Industry Electricity Behavior Analysis Based on Electricity Index with Seasonal Adjustment

In order to analyze short-term power consuming behavior of large enterprises, and explore regional economic trends, this paper proposes an electricity index learning from the stock index. Firstly, according to prescribed rules, some representative samples of enterprise users are selected, whose weights describing the differences from their industries and individuals are determined by fuzzy expert evaluation and some other. Then, the selected users’ daily power consumptions are decomposed by STL(seasonal-trend decomposition procedure based on loess) to obtain the periodic components and the working intensity coefficient. Then, each adjusted daily power consumption is weighted and integrated. After that, based on a selected day’s value, the power consumption trend can be displayed. Finally, the analysis shows that the working strength coefficient can reflect the actual production behavior, and the electricity index can show the short-term power consumption change. Moreover, after excluding the influence of temperature, the index has a strong correlation with social economy, which can explain the relationship between investment, output and production, and represent the economy of social subject.