State Cognition And Oscillation Identification Of Grid-connected Renewable Energy Based On Big Data Technology

With the development of big data technology,power system has entered the era of data analysis.With the help of the massive data provided by the wide area measurement system and SCADA system,the power system can be easily evaluated,and the abnormal operation status can be detected and positioned.As the increase of renewable energy permeability,more new abnormal operating status have appeared in the system.Aimed at the abnormal operation state in the development of new energy,this dissertation studies the problem of photovoltaic mismatch,the identification the oscillation types of wind power system and the location of system oscillation source via big data technology.Using data mining and data fusion methods,the status information hidden in monitoring data is explored.In photovoltaic system,mismatch is the key factor affecting photovoltaic conversion efficiency.This dissertation proposes a mismatch location method based on clustering using real-time operation data,and give the response strategy according to the different causes.In the wind power generation system,the sub-synchronous oscillation problem is getting more and more attention,and more and more research on its mechanism is done in recent years.Based on the small signal analysis results of multimachine wind farms,this dissertation proposes an oscillation modal identification method using correlation analysis.The measured data of multi-machine wind farms is Prony analyzed first and the Apriori Algorithm is used to mine the correlation between between the oscillation modes and the wind speed/voltage factors.Moreover,a prediction scheme of the oscillation modes is proposed based on the correlation analysis.The location of oscillation sources is also a key issue in oscillation study.By locating oscillation sources,the oscillation can be suppressed from the source quickly.Aiming at substation level,this dissertation proposes an oscillation location scheme based on evidence theory and support vector machine,which makes up for the limitation of single oscillation location method.The result of location analysis of oscillation energy method,oscillation phase difference method and forced oscillation phase difference location method is fused by evidence theory,and the final criterion is given.The strategy is tested in modal cases based on WECC179 bus power system.