Abstract:In order to solve the problem of low quality and effectiveness of monitoring work and information inundation risk, when manual monitoring up to millions of information points, after centralized monitoring of power grid substations, to avoid missed monitoring signals and misjudgment signals, which will seriously affect the safe operation of the power grid. This paper considers the mechanism of information occurrence and analyze the law of power grid monitoring information occurrence, proposes the concept of monitoring information event. With big data-driven as its core and causal mechanism analysis, two-stage intelligent monitoring scheme is put forward which concludes the historical information mining, real-time information application. First of all, multi-time precision Apriori algorithm are put forward , which come true by programming with R language to association-mine historical monitoring information, to get event library with information form power grid AVC system，defect management system and so on. Through establishing the characteristic vector for event library and putting forward matching degree concept for different length vectors, intelligent real-time monitoring alarm model is established to form intelligent monitoring scheme based on the split-window display and event list-push. The scheme reduces the unhandled alarm information to about 30%, and significantly improves the power grid.