A Data Processing Method for Load Data of Electric Boiler with Heat Reservoir

Feng Xiao, Zhenyuan Li, Baoju Li, Chang Liu, Yuhang Qiu, Shang Wang, Tao Peng

Abstract: Data preprocessing is the first and critical step in various tasks due to existed quite a few dirty records. Although there are some corresponding methods for data from electric power system, those methods are less systematic and hard to handle various situations. In this paper, we integrate a data cleaning procedure that can handle various dirty records types present in the heat-storage electric boiler load data, including duplicated values, missing values, and abnormal values. The procedure is clear and consists of three modules connected in the pipeline, including data filtering, data cleaning, and data normalization in turn. The data cleaning is the core module in this procedure and includes missing points complementation and outliers processing. In order to demonstrate the effectiveness of the procedure, two real load power datasets from the heat-storage electric boiler are used in our experiments and the results demonstrate the effectiveness and availability of this work.