| Recently, with the rapid development of society and economy the scale of intercon-nected power grid becomes greater and greater. The dynamic property of entire interconnec-tion has been changed to be more complicated with more interconnection of power grid which also could improve the economy for power system operation. Besides, the power load rising also brings the significant challenge of electric power transmission. Under this back-ground of electric power industry, power system low frequency oscillations which is one of the hot research topics in the area of power system stability and control becomes more and more important. Power system low frequency oscillations can limit the backup of power in different areas and exchange of electric power. What is more, in severe cases it will result in the interconnection breakup, blackout and other accidents.Because of the importance of issue on power system low frequency oscillations, the in-dustry and academia has paid a lot of attention on it and a lot of different methods on ana-lyzing low frequency oscillations can be obtained. According to the varied data source for analyzing low frequency oscillations, methods on low frequency oscillations can be classi-fied into two types:the method based on system model and the method based on measure-ment. The first method relys on the accurate model of power system and determinate opera-tion mode which is rather difficult to acquire resulting from the complexity of power grid and uncertainty of operation mode in the case of large interconnected power grid. Actually, at the time of oscillations power system operation data contain rich oscillations information directly. Thus, the method for analyzing power system low frequency oscillations by utiliz-ing the field measurement data has been widely applied.Wide area measurement system (WAMS) can obtain real time operation parameters of power grid from synchronized sampling by phasor measurement unit (PMU) deployed in wide geography area of power grid. Then operation parameters are transmitted to data center via the high speed communication network and the global dynamic information of entire power grid can be obtained accordingly. Wide Area Frequency Monitoring Network (FNET) is a novel WAMS by using the distribution power outlet as its power source and data source. It have characteristics of low cost, easy installation and high accuracy. Nowadays, it has been widely deployed in the North American power grid.A lot of problems occurred when the field measurement data recorded by FNET was applied into analysis of low frequency oscillations. The problems introduced above are as follows:selection of proper data frame of input signal considering the effects of non-linearity of power system, selection of sampling frequency of input signal, time length setting of input signal, effects of varied signal-to-noise ratio (SNR) of multiple input signal on multiple signal modal analysis algorithm, presetting processing parameters of algorithms corresponding to different SNR of input signal. Based on those problems this paper obtains following significant achievements by theoretical research and innovation:improving Prony algortihm, introducing rules of presetting processing parameter of algorithms, presenting the guideline of algorithms application, acquiring the optimal range for parameter setting of in-put signal, appling time frequency technology for choosing the proper data frame of input signal, providing a filtering scheme in the case of multiple input signal with varied SNR, and a whole scheme including all the module above for analyzing the low frequency oscillations from field measurement data. The result of simulated case and practical case shows that the proposed whole scheme is obviously much better than existing methods. Therefore, the re-search of this paper has a certain theoretical value and practical significance. Finally, a sta-tistical analysis for low frequency oscillations from 2009 to 2010 occurred in American eastern interconnection is made by applying the proposed scheme. At the same time, random selected cases of eighteen are analyzed in detail and corresponding modal information in-cluding mode frequency, damping ratio and mode shape are also introduced.
|
PDF Full Text Request