Research And Application Of Power Grid Alarm Analysis Based On Data Mining Technology

In recent years, upload power grid alarm signals increased significantly, showing more complicated alarm information to the dispatcher. These power grid alarm information exit a large amount of noise data, seriously affecting the speed and accuracy of incident handling. Meanwhile, dispatchers can not extract truly important information from the alarm information at the first time because of the lack of rule-based reasoning. Therefore, research of power grid alarm analysis is of great significance. This paper studies the application of data mining techniques in power grid alarm analysis.Firstly, the paper deeply studies the technology of ETL and the data warehouse, and describes the establishment of the power grid alarm data warehouse process in detail.Then, the paper applies bayesian decision tree algorithm into remove noise data from power grid alarm information and specifically addresses the implementation process of achiving denoising classification tree and denoising classification rules. The example shows that the algorithm can improve denoising rate.Lastly, the paper applies association rules method into dig out the relationship between line fault type and fault alarm information. And then the paper can achieve line fault type diagnosis based on matching the real-time alarm information and the association rules, providing references to the dispatcher and preventing the further development of the line fault.Result of the method proposed above is achived by programming. And the examples can effectively remove noise data and achieve line fault type.