Research On Fault Classification And Stability Of The Smart Grid Considering Communication Performances

The smart grid is an ideal form of future power grid,and its scale and complexity will greatly increase the difficulty and vulnerability of power system operation and control.Moreover,due to the large-scale access to the grid by intermittent power supplies,flexible AC transmission technologies,and flexible loads associated with real-time pricing,power system operation and control will be highly dependent on information feedback and decisions.As a result,the smart grid will develop into a complex combination of the physical energy system and the information system,namely the cyber-physical power system(CPPS).Under CPPS framework,energy flow and information flow will be deeply integrated and interacted to determine the functional and behavioral characteristics of the power system.It can be seen that the analysis and control technology of the smart grid will be inseparable from the influence of the power cyber system.At present,there is no systematic modeling and analysis method for CPPS.This paper envisages that the future wide-area protection and control system will be based on the existing wide-area measurement system(WAMS),and thus takes WAMS as the representation of the cyber system under the CPPS framework.Further,this paper explores the modeling of information flow,fault classification based on time-domain sampled data and the coupling dynamics of cyber-physical system.This paper first expounds the composition of time delay of the wide area measurement system(WAMS),establishes the mathematical model of time delay in WAMS,and improves the model in view of the existing model without considering the existence of processing delay in the PDC.Moreover,based on the double adjacency condition search algorithm,a matrix cyclic search algorithm is proposed to figure out the transmission paths from the PMU substation to the WAMS control center.The routing scheme is determined according to the communication delay and reliability of each path.Then,to cope with packet dropout that may occur in the communication network,the interpolation algorithm is applied to predict the missing sampled data.On this basis,the genetic algorithm is applied to optimize the initial weight and threshold of the BP neural network,so as to classify the fault according to the time-domain sampled data.The fault classification ability of BP neural network optimized by genetic algorithm is verified in a demo where transmission line fault is executed.Finally,the time-domain simulation of power system is carried out by using the power system toolbox packages based on MATLAB platform.The effects of protection control communication delay and excitation control delay on power system stability are respectively investigated with a single-machine-infinite-bus system and a three-machine-nine-bus system.Additionally,suggestions are made on how to reduce the protection control communication delay and the excitation control delay.