Power System Transient Stability Assessment Based On Machine Learning

As the economy of China continues to grow,the continuous development of power systems,power toward the long-range,high pressure or UHV direction.The scale of running power systems is becoming more larger,the physical changes are becoming more complicated.As a result of the rapid development of wide-area measurement system and machine learning technology,power system transient stability assessment using data mining and machine learning method is realized Two different kinds of power system transient stability assessment were proposed in this paper.The main research results were summarized as follows:(1)This paper used power system simulation software for transient stability simulation,thus obtained form a sample set of the original parameters.Based on the transient stability characteristics,using rough set theory to reduce initial set of transient stability characteristics,an optimal group of dimension feature attributes was gained,then using least square support vector machine as the classifier.The result is good and effectiveness of the proposed method is verified by simulation.(2)The characteristics of mechanical fault are closely associated with the properties of fault by itself which was considered.It greatly increases difficulty and uncertainty of feature extraction and optimization due to human involvement,and increases the difficulty of the transmission mechanical fault recognition,which weakens the intelligence of machine learning.In this paper,the concept of deep learning is introduced,and deep belief network is applied in power system transient stability assessment.Deep belief network is a typical deep learning method,which can achieve an abstract representation through the combination of low-level features to discover the distributed characteristics.Given the characteristics of deep belief network,this paper proposes a novel method to power system transient stability assessment directly from the original parameters.Finally the effectiveness of the proposed method is verified by simulation.(3)With the penetration level of the wind generation increased,the power system stability will change gradually.The generator rotor angle directly reflected the transient stability of the system,in this paper,the trajectory of rotor angle before fault removal was used as the input of deep belief network to predict generator rotor angle in the process of transient state.