The phase-selected closing accuracy of AC filters in high-voltage direct current(HVDC)transmission systems is influenced by variations in operating conditions

measurement deviations

control delays

and switching dispersion.The resulting switching phase error(SPE)poses potential risks to equipment safety and system stability.To address this issue

a systematic solution is proposed.First

a multidimensional dynamic accuracy evaluation framework and a comprehensive error propagation model are established to quantitatively analyze SPE.Then

an online error identification and compensation algorithm is developed

in which an improved adaptive extended Kalman filter(EKF)is employed to fuse multi-source information in real time for identifying error components.An adaptive feedforward strategy combined with EKF-based prediction is further used to dynamically correct closing commands.Rigorous simulation and hardware-in-the-loop(HIL)verification demonstrate that the proposed algorithm significantly enhances closing accuracy

reducing the root mean square value of SPE

while exhibiting strong robustness under complex operating conditions such as voltage fluctuations and harmonic disturbances.This study provides effective theoretical methods and technical support for improving the reliability of HVDC projects.