The Research Of Determing The Harmonic Contributions In The System With Multiple Harmonic Sources

Harmonic in power system has become one of the important indexes to measure power quality. Classifying the harmonic responsibility of the concerned bus produced by harmonic sources quantitatively is helpful to improve the initiate of the users to control harmonic and has important significance to improve power quality. The research on harmonic responsibility division focuses on determining whether the system side or the user side is the main source of harmonic interference, but this can not quantify all the harmonic sources' contribution. In the actual system, there are many harmonic sources, and the harmonic pollution of the concerned bus is caused by all the harmonic loads. Only quantify each harmonic source's contribution and sort them can ultimately determine the primary goal of governance and take rewards and punishment measures.This paper established the mathematical model to evaluate the harmonic responsibility of the harmonic sources when the main harmonic sources were known and applied the least squares estimation to the model. When the sampling data has the abnormal value and the coefficient matrix is not full rank, the least square estimation is not very good, and the M estimation and ridge estimation are proposed for these two kinds of cases. The simulation data showed that the M estimation and ridge estimation were better than the least squares estimation in the problem. When there are unknown harmonic sources, using the model and method of the previous will lead to the wrong solution of the harmonic impedance and be unable to calculate the harmonic responsibility accurately. Therefore, this paper put forward semi-parametric ridge estimation to solve this problem. The simulation data showed that the method and model were better than the previous when there were a certain amount of unknown harmonic sources.