PROJECT TITLE :

Raw Wind Data Preprocessing: A Data-Mining Approach

ABSTRACT:

Wind energy integration analysis usually relies on complicated sensors located at remote sites. The procedure for generating high-level artificial info from databases containing massive amounts of low-level information should thus account for attainable sensor failures and imperfect input information. The data input is very sensitive to knowledge quality. To handle this drawback, this paper presents an empirical methodology that can efficiently preprocess and filter the raw wind data using solely aggregated active power output and the corresponding wind speed values at the wind farm. Initial, raw wind data properties are analyzed, and all the information are divided into six categories in step with their attribute magnitudes from a statistical perspective. Next, the weighted distance, a completely unique concept of the degree of similarity between the individual objects within the wind database and the native outlier factor (LOF) algorithm, is incorporated to compute the outlier factor of each individual object, and this outlier issue is then used to assess that category an object belongs to. Finally, the methodology was tested successfully on the data collected from a giant wind farm in northwest China.