Research Of The Aggregation Model Of Direct-Driven Permanent Magnet Wind Farms

The purpose of the modelling of direct-driven permanent magnet synchronous generator(PMSG)is to precisely reproduce the steady-state and dynamic characteristics.This model also provides a reliable theoretic and technical support for the research on wind power integration,dynamic and stability analysis of power system.In this thesis,a detailed mathematic model of PMSG wind power system is established.Based on the time scale of power system,time constant of each component in wind turbine and its multi-time scale model,a set of simplified models of PMSG wind turbine are deduced for power system analysis.The models include static model,electromagnetic transient and electromechanical transient model.Accuracy assessment of these simplified models is applied and the validity of the proposed models is verified through simulation.The proposed models lay a foundation for the study of the aggregated model of PMSG wind farm.After the in-depth study on the significance of wind farm aggregation model,the principle of equivalence and the method of wind farm aggregation,a one-machine aggregation method is proposed.The equivalent wind speed of this one-machine aggregated model of wind farm is calculated by wind turbine power curve.This wind farm aggregated model is established by using the electrical parameters and the output power data of the wind farm.In the DIgSILENT PowerFactory software,the availability of the one-machine aggregation method is verified by comparing the simulation results of this one–machine aggregated model with the detailed model and the multi-machine aggregated model of one actual large-scale wind farm under different conditions.The error between the aggregated model and the detailed model can be used as a reference for the correction of simulation results,which can be used for reference in engineering application.In this thesis,the purposed one-machine equivalent method based on the wind turbine power curve can fully reflect the dynamic behavior of the wind farm while taking into account the influence of the wake effect of the wind farm.The average error of the model is within 2.5%.The results show that the precision of the one-machine method is slightly lower than that of the multi-machine method,but the dynamic performance of the wind field at the grid point still can be reflected by using the onemachine method which can reduce the computing resources and save the calculation time with high practicability.Based on the one–machine aggregated model of wind farm,the concept of virtual wind turbine is put forward.An adaptive equivalent model of wind farm based on virtual wind turbine is established.The static power curve of the virtual wind turbine is obtained by BP neural network.In the equivalence model expressed in per unit value,the online modification of the model can be realized by setting the actual number of wind turbine in the wind farm.The results show that the static model of wind farm based on BP neural network can effectively reflect the static characteristics of wind farm.Furthermore,the internal lines influence on the performance of an aggregation model is discussed by simulation.It is shown that under a certain requirement on accuracy,the wind farm internal overhead lines can be ignored in the dynamic study but the internal cables must be taken into account in the electromagnetic transient simulation.