| With the increase of the interconnection scale of power grid and the high penetration of renewable energy,single type oscillation and even broad-band oscillation phenomenon appear,which affects the safe and stable operation of power system.Once the oscillation cannot be suppressed in time,it may cause serious impact on the power system.At present,oscillation mode identification research is still in the stage of using signal analysis methods.And the most of oscillation location methods can only locate the disturbance source generator,which needs further research.The specific research contents of this paper are as follows.A data-driven based mode identification method of broad-band oscillation is proposed based on massive power system oscillation data.Firstly,an improved piecewise aggregation approximation(IPAA)method is used to reduce the dimension of the high-dimensional oscillation data.Then,k-Shape clustering algorithm is used to sort out the massive oscillation data.Finally,mode identification models can be obtained based on Random Forest(RF)algorithm.The identification results of two simulation oscillation cases and an actual oscillation case show that the proposed method can accurately identify the oscillation mode parameters of broad-band oscillation and has higher accuracy than other methods.Aiming at the problem of oscillation location,an oscillation type judgment method based on upper envelope is firstly proposed,which provides basis for the selection of oscillation location method.Secondly,to solve the problem of negative damping oscillation location,the oscillation energy of the generator under different oscillation modes is defined based on data-driven mode identification method and damping torque analysis.And the fault generator can be located according to the variation of oscillation energy.Finally,a fault generator location method based on delay time estimation is adopted to solve the problem of forced oscillation location.In order to further determine the specific location of the fault,a device-level fault location method based on transfer function is studied in combination with the multi-machine system model with generator control equipment.The applicability and accuracy of the two oscillation location methods under different faults are verified by simulation. |