Research on Intelligent Inspection Technology for Gas Leakage

Due to the large number and vast spatial distribution of such equipments, manual inspection of gas leakage defects has poor efficiency. An automated inspection system is developed to quench this conflict. First, a Fourier transform spectrometer is integrated to patrol robot, so infrared spec-tra which contains gas leakage signal could be captured remotely while the robot cruise in inspection area. Second, a spectra processing method was established, which uses brightness temperature spectra to eliminate background interfer-ence, then extract gas spectral feature and finish recognition. Such data processing method give this system capability to operate in complex outdoor environments to give alarm signal when gas leakage is detected. Validation experiments sug-gests that the sensitivity of this system is $0.5\mu(\mathrm{L}/\mathrm{s})$ (rate of leakage) at a distance of 25m. The spatial distribution of gas leakage could be mapped to give exact leakage point by using wide field visualization and data assimilation techniques. This patrol system for gas leakage inspection was developed for the first time in electrical industry, and has successfully given early warning to a gas insulated electrical instrument leakage defect.