field. Study showed that the partial least square regression method had superior performance, which was suitable for popularization and application in practice.Based on above theory research, the harmonic impact assessment software was developed according to the proposed methods. And the application of the research in the engineering practice was explored utilizing the combination of hardware and software. The development process of the assessment software and the combination scheme of the hardware device were introduced in detail. On this basis, the evaluation results were provided based on the operation data in the substation.