Research On False Data Attack Detection Method In Smart Grid

The smart grid is a coupled network system composed of physical power system and information communication system.The control center through the collection and analysis of data in power system,so as to make a decision to ensure the security and stability of the smart grid.At the same time,the power system is an important basic resource facility,ensuring its safe and stable operation plays a key role in maintaining social stability and normal development of national economy.False data injection attack(FDIA)is a new kind of attack method in recent years.takes advantage of the vulnerability of traditional bad data detection in power system,by injecting false data into the metering device distributed in the power grid,resulting in the deviation of the state estimation result from non-attack state.Thus,the control center can get the wrong system operation state,make the wrong decision,then harm the normal safe economy operation of the power system,and even cause the whole power system to collapse.The current research on false data injection attacks is still in its infancy.therefore,in the smart grid environment,the data holes exist in the analysis system,and the corresponding detection and prevention methods are studied,which has very important theoretical and practical significance.In view of the problem of false data attack detection,this paper mainly does the following work:1.Studies the false data attack(FDIA),the basic principle of the false data attack may cause adverse consequences are analyzed,and design a false attack model data,by using the model for the false detection mechanism to provide the training sample data.2.Designed a real-time false data detection mechanism based on Deep Belief Network(DBN).This mechanism use improved Deep Belief Network,with history data as samples of the false data for identification.3.In the simulation experiment,we verify our work by running four cases on the IEEE-118 and IEEE-300 bus structure.respectively in different affected measurement value quantity,environmental noise level,system and the traditional mechanism of bad data detection value under the influence of the critical value,the design of the real-time detection mechanism based on DBN detection accuracy.