Wind Turbine Condition Monitoring based on SCADA Data Analysis

There is a correlation between two parameters in the data collected by the SCADA system of wind turbine. In this paper, SCADA parameters were related to each subsystem, and their correlations can be found according to the principle of wind turbine operation. The Polynomial Regression was used to establish normal relations matrix between two parameters in different wind speed. The distance between undetermined SCADA data relationship matrix and normal SCADA data relationship matrix was treated as the evaluation criteria. Based on SCADA data collected before a failure occurs of two wind turbine located in a wind farm of Zhangbei County, MATLAB was used to verify the validity of this method. The result confirmed that this method can reflect faults two months ahead of the SCADA system.