Research On Key Techniques For Load Frequency Control In Renewable Energy Power Systems With Coupling Cyber-Physical Effects

Integration of high-proportional renewable generation systems and application of large-scale opening-sharing communication networks are two main characteristics of the modern cyber physical power systems.Under this background,the frequency stability problem is challenged by both physical and cyber layers.More specifically,1)The large-scale access of renewable generations with weak inertia reduces the proportion of traditional synchronous generator units with rotating inertia.As a result,the power systems lack sufficient inertia support and enough frequency regulation capacity in the presence of dual-stochastic fluctuations of load demands and renewable generation outputs.Consequently,the frequency deviation may be too large or even beyond the limit.2)With the continuous expansion of the scale of interconnected power systems and the number of devices participated in frequency regulation,the statuses of generating devices engaged in frequency regulation and the control instructions sent by the control center should be transmitted over the communication networks.However,the delay,packet loss and malicious network attack risks in open communication networks will lead to the decrease of the stability margin or even instability of frequency control system.3)Besides the malicious hacking attacks at cyber side,sudden equipment failures at physical side may also lead to structural mutation and large-scale operating parameter transfer.The above issues put forward higher requirements for the design of load frequency control(LFC),which is an important approach to maintain real-time active power balance and fast frequency deviation damping performance.In this thesis,the load frequency control strategy is studied from the following four aspects: 1)inertia ehnamancement and frequency regulation capaibility improvement at physical layer;2)data flow transmission characteristics analysis;3)networked control strategy design;4)fault-tolerant ability improvement in the presence of cyber/physical failures.Main research contents and corresponding contributions are as follows:1.From the physical dimension,a load frequency control scheme which utilizes virtual synchronous generator(VSG)based electrical energy storage systems(ESSs)to provide auxiliary frequency regulation service is proposed.A real-time allocation method for frequency regulation task considering limited state-of-charge(So C)of ESSs is proposed to improve the frequency regulation capacity of LFC system.Secondly,on the basis of estabilishing the small-signal model of the VSG-ESSs,a LFC framework containing renewables,loads,ESSs ans traditional synchronous generators is proposed.Meanwhile,the corresponding state-space model is deduced.Finally,condsidering the effects of modelling errors and external power disturbances on the frequency deviation damping performance,a linear quadratic Gauss based controller is designed.The simulation results show that the proposed method can effectively improve the frequency devation damping performance in the active power mismatch case.Moreover,the overcharge or overdischarege of ESSs can be effectively avoided.The corresponding service lives of ESSs can be prolonged.2.From the cyber dimension,a networked LFC strategy based on deterministic network calculus and jump control theory is proposed.The upper delay bound of communication network with multi competing data flows is deduced with deterministic network cauculus theory.Then,the mapping relationship between the dynamic dynamic response characteristics of LFC system and the bounded stochastic transmission delay is accurately described based on switching control theory.The asymptotic stability constaints with bilinear matrix inequality form is deduced via constructing a Layunov function.In addition,to minimize the objective function about the peak valve,peak time and adjustment time of frequency deviation,the controller design problem is transferred into a constrainted optimizaiton problem.Simuliation results with a mutiarea LFC system show that the proposed method can effectively avoid the optimistic or conservative design and obtain a satisfactory dynamic performance of frequency deviation damping.3.Considering that the injection of abnormal data flows in a distributed fashion will cause the significant increases of the transmission delays of the normal data flows and the failure of the control law designed according to normal transmission delay and packet loss,an adaptive load frequency control strategy based on deterministic network calculus is porposed.Firstly,with DNC theory,the analytic relationships among the upper bound of tranimission delay and key parameters of abnormal data flows and communication networks.Secondly,the LFC system is divided into several operation scenarios according the time delays.Second,via construcing Lyapunov-Krasovskii function,the constraints guaranteeing asymptotical stability of the closed-loop system in different scenarios are derived.Finally,the controllers corresponding to each operation scenario constitute a set.The most matching controller gain is selected to be switched into the control loop according to the actual transmission delay of data packets.Compared with the traditional fixed-gain based LFC schemes,the porposed adaptive control can obtain better dynamic performance different delay-attack intensities.Simulation results in Three-area interconnected powery systems show that the proposed method ensure that the LFC system can effectively resist the distributed traffic attacks and thus enhancing the operational reliability of the power cyber physical system.4.Frutherly,considering the impacts of sudden physical faults in physical layer and malicious network attacks in information layer on the safe and stable operations of load frequency control in power systems,an active fault-tolerant load frequency control strategy based on gain scheduling is proposed.Firstly,for the possible faulty operating scenarios,the targeted controllers are designed in a one-on-one way and formed into a basic controller library.Secondly,in order to solve the problem that the traditional gainscheduling based AFTC methods rely on accurate prior system parameters to construct the fault identification module,a dynamic behavior reconfiguration scheme and a onetime switching method for the most matching controller are proposed.Finally,the simulation results show that the proposed method can achieve the controller effect similar to the traditional observer-based AFTC methods without relying on the prior system parameters.Therefore,a better practicability can be obtained.