Phase Theoretical Line Calculation And Analysis Based On Pattern Recognition Of Daily Load Curves

The line loss is an important index to measure and examine the economic effectiveness of power supply enterprises. At the present time, comparative analysis results of theoretical line losses calculation and statistical line losses obtained using a time section or typical are not enough accurate, and not enough to describe long-term operational performance of power grid. Then a method about phase theoretical line loss calculation based on pattern recognition of daily load curves is proposed in this article.Firstly, the impact of the network topology changes to flow distribution and the theoretical line loss is analyzed and a brief analysis of the grid structure identification’s significance and research methods is given. The time is divided into several sub-stage time periods with the identification of network topology’s changes, if there is big change of network topology when the theoretical line loss is of a certain time is calculated. This paper defines the influence’s degree of the increase or decrease plant’s nodal injection power and increase or decrease line to the grid structure, and identifies the network topology with the correlation matrix.The daily load curves with the same type of week have similar lever and similar fluctuation trend in the same sub-stage time period, if the weather, temperature and other external is similar. The maximum of daily load curve, the average of daily load curve and the shape coefficient of daily load curve is introduced as section feature vectors in order to perform pattern recognition of daily load curves and made the daily load curves with same lever and similar fluctuation trend in the same category.Finally, the work section pattern recognition to every daily load curve pattern’s 24 sections data is analyzed. Because the total load of the sections at the same hour in the same pattern is similar and the user load doesn’t fluctuate greatly in short term, so the flow distribution and the theoretical line loss of the same hour is the same pattern if the main power plants’ output doesn’t change largely. And then on the basis of the identification of power plant output and load shedding, the work section pattern recognition to every central daily load curve’s 24 sections data laterally and the section pattern classification to other curves’ same time sections is proceeded using rough set theory with setting up dual-threshold. After pattern recognition, the phase line loss is calculated with correction of the network loss sensitivity.