Theory And Method For Determining The Contribution Of Harmonic Loads

With the development of economy, the requirement for power quality is more and more strict. Harmonic is one of the main power quality problems. According to the current power system harmonic management, every harmonic source have the right to inject the harmonic into the grid and no one has the right to stop other harmonic sources to do so as long as the harmonic does not exceed the national standard. This situation led "who should be responsible, who are irresponsible", causing the deterioration of the harmonic problems and restricting the improvement of power quality. It is the premise of improving the power quality harmonic management to determine the harmonic contributions for harmonic loads. Power quality monitoring data will be saved by the power quality data interchange format(PQDIF) or by the wave data. These two types of data are different. The former is the statistical value and the latter is the instantaneous value. The theory and method of harmonic contributions determination of the grid and the power customers at the point of common coupling are discussed in the thesis, which is helpful to improve the harmonic management and meaningful for achieving high quality power supply to the grid.For the wave data, complex linear regression model and its complex least squares solution are proposed for harmonic impedance estimation under the background harmonic remain constant, which avoids the shortcoming that the traditional methods can not get the optimal result in the real domain. On this basis, the harmonic contributions of the harmonic sources are determined. The result shows that the method proposed by the thesis can get the global unique optimal solution of the harmonic impedance. The fitting parameters error calculated by the complex linear regression method is smaller compared with the traditional methods. This method is suitable for the harmonic contributions determination for the wave data when the background harmonic remain constant.As for the bulk power system, the background harmonic is variant. In this case, sample data can not be modeled by the complex linear regression. To solve this problem, complex maximum likelihood estimation is proposed afer the probability density function of the complex Gaussian distribution deduced. And then, the harmonic contributions of the interface sources are calculated. This method is suitable for the harmonic contributions determination for the wave data when the background harmonic fluctuate.The harmonic data is monitored at a single point for the above two cases. Harmonic data is not rich under the single point measurement. The neighborhood multipoint measurement method for harmonic contribution determination is proposed for the further improvement of the estimation accuracy of the harmonic contributions. Due to the interaction between the electrical quantities, the measurement data is easy to show a high correlation. The data correlation will affect the calculation accuracy of the harmonic impedance. Complex partial least squares method is presented in the thesis to achieve the removal of the data correlation. On this basis, the harmonic contributions are determined. Compared with the single measurement method, the neighborhood multipoint measurement method can get more accurate result.Compared with the wave data, data recorded by PQDIF is statistical value. PQDIF data has the information of harmonic voltage and harmonic current, while lacks the accurate phasor information. Therefore, the statistical data shows the feature of large data and poor information. In order to calculate the harmonic contributions in this case, a rough estimation method is proposed based on the national standard for harmonic management. Compared with the traditional methods for the wave data, the calculation accuracy declined, but the result is still acceptable. This method is suitable for the rough estimation of the harmonic contributions for the harmonic data by PQDIF.The rough estimation method for PQDIF data is simple and easy to realize, which can be used for rough estimation. But the estimation accuracy is not enough when taking the harmonic contribution as the criterion for harmonic economic management. The partial linear model is proposed for PQDIF data for more accurate result. Harmonic contributions of the concerning harmonic sources is calculated after solving the partial linear model for harmonic impedance estimation. Compared with the rough estimation method, this method is more complicated but can get a more accurate result. The method is suitable for the accurate harmonic contributions estimation for PQDIF data.