Method and system for identifying and modeling multi-time scale characteristics of power grid frequency fluctuation

Abstract

The invention discloses a power grid frequency fluctuation multi-time scale feature identification and modeling method and system. The method comprises the steps of obtaining a frequency fluctuation sequence; performing detrending fluctuation analysis on the frequency fluctuation sequence to obtain a fluctuation function; acquiring a local maximum value of the fluctuation function, and dividing second-level, minute-level and hour-level time scale intervals; in different time scale intervals, a second-order autoregression model is adopted to carry out modeling on frequency fluctuation; all the second-order autoregression models are superposed, a power grid frequency average value in a set period is added to obtain a power grid frequency preliminary comprehensive model, and the preliminary comprehensive model is corrected based on historical accumulated deviation in a set long-time scale; the characteristic equations of all the second-order autoregression models are solved, different characteristic parameters are selected for calculation or acquisition according to the types of the characteristic roots, and the different characteristic parameters correspond to different power grid adjusting mechanisms. According to the method, accurate quantification of each fluctuation source is realized, and specific key parameters such as relaxation time, oscillation period and noise variance are identified.

Images