Data mining with improved Apriori algorithm on wind generator alarm data

Alarm system is an important subsystem of Supervisory Control and Data Acquisition (SCADA) system in wind turbines. Because of the bad operating environment and condition switching frequently, this system has the problems of alarm too frequently and too large amount of alarm data which greatly reduce the effectiveness of the alarm system. Remove the excessive redundant alarm and refine the valid information are significant to early find wind turbines abnormal operation, to research the causal connection between different faults and to reduce the operators' workload. Data mining is an effective way to solve this problem. This paper used association rules on improved Apriori algorithm to analysis the alarm information which happened before and after blade angle asymmetry fault. Combined with the running mechanism of variable-pitch systems, we find the implied causal relationships between faults, then filter out minor redundant information, refine effective leading fault alarms and at last greatly reduce the number of alarm, improve the operators' work efficiency.