Research On Smart Grid Intrusion Detection Model Based On LSTM

With advances in monitoring,sensing,signal processing,control,and communications,advanced technologies are being integrated into next-generation power systems,the smart grid.On the one hand,the upgrading of the smart grid has made the connection point between the power system and the Internet increase,becoming a more automated power transmission network;on the other hand,in the smart grid coupled with the Internet,how to be the intelligent terminal and the main Providing reliable communication between stations to ensure the stability and security of the power system has become a major difficulty at this stage.Advanced Metering Infrastructure(AMI),the basic information platform for smart grids,is responsible for collecting,measuring,and analyzing energy usage data,and transferring this information from smart meters to data concentrators and then to the utility side.The system,as one of the important components of the smart grid,has become the target of attackers,suffering from the attacks of false injection,distributed rejection and so on.Intrusion Detection System(IDS),as a means to improve the security of the network system,can play a role in ensuring the security of the AMI network.It can monitor the running status of the network and query the status characteristics,so as to hope to find the target of the attack as early as possible,so as to make early warnings,formulate relevant plans,implement protective measures,and ensure the confidentiality,integrity and availability of the system.This paper aims to combine the most popular long-term and short-term memory network(LSTM)with the intrusion detection system,and propose an LSTM-based intrusion detection model,which is applied to the AMI network of the smart grid.The timing and memory characteristics of the LSTM network are utilized to improve the detection of the intrusion detection system.By classifying and filtering normal data and malicious data in the AMI network,malicious attacks are removed,and abnormal network requests are filtered to reduce the security threat of the AMI network.Finally,we give the performance test results of the proposed model through performance test.The experiment shows that the model has better detection effect when the learning rate is 0.01 and the hidden layer is 80.Therefore,we use this learning rate and hide.The layer performs subsequent simulation experiments.Finally,the experimental results are compared with the effects of other existing neural network intrusion detection models.The experimental results show that the detection rate of this paper is about 98%.Other neural network models are lower than the experimental model in this paper,which proves that based on LSTM.The effectiveness of the network's intrusion detection system for AMI network data monitoring.