A method for detecting false data attacks in smart grid based on k-means

False data injection attacks (FDI as) have become the most harmful attack means in smart grid. On this basis, this paper adopts the concept of “node voltage stability index”, and analyzes the influence of FDI as (false data injection attacks) on network node NVSI (Node Voltage Stability Index), and establishes the relationship between them. On this basis, K-means+ clustering algorithm based on CFPSO (constraint factor particle swarm optimization) optimization is used to cluster the nodes in the network and identify the weak links in the network. This project intends to study the vulnerability detection algorithm based on SFD (State Forecasting Detection), and give priority to the detection of nodes with high vulnerability level. In the IEEE30-bus system with a small percentage of attacks, three methods (SFD detection method, conventional target detection method and maximum residual method) can detect few FDIAs; With the increase of the proportion of being attacked, the probability of being detected also increases. Under the condition that every system is detected, the detection rate of SFD can exceed 70%. This paper helps to ensure the safe operation of smart grid.