Method for estimating dynamic states of power generators on basis of unscented particle filtering theories

Abstract

The invention discloses a method for estimating dynamic states of power generators on the basis of unscented particle filtering theories. The method includes utilizing fourth-order dynamic equations of the power generators as state equations for estimating the dynamic states of the power generators, simulating PMU (phasor measurement units) by the aid of power system analysis software to acquire measurement data of power angles, angular speeds and the like of the power generators, and creating measurement equations for the power generators; acquiring static estimation values at state estimation initial moments, utilizing the static estimation values as initial values for the power generators at dynamic state start moments, generating initial particles adjacent to the initial values, carrying out tracking filtering on state variables of the power angles, the angular speeds and the like of the power generators by the aid of unscented particle filtering algorithms to ultimately obtain estimation values of the state variables of the power generators. The method has the advantages that the quantity demands on the particles can be lowered, and the filtering accuracy and the computational efficiency of the method are superior to the filtering accuracy and the computational efficiency of the traditional particle filtering processes; the dispersibility of the particles is improved by the aid of the method, and accordingly the robustness of the method is superior to the robustness of the traditional particle filtering processes and unscented Kalman filtering processes.

Description

technical field [0001] The invention relates to a generator dynamic state estimation method based on the unscented particle filter theory, which belongs to the technical field of power system monitoring, analysis and control. Background technique [0002] With the increase of the grid scale and the complexity of the power system, and due to the measurement errors of the existing measurement equipment, it is difficult to obtain the real state information of the power system by means of direct measurement. The real state of the power system is an important reference for the analysis, control and decision-making of the power system, and the state estimation can filter out the error of the measurement data of the power system, and obtain an approximate value as close as possible to the real state of the system. [0003] The traditional state estimation often adopts the static state estimation method represented by the least squares algorithm to iteratively solve to obtain the st...

AI Technical Summary

Problems solved by technology

However, UKF is sensitive to the initial value, and there is no definite selection principle for parameters. Although CKF does not need to select parameters, the filtering improvement is limited

PF filtering has high precision, but requires a large number of particles for calculation, and the calculation efficiency is low, and when the overlap between the prediction prior and the likelihood function is small or the accuracy of the measurement model is high, the filtering may fail.

Images