In order to solve the problem of filtering effectiveness degradation caused by inaccurate predictive values with too large errors of historical data and filtered historical values

this paper proposes a data adjustment method for distribution internet of things（D-IoT） based on AF/UKF being combined filter.In the method

adaptive filter（AF） is introduced into the prediction part of unscented Kalman filter（UKF） and a judgment based on statistical laws is added

thus

the filtering system has a certain ability to judge the transient or fault time. Simulation verification is performed by taking an actual 85-node system of actual power grid as an example through steady-state SCADA data acquisition system simulation and transient μPMU system simulation. Experimental results show that the proposed method has higher estimation accuracy than traditional unscented Kalman filter. After treatment of AF/UKF being combined filter

the average and maximum absolute error of the estimated values are smaller

and the system has better tracking performance when the system is in drastic changes.