Pre-processing and audit of power consumption data based on composite mathematical statistics model

Abstract

With the promotion of intelligent power consumption information acquisition system, power companies can easily obtain a large number of real and effective data sets from the acquisition system. These data sets can help power companies fully grasp the user's power consumption and then analyze user behaviors and power consumption characteristics. However, due to signal interference and other factors, in the currently running acquisition systems, some “dirty data” is inevitably mixed in. Due to the huge amount of data collected by the power system, the total amount of these contaminated data cannot be ignored, and the data extraction and analysis work caused a great deal of interference. Traditional outlier detection methods rely on the intuitive judgment of experienced staff. This paper attempts to design a power data preprocessing model based on mathematical statistics and using detection methods commonly used in mathematical statistics. This can help power grid workers improve the efficiency of auditing original abnormal data and optimizes the quality of the data sets.