Control Strategy For Damping Low Frequency Oscillations Of Power System With Large-scale Photovoltaic Plant

In recent years,the construction of power system interconnection has been actively promoted in China,and tasks of electric power transmission in long distance have been increasingly difficult.At the same time,in order to optimize the energy structure,new energy power generation has been vigorously developed,especially the state of art of large-scale photovoltaic(PV)plants are very different from a decade ago in terms of either capacity or penetration.Long-distance transmission and grid-connection of large-scale PV plants will significantly change the power system characteristics,making low frequency oscillations more prone to occur.In this context,in order to strengthen the ability of power system stability control,and fully explore the power regulation capacity of PV plants,this paper studies the feasibility of using PV plants to suppress low frequency oscillations,and puts forward damping control methods.Based on electromechanical transient models of single-machine infinite bus system and multi-machine system integrated with PV plant,the influence of PV integration on power system stability is investigated by Lyapunov linearization theory and eigenanalysis.Solutions to weak damping of the power system and low frequency oscillations caused by big disturbance are proposed,which establish the foundation for further research of large-scale PV plants participating in control of power system stability.First,the influence of grid-connection of PV plant on the steady state and transient of the power system,the control performance limitation of PV damping control and the design of controllers are studied in the simplest test system----single-machine infinite bus system.Based on the multiple-input multiple-output small signal model with PV active and reactive power as the inputs,power angle and voltage at the point of common coupling(PCC)as the outputs,the influence of the location of PCC on the steady state and transient of the system is analyzed.The damping control performance limitation due to the directivity of the multiple-input multiple-output system is studied.The decentralized controller for both power angle damping control and PCC voltage control is designed and the control performance is analyzed in frequency domain and time domain.Second,research objects are extended into multi-machine systems.The damping of the system without and with PV integration is compared in 4-machine 2-area system and 16-machine 68-bus system.The results show that the effect of PV plant on the damping of the system can be either positive or negative.Not only PV penetration will affect damping,but different dispatchment of generators in the same PV penetration and different characteristics of loads will also affect damping.Third,mixed-sensitivity damping control is introduced for increasing damping ratios of low frequency oscillations to expected levels while reducing the possibilities of saturation of PV power regulation.Solving the H? control problem in the frame of linear matrix inequality can solve the control design problem with the specific closed-loop damping ratio control target.Combining with pole placement,the additional design requirement can be transformed into the constraint condition to solve.In the 4-machine 2-area system and 16-machine 68 bus system,damping controllers designed by using traditional residue method and S/KS mixed-sensitivity H? control are analyzed and compared.The results show that the controllers designed by mixed-sensitivity approach can adjust multiple weak damping modes once,while only a single mode can be controlled using residue method.In the study case of 4-machine 2-area system,it has been found that the use of mixed-sensitivity damping control will produce a new low frequency inter-area oscillation mode,which occurs between the PV plant and the synchronous generator group in another area,and the frequency is even lower than ordinary inter-area modes.Considering that controllers designed based on a single operating point can not cope with the large deviation of operating point after large disturbances,the multi-model adaptive damping control strategy is proposed.Low frequency oscillation modes being the multidimensional features,K-medoids algorithm is used to cluster models associated with operating conditions,and damping controllerers are designed based on individual centroid models of the clusters respectively.Based on the deviation between the output dynamic responses of the actual system and models,the probability of each model representing the actual system in real time is calculated through Bayesian approach.Weights of controllers' outputs are updated,and the weighted average of the individual controllers' outputs is attached to the reactive power control of the PV plant as the resultant damping control signal.The time domain simulation conducted in the 4-machine 2-area system integrated with the PV plant shows that the proposed control strategy is efficient for damping low frequency oscillations caused by unexpected disturbances.Robustness against time delays and random fluctuation of irradiation is also verified,and the control effort is lower than traditional individual controllers with similar control performance.