Research On State Estimation Of Distribution Network Based On Volumetric Particle Filter

The reliable and stable operation of distribution network is the basis of safe and stable operation of power system.However,the monitoring and control of distribution network becomes difficult due to the access of the DG.In addition,there are errors in the original measurement data.If the raw data are used directly to analyze the operation state of the distribution network,it will cause erroneous judgment,and affect the safe and stable operation of the distribution network.State estimation which is the core of DMS is the premise and foundation of safe and stable operation of distribution network.Therefore,it is necessary to use state estimation to process the raw data and improve the accuracy of the data to obtain more accurate operation state of distribution network.This paper mainly studies from the following aspects:1)The basic state estimation algorithm-KF is introduced,and the three-phase model of distribution network is established according to the characteristics of three-phase asymmetry of distribution network.On this basis,the mathematical model of distribution network dynamic state estimation is given,and the heavy-tailed distribution is used to represent the measurement error obeying non-Gaussian distribution.2)The filtering method based on the Kalman framework only has the best estimation results for the Gaussian system,while the PF algorithm can be applied to any system.However,PF is so excessively dependent on the system model that the accuracy of state estimation is insufficient.The CPF algorithm is introduced,which uses the estimation results of CKF as the recommended distribution of particle filter,and improves the state estimation performance of particle filter.Therefore,the CPF algorithm is applied to the state estimation of distribution network.The simulation results in the IEEE33-bus system show that the estimation accuracy of CPF algorithm is the highest compared with CKF and PF algorithms under Gaussian noise or non-Gaussian noise.3)In view of some abnormal phenomena in distribution network,the cubature particle algorithm is improved.Such as bad measurement data in the system,the RCPF algorithm was created by introducing a time-varying,multi-dimensional scale factor into the CPF algorithm and is proposed.The value of Kalman filter gain is adjusted online by the scale factor,and then enhance the robustness of the RCPF algorithm to bad data.Such as the sudden change of system load,an ACPF algorithm was created by introducing an adaptive factor into the CPF algorithm and is proposed.Finally,the effectiveness of the proposed improved CPF algorithm is verified by simulation analysis in IEEE33 bus system.