Abstract：In the multi-source measurement system of distribution network，the sampling frequency and time

stamp of multi-source measurement equipment are not synchronized，as well as the bad data in the system，which

will lead to the bias among the measurement data，thus affecting the accuracy of state estimation. To this end，a

dynamic state estimation method based on adaptive extended Kalman filtering （EKF） under multi-source

measurement data was proposed. Firstly，to address the non-synchronization problem of multi-source measurement data，a multi-source data timestamp alignment strategy based on dynamic time warping（DTW）was proposed to realize the synchronization of measurement data. Secondly，for the bad data in the system，an EKF state estimation method integrating the bad data adaptive detection and filtering link was proposed to overcome the effect of bad data on state estimation. Finally，an arithmetic test was performed in an IEEE 33 node system and compared with a conventional EKF method that did not consider the fusion of multi-source metrology data and outlier detection. The results show that the proposed method improves the robustness and reliability of the estimation results.

      Key words：distribution network；state estimation；multi-source data fusion；bad data detection；Kalman

filtering（KF）