Identification Of Low-frequency Oscillation Characteristics Of Power System Based On Prony Algorithm

Under the trend of power grid interconnection,the problem of low-frequency oscillation is becoming more and more serious.Low-frequency oscillation will not only reduce the transmission capacity of the system,but also affect the stable operation of the whole network.So it is necessary to study the low-frequency oscillation.In recent years,the atypical power oscillation of the power system has become a new problem due to the generator in the power system.It is of practical significance to study the power fluctuation characteristics of the generator.With the development and application of phasor measurement unit(PMU),it can acquire the multi-point dynamic information synchronously,which provides a more convenient foundation for the study of power system low-frequency oscillation and the generator power fluctuation characteristics.At present,study low-frequency oscillation are mostly based on the Prony algorithm,Prony algorithm can identify the characteristic parameters of the system and is simple in application,but it is sensitive to noise.Therefore in this thesis,the improved Prony algorithm as the starting point,study low-frequency oscillation and the generator power fluctuation characteristics.Firstly,this thesis discusses the power system oscillation model and the principle of traditional Prony algorithm.And then it improves on the basis of the traditional Prony algorithm,proposes the static average Prony algorithm.The analysis results prove that the static average Prony algorithm increases the anti-noise ability by adopting the method of summing and averaging.Secondly,based on the static average Prony algorithm,adding the moving window technology,the Prony moving average window algorithm is obtained.The relevant examples show that the algorithm can not only weaken the influence of noise but also make full use of the collected data,and also can obtain the time-varying characteristics of the low-frequency oscillation parameters,which provides more reliable support for the analysis of low-frequency oscillations.In addition,according to the different degrees of influence on the system in prophase,metaphase and anaphase of the low-frequency oscillation,this thesis combines the oscillation energy function and time-varying function,the time-varying oscillation energy function is obtained,and the analysis results verify that the time-varying oscillation energy function can effectively improve the dominant position of the dominant mode and can more accurately choose the dominant mode of the low-frequency oscillation in every period of time.Finally,through the above parameters identification method,the characteristics of the power fluctuation of the generator are obtained,then this thesis can make a comprehensive assessment of the volatility characteristics and sort the comprehensive scores,which can help staff to improve the poor generator to raise the overall safety of the system.This thesis verifies the effectiveness of the comprehensive assessment scheme based on 4-machine-2-area system and actual grid data.