GPS Spoofing Attack And Defense For PMU

The global positioning system(GPS)receiver of the phasor measurement unit(PMU)uses radio frequency(RF)to receive civilian GPS signals from different satellites.Since the detailed information of civilian GPS signals is public,malicious users can build devices that generate fake GPS signals and carry out GPS spoofing attacks on the power grid.The attacker constructs a pseudo-GPS signal and induces the GPS receiver to receive the pseudo GPS signal,and introduces the wrong phase angle into the PMU measurement value,which causes the state estimator to estimate the incorrect system state,thereby affecting the normal operation of the entire power system.Therefore,it is of great significance to study GPS spoofing attacks and defense methods for PMU.Aiming at the GPS spoofing attack in smart grid,this thesis adopts the detection method based on CapsNet-GRU,and proposes a defense model based on improved generation countermeasure network to realize the effective defense against GPS spoofing attack.The specific research contents are as follows:(1)The dynamic model of power system and the principle of GPS spoofing attack are analyzed,and the influence of GPS spoofing attack on PMU measurement is analyzed based on this,which provides data support for the following detection and defense research.(2)An improved CapsNet is obtained by adding the gated recurrent units(GRU)structure to the front of the full connection layer of capsnet.Then,a GPS spoofing attack detection model based on CapsNet-GRU is constructed.Finally,experiments are carried out in different test systems.The experimental results show that compared with other detection methods,the method in this thesis has better performance.(3)This thesis studies the nonlinear regression model of measurements,introduces deep learning parameters into the model,and constructs a new network physical model,which can use historical data to calculate the measurements at the current time.A time attention module is integrated into SAGAN's generator and discriminator networks,and an improved SAGAN defense algorithm is proposed to implement the network physical model.Finally,the good performance of our defense mechanism is verified by experiments.