The Research On Reliability Of Networked Control System

Networked control system(NCS)plays an indispensable role in some safety-critical systems.How to effectively evaluate the reliability of NCS is important.Recently,the research on the reliability of NCS has attracted more and more attention.The complexity of the NCS lies in the dynamic characteristic of the control system,the internal complexity of the communication network,as well as the interaction between the communication network and the control system.The traditional reliability analysis method is no longer applicable for the complex NCS.Therefore,the purpose of this thesis is to study how to estimate and optimize the reliability of the NCS when there exist transient faults(transmission delay and packet loss)and malicious cyber attacks in the communication network.The innovation of this paper is to propose a reliability analysis framework to conduct dynamic reliability modeling and statistical analysis of the NCS by combining reliability engineering and control theory.Meanwhile,the influence of unreliable communication network on the control system can be studied and analyzed.This work plays an important role in the early design phase of NCS.In the first place,this thesis proposes a reliability analysis framework for the NCS.The difference equation is used to model the NCS,which is convenient to analyze the occurrence of transmission delay and packet loss in each period and their influence on the mathematical model.A test platform based on TrueTime simulation is built to simulate the NCS and the actual network environment.In view of the performance index of the control system,functional requirements are put forward from the perspective of the control theory.Therefore,the reliability of the NCS is defined as the ability of the system to meet the functional requirements under the unreliable communication network.Based on this definition,the Monte Carlo simulation method is used to obtain the system reliability through simulation and statistical analysis of the NCS,while meeting certain accuracy requirement.Secondly,this thesis considers transmission delay and packet dropout in a more actural situation.The network degradation is modeled and integrated into the reliability analysis framework.Different from previous assumptions that network degradations are independent of each other,we assume that the network degradations are time-correlated.Wiener process is used to describe continuous time delay,and independent homogeneous Markov chain is used to model the packet loss,while the transition probability matrix with polytopic uncertainty is considered.A particle swarm optimization(PSO)algorithm based on Monte Carlo simulation is proposed to estimate and optimize the system reliability.The objective function of the PSO is designed.By searching for optimal controller parameters,the reliability of the system is optimized and the performance of the system output is improved.Finally,this thesis studies the influence of common cyber attacks on the performance and reliability of NCS.Denial of service attack(DoS attack)and deception attack are considered.The optimal DoS attack strategy is analyzed from the attacker's point of view.Semi-markov chain is used to model the intrusion process,the mean time to failure of the system is obtained.The impact of various cyber attack scenarios on system performance is analyzed,which is important in the reliability growth design.