Research On Low Frequency Oscillation On-Line Monitoring And Forced Oscillations Suppression In Power System

Scholars and researchers have made many achievements in power system oscillations, which were widely used in field and achieved significant benefits. But with the development of power system, which contains the sharp increasing numbers of components and a higher proportion of new energy, a variety of different types of low-frequency oscillation have founded, the challenge of analysis and control of low-frequency oscillation is becoming more difficult. Considering the increasing complicated trend of low frequency oscillations problems, this paper further studies on-line monitoring of low frequency oscillation and forced oscillation suppression.Firstly, an overview of the reason, major monitoring and control methods about low frequency oscillation are proposed. Secondly, based on the linear state space model, modal superposition and virtual excitation method are introduced to analyze characteristics of mode and modal shape of free oscillation, forced oscillation and random oscillation, which lays the foundation of oscillations identification. Then, the stochastic subspace identification characteristics are summarized to verify the possibility of the whole process of power system oscillations identification, and stability diagram based on quantum clustering is proposed to decide the order of identification, and some cases are introduced to verify this method. Meanwhile, to reduce calculating time and adapt to the requisition of real-time online monitoring, a subspace tracking method with an adaptive scheme is proposed to improve the rapidity and robustness of the tracking. Finally, a method combing ADRC and multi-objective control is proposed to achieve forced oscillation suppression.The analysis of oscillations is in favor of understanding presentations of these oscillations and provides the theory foundation of the identification. The research of whole process of on-line monitoring based on stochastic subspace can be used to take advantage of PMU data to monitor the situation of low frequency oscillation in real time. The ADRC method is proposed to suppress the forced oscillation, which can reduce the impact on system stability limit and the risk of cutting machines and splitting measures caused by forced oscillation. This paper can be helpful to understand features of three typical low frequency oscillation, and makes attempt to perceive the situation of low frequency oscillation and suppress forced oscillations, so this thesis has some theoretical significance and project benefits.