Fault diagnosis for wind power generation system based on association rule mining

Wenju Ji, Jianwen Wang

COMPUTER MODELLING & NEW TECHNOLOGIES 2013 17(5D) 105-108

Inner Mongolia University of Technology, Hohhot, 010051

This paper concentrates on the problem of fault diagnosis for wind power generation system, which is a crucial problem for wind power industry. Firstly, framework of the fault diagnosis system for wind power generation is presented. This framework is made up of two main parts, that is, “local device module” and “remote diagnosis center”. In the local device module, wind turbines are connected to other servers through lower computers, and then data is transmitted to the remote diagnosis center. Furthermore, the remote diagnosis center can receive the data transmitted by the local devices and then discover faults by the proposed association rule mining algorithm. Secondly, in the proposed association rule mining algorithm, the opportunity and effectiveness of a specific rule is represented as the number of chances to utilize this rule and the average utilization ration of this rule, and then the rules with higher probability are preserved to conduct the Fault Diagnosis. Finally, specific wind power generation equipment is used to test the effectiveness, and experimental results show that the proposed method can discover different kinds of faults in the power generation system with high accuracy.