Research On Detection And Location Technology Of Subsynchronous Oscillation In Wind Power Integrated System

Subsynchronous Oscillation(SSO)is an important stability problem in wind power integrated systems.Several major SSO accidents caused by wind power have occurred both at home and abroad.SSO caused by wind farms may cause damage to wind farm equipment,causing break down of renewable energy units,inducing shaft torsional vibration,etc.,which affects power system stability greatly.Therefore,it is necessary to identify and locate the SSO mode in wind power integrated systems accurately and timely,to provide a basis for further control measures.In this paper,the main theme is the SSO problem in wind power integrated systems and research in system modeling,SSO detection technology and SSO location technology are carried out,and the main works are stated as follows.(1)Modeling of wind power integrated system for SSO analysis is studied.The structure of Doubly Fed Induction Generator(DFIG)and Permanent Magnet Synchronous Generator(PMSG)is investigated.The mathematical models of the shaft system model,the generator and the transmission line with series compensation capacitors are derived,and the models of the rotor-side and grid-side controllers are established.In addition,the DFIG and PMSG integrated systems are established on the MATLAB/Simulink based on the above models.Furthermore,the typical SSO phenomena mentioned in literatures are simulated.(2)A method of SSO signal detection based on higher-order accumulation and ESPRIT is proposed.Firstly,the signal pre-processing method is improved by using a seven-point second-order algorithm to to identify the anomalous data,and then sequentially perform the correction,filtering and de-DC processing.The original signal is replaced by Higher Order Cumulant(HOC),and then is identified using the Estimating Signal Parameter via Rotational Invariance Techniques(ESPRIT)algorithm.The results show that the proposed method significantly improves the computational speed and noise immunity compared to the original ESPRIT oscillation detection method,laying a good foundation for online monitoring of SSO in wind power integrated system.(3)A correlation analysis method for oscillation mode based on bispectrum-bicoherence estimation is proposed.Firstly,bispectral estimation is introduced to identify the SSO signals in the system,to obtain the possible quadratic phase coupling relationship between each oscillation component.Furthermore,bicoherence analysis technique is used to identify the authenticity of the quadratic phase coupling relationship mentioned above.Results of the case study show that the bispectrum-bicoherence estimation method can effectively extract the inter-mode correlation of SSO in the system,which can provide technical support for further cognition of the multi-mode SSO problem.(4)A method based on transient energy flow and Naive Bayes Classifier(NBC)is proposed for locating SSO disturbance sources,and the SSO disturbance source contribution degree is assessed based on fuzzy inference.Firstly,the transient energy flow method is applied to locate SSO sources in wind power integrated systems and the practical location method based on NBC is proposed.Furthermore,a Fuzzy Inference System(FIS)based SSO source contribution rating method is proposed.The method obtains SSO disturbance contribution degrees for each area by constructing,calculating and judging proposed SSO indicators using FIS.Results of cases show that the location method combining NBC and transient energy flow theory can quickly locate SSO disturbance sources,and the FIS-based method to evaluate SSO source contribution can clearly perceive the SSO contribution degree in each area,identifying the dominant oscillation disturbance sources,which can provide operational reference for emergencies.