Research On Dynamic Equivalence Of Wind Farms Based On Hierarchical Clustering Of Dominant Variables

With the gradual change of energy structure,the importance of electric power production is becoming increasingly prominent.The third generation of large power network began to develop large-scale new energy power.As clean,technically mature and widely used new energy sources,the scale of wind power grid connected is increasing year by year.However,the output of wind power is not stable because of its randomness and volatility,which has a great influence on the security and stability of the power network.So,it is necessary to deeply analyze the operation characteristics of the power system,after the large-scale wind power is incorporated into the power grid.For the research on the safe and stable operation of power system after wind power integration,the most important issue is to set up a more precise wind farm model.Building a detailed wind farm model will increase the difficulty of modeling the whole power system,and make it difficult to correct the model,and greatly increase the simulation time in the time domain.In practical engineering applications,the whole wind farm is often used as a micro grid,and its influence on the power grid is considered.So,it is necessary to build an equivalent model to characterize the running state of the whole wind farm.In this paper,a dynamic equivalent modeling method for doubly fed induction generator is proposed on the basis of the accuracy of fan equivalence.(1)The state variables describing the operation of the wind turbine are obtained with the method of building a mathematical wind turbine model.The factor that the actual operation state of wind farm units is affected by wake effect,time-delay effect and wind motor control mode,are taken into account in these variables.(2)The principal component analysis method is used to extract the rotation difference and the torque of the wind turbine which can represent the 94% oscillation characteristics of the state variable.(3)The state variable,leading variable and wind speed are taken as the cluster index of wind farm,and the same unit is divided into a typhoon motor by using hierarchical clustering algorithm for homology unit clustering.(4)According to the principle of capacity weighting and voltage invariance,the generating and collecting parameters of equivalent wind motor are calculated.(5)The equivalent model of 24 wind turbines was built on the PSCAD platform to simulate the fault and wind speed fluctuation.The simulation results show that the two leading variables of slip ratio and wind turbine torque are reasonable and effective as clustering index.Compared with the wind speed equivalent model,the dominant variable equivalent model is more comprehensive and more accurate.Compared with the detailed model,the calculation complexity of the dominant variable equivalent model is low,and it can be suitable for the research on power system after wind power integration.In this paper,an equivalent model of wind farms based on hierarchical clustering of dominant variables is proposed.The model is low in complexity,easy to be realized and widely applicable,and can satisfy the actual demand of dynamic analysis of wind power field.The research results provide a good theoretical basis for analyzing the security and stability performance of power system after wind power integration,and it has certain reference value and reference significance.