Subsynchronous Oscillation Characteristics Analysis Of D-pmsg Based Wind Farm Connected With Power Grid And Its Model Order Reduction

The large-scale wind power connected to the power system has caused a new subsynchronous oscillation（SSO）problem.Solving this problem is conducive to the safe and stable operation of the system.Due to the characteristics of "large number of units,various operating conditions and high model order" of wind farm,it is an inevitable trend to reasonably reduce the order and simplify the detailed model of wind farm when SSO analysis is carried out.Therefore,aiming at the SSO problem of direct drive permanent magnetic synchronous generator（D-PMSG）based wind farm connected to the power grid,based on the analysis of SSO characteristics,this paper carries out the research on model order reduction of D-PMSG based wind farm for SSO analysis.The main research contents and innovative work of this paper are as follow:（1）The D-PMSG based wind farm connected with weak power grid is prone to SSO,and the influencing factors are complex.The full order state space equations of D-PMSG based wind farm connected into power system is established.The SSO mode is identified by the characteristic structure method,and the dominant factors affecting SSO characteristics are determined according to the participation factors.On this basis,the general law of the variation of SSO characteristics when different factors change is analyzed.The correctness of analysis results is verified by timedomain simulation with PSCAD/EMTDC.（2）It is difficult to carry out SSO analysis when the structure or parameters of D-PMSG are unknown.The state space impedance model of D-PMSG is deduced and an identification method of D-PMSG impedance characteristics based on polynomial fitting is proposed.Different from the traditional idea of taking the transfer function of the unit’s impedance model as the identification object,this method takes the impedance characteristics of D-PMSG under subsynchronous frequency band as the identification object,and replaces the transfer function with curve fitting.This method is simpler and faster than the traditional one,and has clear indicators to measure the accuracy and help to determine the order.Then,based on the identification model,the impedance characteristic model of the whole system is established,and SSO characteristics is analyzed through the impedance determinant criterion.The effectiveness of the method is verified by comparing with the previous conclusions and building a simulation model in PSCAD/EMTDC.（3）The order of detailed wind farm model is high,which makes SSO analysis difficult.Based on the balanced truncation method,the order reduction of D-PMSG based wind farm model for SSO analysis is studied.For the two scenarios of stability and instability,the chain and parallel collector network structures,and the scale of 50 and 100 D-PMSGs-based wind farm,the reduced order model which can maintain the characteristics of SSO is obtained.The application scenario of balanced truncation method is given according to the calculation time.Compared with the existing studies,this study can fully take into account the differences between unit working conditions and collector lines.（4）The "curse of dimensionality" will be encountered when calculating the eigenvalues based on the detailed model of wind farm.A reduced order calculation method of SSO mode of D-PMSG based wind farm connected with power grid based on Arnoldi method is proposed.Firstly,the average wind speed is used to form a single machine equivalent model to obtain an approximate SSO mode,and then the characteristic equation of the original system is displaced with this mode as the reference point.Finally,the SSO mode meeting the accuracy requirements is obtained by iteratively increasing the target order.The effectiveness,applicability and rapidity of the proposed method are verified under stable and unstable states,different collector network structures and different wind farm scales.