An Adaptive Decomposition Scheme for Wideband Signals of Power Systems Based on the Modified Robust Regression Smoothing and Chebyshev-II IIR Filter Bank

Distributed energy resources and new types of electrical loads are morphologically changing the signals of power grids. Grasping the wideband information of field data will play an important role in helping smart grids to operate in a safe manner. A wideband signal model composed of deterministic and stochastic components is proposed to enhance the wideband measurement. Based on the model, an adaptive decomposition scheme is designed for the offline analysis of wideband signals. In the scheme, a modified robust local regression smoothing method is presented to denoise the wideband signal. A totally data-dependent threshold is established to extract adaptively interharmonic components. Further, the Taylor-Fourier transform algorithm is used to estimate the parameters of deterministic components based on the subsignals extracted by the Chebyshev-II infinite impulse response filter bank. Finally, the stochastic components are extracted using the adaptive threshold based on the signal that the deterministic components are removed. The performances of the decomposition scheme under different conditions have been validated. In the end, the proposed decomposition scheme is carried out and verified with the field data of the 24-h current signals at one residence community.