Analysis of Power System Harmonic Effects Based on Data Features and Gene Expression Programming

In order to quantitatively recognize the harmonic impacts of harmonic source users at the common connection point (PCC), a method to calculate the system harmonic impedance and quantify the harmonic impacts by using gene expression programming to analyses data features was proposed in this paper. Firstly, the harmonic voltage and current data satisfying the analysis conditions were selected by using the time series segmentation method. The system harmonic impedance phase angle was obtained by using the characteristic of zero covariance of independent random variables, and then the harmonic phase angle was implanted into the existing data correlation analysis model. When the user harmonic current fluctuated, the accurate system harmonic impedance module was calculated by the user harmonic current fluctuation. When the user’s harmonic current was stable, the harmonic impacts was quantified by the fluctuation of system harmonic voltage. In this paper, the user harmonic impedance was also given in thoughts in the entire process, which reduced the error caused by ignoring the user harmonic impedance in the traditional method. Some potential Evolutionary Algorithm based solutions which could be further applied in this domain was also reviewed. As a particular novelty of this work, the possibility and feasibility of employing gene expression programming into the conventional data correlation analysis work in power system. Simulation analysis and practical engineering examples showed that this method can effectively suppress the influence of system harmonic change and user harmonic impedance compared with the existing methods and obtained more accurate harmonic responsibility division results.