In this paper,a combined Characteristic Ellipsoid(CELL)and Decision Tree(DT)method for fast classifying the transient stability of power systems after a large disturbance is proposed.The proposed two-stage method invo...In this paper,a combined Characteristic Ellipsoid(CELL)and Decision Tree(DT)method for fast classifying the transient stability of power systems after a large disturbance is proposed.The proposed two-stage method involves projection of the PMU measurement data after the disturbance on a multidimensional space to form the CELL and then classification of the transient stability using DT which takes the characteristic attributes of CELL under different fault scenarios as input features.The dynamic behaviors after a disturbance for both stable and unstable situations are identified from the variation of the CELL’s shape.The database,composed of the geometrical properties of the CELL such as volume,eccentricity,center and change rate of volume,is used to train a DT for transient stability classification.Case study on a IEEE 39-bus system demonstrates the feasibility of the proposed method.Investigations show that the proposed method requires less information from the system to fast classify the transient stability with high precision.展开更多
基金supported in part by the National Natural Science Foundation of China(NSFC Project,No.51437003).
文摘In this paper,a combined Characteristic Ellipsoid(CELL)and Decision Tree(DT)method for fast classifying the transient stability of power systems after a large disturbance is proposed.The proposed two-stage method involves projection of the PMU measurement data after the disturbance on a multidimensional space to form the CELL and then classification of the transient stability using DT which takes the characteristic attributes of CELL under different fault scenarios as input features.The dynamic behaviors after a disturbance for both stable and unstable situations are identified from the variation of the CELL’s shape.The database,composed of the geometrical properties of the CELL such as volume,eccentricity,center and change rate of volume,is used to train a DT for transient stability classification.Case study on a IEEE 39-bus system demonstrates the feasibility of the proposed method.Investigations show that the proposed method requires less information from the system to fast classify the transient stability with high precision.
