A self-tuning optimised unscented Kalman filter for voltage flicker and harmonic estimation

摘要

This paper presents a novel technique for estimation of two important power quality problems like voltage flicker and harmonics in power networks using an unscented Kalman filter (UKF) algorithm. The fluctuating voltage of a power network resulting in voltage flicker is tracked for estimating its envelope, voltage magnitude, and frequency. Further, the unscented filter is used to estimate harmonics in power network and in a special case harmonic voltage in a voltage-source converter-based HVDC (VSC-HVDC) system to be used for power quality improvement in distributed generation system. The UKF is found to be superior to the conventional extended Kalman filter (EKF) as it overcomes the difficulties in linearisation and derivative calculations used in the latter for computing the Kalman gain. As the design of noise covariances in the signal and measurement models play important roles in stabilising the performance of the filter, a particle optimisation technique is used to obtain initial optimal values of the filter parameters. Once the filter is initialised, a self-tuning procedure is adopted here to vary these iteratively for improved tracking performance.