Laser-induced breakdown spectroscopy; SF6 oxygen concentration; Wavelet transform; Partial least square; Quantitative analysis;DISCHARGE

Sulfur hexafluoride (SF6) is widely used in gas-insulated high-voltage equipments due to its excellent insulation and arc extinguishing performance. However, when impurities such as trace water and oxygen are present in SF6, the by-products, which are formed due to equipment faults and partial discharge, can react with these impuriti to form stable sulfur oxyfluorides. The equipment insulation performance can be degraded significantly by those stable sulfur oxyfluorides, causing threat to the safety of the power system. The detection and analysis of the impurities such as water, oxygen and decompositions in SF6 are therefore of great importance. In this paper, laser-induced breakdown spectroscopy is employed to measure trace O in SF6. CaF2 is used as window material to solve the degradation of the excitation energy caused by the corrosion of the window material by SF6 decompositions, to eliminate the pollution problem caused by the reaction between the window material and the breakdown products of SF6 gas, and to reduce the change of plasma state caused by the change of excitation condition. By correcting the spectral baseline with an iterative wavelet transform and suppressing the noise with a wavelet transform with soft thresholding, a limit of detection of 38ppm is achieved from the experimental calibration curve. Furthermore, a quantitative model based on partial least squares (PLS) is developed to achieve better stability and precision.

laser-induced breakdown spectroscopy; time-resolved; plasma temperature; electron density;INDUCED BREAKDOWN SPECTROSCOPY; AIR PLASMAS; GAS; DISCHARGES; MIXTURES

As an important analytical tool, laser-induced breakdown spectroscopy (LIBS) has been widely used in material analysis, environmental monitoring, and other fields. In recent years, due to increasingly serious air pollution, various LIBS-based on-line air pollution detection techniques are being developed. The temporal evolution of nitrogen plasma characteristics is of great importance for investigating the atmospheric plasma dynamics and developing the LIBSbased air pollution monitoring techniques. Temperature and electron density, which are the most important parameters of a plasma state, directly influence the kinetic behaviors of plasma formation, expansion and degradation processes, as well as the energy transfer efficiency in plasma. In this paper, the temporal evolutions of continuous background radiation, molecular spectral strength, and signal-to-background ratio (SBR) are studied based on time-resolved spectra. The results show that the lifetime of the continuous background radiation is about 700 ns, the N-2(+) (B-2 Sigma(+)(u) -X-2 Sigma(+)(g), v: 0-0) transition line strength reaches a maximum value within 12-15 mu s, the SBR first increases and then stabilizes. Accordingly, the optimal observation period for N-2(+) (B-2 Sigma(+)(u) -X-2 Sigma(+)(g)) band system based plasma temperature investigation should be selected to be between 10 and 25 mu s. The temporal evolution of plasma temperature is determined by fitting experimental spectra to theoretical ones simulated by LIFBASE (a spectral simulation program). As the radiation loss is less than the loss due to the collision cooling, the plasma temperature decays exponentially from similar to 10000 K to similar to 6000 K within 10 28 mu s. By taking into account the instrumental broadening lineshape (Voigt lineshape), the temporal evolutions of Stark broadening and Stark shift of N 746.831 nm atomic line are obtained via Nelder-Mead simplex algorithm, and then the electron density is calculated accordingly. The results show that the electron density decays between 10(17) and 10(16) cm(-3) in magnitude. By comparing the experimental electron decay rate with theoretical values calculated from different mechanisms, it is concluded that a three-body collision recombination is the main mechanism of electron decay.