Comparison of adaptive filter performance in estimating the noisestatistics for harmonic models

摘要

Electric power system harmonics and interharmonics having randomvariations in magnitude are modeled as state variables of adiscrete-time stochastic system. The problem of adaptively estimatingthe statistical characteristics of the process and measurement noises isaddressed. Incorporating the unknown noise covariance in the systemleads to a state-multiplicative-noise model resulting in a nonlinearestimation problem. The main contribution of the paper is theperformance comparison, by Monte Carlo simulation, between the linearminimum-variance filter and the extended Kalman filter. Conclusionsbased on these simulation results are presented