A New Method Of Online Recognition Of Coherent Generators Based On Wide Area Information

With the development of large-scale interconnected power system,the scale of power grid is constantly expanding and the number of electrical components contained in a power system is lager.As a result,the analysis of power system dynamic process becomes more difficult.By identifying the coherency generators in the power system and merging them into one equivalent generator,the scale of power system can be effectively reduced,the efficiency of system dynamic analysis process can be greatly improved.In addition,some large interconnected power grid events about cascading failures will cause immeasurable loss,so there is an urgent need to improve the construction of the "three lines of defense".Out-of-step splitting control of power system is the last line of defense to guarantee the security and stability of the power system,and identifying coherency generators rapidly and accurately is the premise for splitting control.Therefore,the research on online coherency grouping has great theoretical and practical significance.The results of traditional coherent grouping always rely on the accuracy of the system model.With the development of wide area measure system(WAMS)and the emergence of communication technology,it provides some new directions for generator coherency grouping study.Phasor Measurement Unit(PMU)in WAMS can provide rich data sources for the power grid.As a result,some coherent grouping methods based on wide area measurement information are developed.An online coherency clustering method based on wide area measurement information is proposed in this paper.It is based on the abundant information contained in real time power angle data acquired by PMU.Two clustering indexes are proposed in this method.Effective values of power angle difference index is suitable for rapid pre-grouping of generators in a short period.The comprehensive clustering index ?Hsim,constructed by combining Hsim function and Pearson correlation coefficient,is suitable for strict further-grouping of generators.It takes the trend difference and distance difference of generator power angle trajectories into account at the same time.On this basis,in order to meet the requirements of online recognition of coherent generators,an improved clustering algorithm combining k-means clustering and QT(Quality Threshold)clustering is proposed.The proposed algorithm can output the clustering result with only one artificial parameter,and the number of clusters requires no consideration.The simulation results of EPRI-36 node system verify the effectiveness of the proposed method.In the pre-grouping stage,by comparing the effective values of the power angle difference among all the generators,the generators with large power angle difference can be clustered into different groups rapidly.In the further-grouping stage,by calculating the comprehensive similarity measure index ?Hsim among all the generators,it is possible to achieve generator grouping results at different scales and levels of detail.