Credibility Evaluation of the Phase Identification Algorithm Based on Data Analysis in LVDN

In the low-voltage distribution network, it's hard to know the identification accuracy of the phase identification algorithm. To solve this problem, this paper proposes a credibility evaluation method which can assess the results of phase identification algorithm, based on data analysis, in the low-voltage distribution network. In this paper, we start from the mechanism of the phase identification algorithm based on the data analysis. Then considering the effect of data quality measured by smart meters, we propose five evaluation indicators of credibility including completion of data, completion of valid meters, unbalance of three-phase voltage, obvious users rate and ratio of time to meters. Next, we build the credibility evaluation model of phase identification result based analytic hierarchy process and polynomial fitting. Finally, simulation results using the real meter datasets of two real low-voltage distribution networks in Guangdong, China to demonstrate the performance of our method.