[Baseline correction method for spectrum signal of SF6 insulating air with optimum wavelet basis]

SF6 gas has been widely used in the power equipments as an excellent electric insulating and arc-quenching medium. In the present paper, a baseline correction method based on the optimum wavelet basis for spectrum detection is proposed to measure the composition content of the SF6 insulating gas to secure the power safety. In this method, the optimum wavelet basis is selected in the wavelet packet according to constructor function on the energy concentration criterion to express the spectrum signal in the time-frequency domain. Then the strong spectrum composition is removed from the spectrum signal with the threshold method to eliminate the interference with the continuous spectrum fitting. Finally we remove the continuous spectrum which is fitting result from the origin spectrum and obtain the useful signal of line spectrum. The intensities of spectral line processed with the proposed algorithm could reflect the concentration of the conponents to be measured in SF6 gas. Experiments to analyze the absorption spectrum of the SF6 insulating gas mixture show that the proposed algorithm can estimate and correct the drifting baseline accurately, and its performance is better than the algorithm based on iterative wavelet.