Research On Interconnected Power System Frequency Fluctuation Control With Large Scale Wind Power Cluster By Predictive Control Method

With the increasingly severe problems of fossil energy shortage and environmental pollution,as a relatively mature clean power generation technology,wind power has been enlarging its penetration in power system and replacing some old traditional thermal power units.However,wind power has characteristics of volatility and uncertainty,and it cannot respond to the change of power system frequency timely because of the power electronic converter operation technology,which decouples the mechanical and electrical parts of wind power units.In this way,the power system frequency stability could be damaged in the case of high wind power penetration level.Recently,several ten-million kilowatts wind power bases have already been built and planned in China,and relevant wind farm integration standards of restraining wind power fluctuation have also been introduced.Therefore,how to flexibly control the active power output of wind power clusters within the effective output range of wind power prediction,making wind power clusters ha-ve the ability of frequency control in coordination with traditional AGC power units in order to restrain power system frequency fluctuation is becoming an urgent problem to be solved.On the basis of a comprehensive review of domestic and overseas research results,and the combination of model predictive control theory,interconnected power system frequency control theory,large systems cybernetics and stochastic programming theory,a predictive control method considering interconnected power system frequency fluctuation control with large scale wind power cluster is established in three control thought levels of distribution,hierarchy and stochasticity in order to solve the problem of frequency fluctuation and improve power system frequency stability.The main research results and innovations are described as follows:(1)In the level of "distribution" control thought,a multi-area interconnected power system advanced frequency control strategy considering large scale wind power cluster integration based on DMPC was proposed.Firstly,a multi-area interconnected power system model with large scale wind power cluster and traditional power was established.The model took into consideration of several nonlinear physical constraints,which included wind power generation rate constraints,traditional power operation dead zone,power generation rate constraints and time delay;Secondly,a Laguerre function based distributed model predictive control strategy concerning the ultra-short-term forecasting information of wind power was proposed.The strategy operated online rolling optimization with the target of output variables minimization and control variables variation weighted restraining in order to realize advanced frequency control;Finally,a Nash equilibrium coordination online optimization algorithm was proposed to solve the distributed model predictive control strategy using Laguerre functions,which searched Nash equilibrium points to realize online fast optimization of multivariable systems with constraints.The simulation results showed that the proposed strategy and algorithm could realize the coordination of large scale wind power cluster and traditional power in frequency control,and satisfied the computing speed requirement of the online rolling optimization in the multivariable system.(2)In the level of "hierarchy" control thought,an integrated frequency control strategy for wind power cluster with multiple temporal-spatial scale coordination based on H-DMPC was proposed.A variable granularity frequency control framework with wind power cluster participating in third frequency control layer,secondary frequency control layer and primary frequency control layer was established in this control strategy.Firstly,prediction horizon,control horizon and the ultra-short-time wind power forecasting data with coordinated temporal scale were used in each layer.Secondly,a rolling optimization model with coordinated temporal and spatial scale was established in each layer.In third frequency control layer,power system topology and economically optimal objective were considered.In secondary frequency control layer,average system frequency augmented model and areal securely optimal objective were considered.In primary frequency control layer,system load stochastic disturbance with small magnitude was reduced by means of dynamic classification of wind farms in areal wind power cluster.Finally,according to the real time operation state of power system and wind power cluster,a feedback correction link was established to realize closed-loop control.Case study results of IEEE 39-bus system demonstrated that the proposed control strategy could effectively realize power system frequency control with wind power cluster participation and gave consideration to both economic and secure objectives in power system.(3)In the level of "stochasticity" control thought,a chance constrained goal rolling programming method for wind power cluster frequency control based on S-H-DMPC was proposed.Firstly,a wind power cluster forecasting error model considering wind power temporal correlation was established.Secondly,a bi-level chance constrained goal rolling programming method considering wind power cluster forecasting error random vector was proposed,in which power system topology and system economic objective were focused in upper-level,and average system frequency augmented model and system secure objective were focused in lower-level.Finally,a solution method based on Monte Carlo stochastic simulation was proposed,in which an affine transformation algorithm was utilized to sample wind power cluster forecasting error random vector,and to handle chance constraints.Case study results of WSCC 3-machine 9-bus system and IEEE 118-bus system demonstrated that the proposed control method could effectively improve the accuracy of wind power cluster participating in power system frequency control,and the feasibility and effectiveness of the proposed control method were also proved.