Forced Oscillation Source Location in Power System Using Wavelet Dissipation Energy Spectrum

Fast and accurate location of the forced oscillation source is the key to eliminate the forced oscillation source. At present, most of the methods for locating forced oscillation sources in power systems adopt the theory of dissipating energy flow (DEF) in time domain via wide-area measurement information. However, when the traditional time-domain DEF is used to determine the oscillation source, the required wide-area measurement information needs to be converted into "time-domain, frequent-domain, time-domain" to extract the time-domain component of forced oscillation in the wide-area measurement information, so as to improve the accuracy of the oscillation source location. The calculation process is relatively complex and the computational efficiency is low. solve this problem, this paper proposes a method of locating the forced oscillation source in the frequency domain based on the wavelet dissipation energy spectrum (WDES), and realizes the forced oscillation source localization in the frequency domain to improve the computational efficiency of the oscillation source localization. Firstly, the wavelet coefficient matrix of each wide area measurement information is obtained by using continuous wavelet transform (CWT). On this basis, the CWT multiplication formula is combined with the traditional time-domain DEF formula to construct the WDES, and the relationship between the WDES and time-domain DEF is proved. Then, the trend of the WDES was analyzed, and the forced oscillation frequency of the system was determined according to its jump characteristics. Further, the WDES of each generator at the forced oscillation frequency is extracted to locate the forced oscillation source accurately. Finally, the proposed method is evaluated by WECC 179-bus test system and ISO New England system. The simulation results show that the proposed method can accurately and effectively locate the single oscillation source, double oscillation source and real oscillation source in WECC 179-bus test system and ISO New England system, respectively. At the same time, compared with the traditional time-domain DEF method and time-frequency DEF method, the computational efficiency of the proposed method is improved by 32.78% and 31.39%, respectively, in the single oscillation source location of WECC 179-bus test system in scenario 1. In the single oscillation source location of WECC 179-bus test system in scenario 2, the computational efficiency of the proposed method is improved by 32.17% and 31.47% compared with the traditional time-domain DEF method and time-frequency DEF method, respectively. In the ISO New England system, compared with the traditional time-domain DEF method and time-frequency DEF method, the computational efficiency of the proposed oscillation source location method is improved by 30.10% and 24.16%, respectively. Obviously, the proposed method has higher computational efficiency of the forced oscillation source. The following conclusions can be drawn from the simulation analysis: (1) the proposed method based on WDES can accurately and effectively locate the forced oscillation source from the wide-area measurement information of the power system. (2) Compared with the localization method of DEF in time-domain and time-frequency domain, the proposed localization method of forced oscillation source in frequency domain based on WDES effectively improves the computational efficiency of forced oscillation source. (3) The proposed WDES is equivalent to the time-domain DEF and the time-frequency DEF. It can be considered as a form of time-domain DEF and time-frequency DEF in frequency domain. From the perspective of frequency domain energy spectrum, it provides a new solution for locating power system forced oscillation source through the field measurements. © 2023 Chinese Machine Press. All rights reserved.