Robustness Analysis And Optimization Of Cyber Physical Power System

With the increasing development of Internet of Things and 5G technology,more and more smart devices are put into the power network,which makes the tradition-al power network gradually evolve into a cyber physical power system(CPPS).Such integration provides 3C function(Communication,Computing,Control,3C)to moni-tor the running status of the power network in real time.However,it also brings new challenges to the safety and the stable operation of CPPS.The types of initial failures in CPPS can be divided into two categories:physical failures and cyber failures.Physical failures mean that the nodes or links directly suffer physical damage,such as the fall out of power stations and the cut off of transmission lines.Physical failures will not only change the load distribution in their own network,will also affect the performance of coupled network,and even trigger cascading failure process.On the other hand,the cyber failures refer to the tampering or absence of measurement data during the transmission.Unlike physical failures,cyber failures do not directly damage network nodes or links,but indirectly affect CPPS by manipulat-ing data.Such stealthy attack also poses a high risk to society and national security.Therefore,this thesis will focus on the modeling,robustness analysis and optimization of CPPS for in-depth and detailed research.Specifically,1.In order to better carry out the research work on CPPS,based on the cascading failures model considering the electrical characteristics,we analyze the robust-ness of the power network.We refine several topological indicators to quantify the global connectivity of the network,the location centrality and the connection efficiency of the generator nodes.Based on these topological indicators,we do the simulation and optimization for different kinds of networks.The results sum-marize the trend of these indicators that can improve the robustness of the power network.Then,we further propose a protection strategy,namely link switch off,to improve the robustness of the power network.By establishing a multi-objective optimization problem,we study how this protection strategy affect the cascading failure process with the objective of robustness and operating cost-s,and quantitatively reveal the efficiency of proposed strategies under different operating costs.2.In terms of CPPS,we propose a more realistic interdependent model which con-sider a weak interdependency,and analyze the influence of several critical factors on CPPS.The proposed weak interdependent model takes into account the realis-tic redundancy protection mechanism,which makes the model more closer to the actual system operation.Based on the model,we study the impact of serval criti-cal factors,such as load redundancy coefficient of nodes and failure propagation probability between two networks.Based on the simulations on IEEE standard data sets and classic complex network models,we fully reveal how these critical factors affect the robustness of CPPS.3.False data injection attack(FDIA)and Denial of Service(DoS)are the most typ-ical cyber attack methods in CPPS.Based on the mechanism of state estimation and bad data detection,FDIA can deceive and mislead the control center to make invalid or even wrong instruction by injecting attack vectors.By sending a large number of pseudo data packets to the channel of the communication network,DoS attack occupies the communication resources between the control center and the remote terminal,which affecting the judgment of control center,and threatening the safe and stable operation of CPPS.On the basis of FDIA,com-bined with the characteristics of DoS attack,a hybrid attack model is proposed to re-evaluate the robustness of CPPS.Finally,we established an optimization to study the impact of this hybrid model on the robustness of CPPS under two attack scenarios.The more effective the proposed model,the greater the scope for improvement.It can also provide some insights for designing more robust algorithms and modules.4.We also propose a robustness improvement optimization,which based on traffic adjustment of the communication network to improve the robustness of whole CPPS.The model considers the electrical characteristics and the traffic law of the power network and the communication network,and the interdependency be-tween these two networks into multiple equations or inequality constraints.By introducing slack variables,the original optimization problem is further trans-formed into a two-level mixed integer programming problem,which can be e-quivalently solved.Finally,we study the impact of the severity of initial failure,the coupling strength,and the topological structure of the communication net-work on the optimization results.