Optical Sensing System Based on Bidirectional Self-Correction Technology: An Online Detection Method for Carbon...

Carbon disulfide (CS₂) and sulfur dioxide (SO₂) are typical indicative gases for the early warning and diagnosis of faults in gas-insulated switchgear. In this study, an optical sensing system is reported for online detection of CS₂ and SO₂ based on a bidirectional self-correction technology (BSCT). First, the differential absorption signals of CS₂ and SO₂ in the wavelength range of 195-230 nm are obtained using the UV differential optical absorption spectroscopy (UV-DOAS) technique. On this basis, a BSCT is proposed to decouple the spectral lines that exhibit significant overlap. This method primarily employs bidirectional difference spectroscopy to mutually correct the spectra of CS₂ and SO₂, combined with spectral reconstruction to extract the single-component absorption signals of CS₂ and SO₂ from the mixed gas spectra. Furthermore, the effectiveness of this novel decoupling technique is validated by comparing the decoupling results with the absorption spectra of single-component standard gases at corresponding concentrations. Finally, the quantitative relationships between the concentrations of target gases (CS₂, SO₂) and the optical parameters are modeled using the least-squares method. The experimental results show that the mean absolute percentage errors of CS₂ (19.00-3735.35 ppb) and SO₂ (0.19-38.77 ppm) are 0.543 and 0.521%, respectively. At an effective optical range of 50 cm, the system achieves the lowest detection limits of 0.5 ppb for CS₂ and 12 ppb for SO₂, representing the best results reported to date for the online detection of CS₂ and SO₂ in the ppb-ppm range.