Research of alarm correlation technique for edge cloud computing in power communication network

Abstract

The demand for electric power and the coverage of power communication network has been gradually expanding, which brings the challenges that need to be addressed in the operation and maintenance of the domain. The deployment of edge nodes provides the availability of data collection, information filtering and computational support, which can heavily alleviate the pressure of management. Alarm analysis is a key and difficult problem in operation and maintenance. The traditional alarm analysis first uses rules to merge alarms, so as to reduce the workload of subsequent processing. However, the completeness of rules requires a lot of expert knowledge and human resources investment, and there are limitations. This paper proposed a novel and lightweight algorithm. In this paper, Unsupervised clustering is introduced into the alarm merging process of power communication network edge cloud computing, and the density based clustering method is combined with the existing merging rules. The experimental results show that the effect of alarm merging can be significantly improved by adding unsupervised learning, which is helpful to improve the accuracy and completeness of subsequent defect location.

Keywords

• unsupervised learning
• dbscan
• alarm correlation
• edge computing