Gas-Sensing Mechanism Research Of InN In SF₆-N₂ Decomposition Characteristic Components Detection Under Partial Discharge

SF₆ is a typical greenhouse gas.In order to reduce the impact on the environment and solve the problem of easy liquefaction in alpine regions,the SF₆-N₂ mixture as gas insulation medium for gas insulated switchgear（GIS）has become a technical solution and has practical engineering applications.Indium nitride（In N）has unique gas sensing properties and has great application potential in the above-mentioned GIS internal partial discharge detection,but there are still many problems to be studied.In this thesis,the structure and electronic properties of pristine In N monolayer were studied.The conductivity and gas-sensing properties were improved by cobalt atom/silver oxide molecular modification,and the potential of In N in the gas-sensing field was explored.Following are the main study contents:Based on gas discharge theory,electrostatic potential theory and molecular orbital theory,the molecular descriptors related to gas insulation strength were screened and calculated.The structural parameters of SF₆,N₂,NO₂,SOF₄,SOF₂ and SO₂F₂molecules were extracted,and the relative dielectric strength of each gas molecule was calculated by quantitative structure property relationship（QSPR）model.The results indicate that the SF₆-N₂ mixture gas has good insulation performance and has significant advantages as a substitute gas for SF₆.Based on density functional theory（DFT）,the electronic properties of pristine In N was analyzed.Results show that the pristine In N monolayer has good gas sensitivity to NO₂ molecule,but the selectivity and sensitivity to other SF₆-N₂decomposition components are limited.On the other hand,the negative electrostatic potential region on In N surface is mainly distributed around N atoms,which means that the metal elements tend to be doped above N atoms.Modify In N monolayer with Co atom and Ag₂O molecule.The gas sensing properties of the pristine and modified In N monolayer to NO₂,SO₂F₂,SOF₂ and SOF₄were systematically investigated.On the one hand,HOMO and LUMO are mainly distributed near the dopant,providing an active region for the adsorption of gas molecules;the element orbitals of the dopant tend to increase the density of states near the Fermi level,which is conducive to the transition of electrons,resulting in an increase in the conductivity.On the other hand,considering the short adsorption distance,the large charge transfer and adsorption energy of gas molecules on Co-In N surface,it is considered that Co atom modification can improve the adsorption performance of In N to NO₂,SOF₄,SOF₂ and SO₂F₂,making Co-In N a strong candidate for preparing adsorbents for SF₆-N₂ decomposition components.In addition,the conductivity of Ag₂O-In N system varies with the adsorption of NO₂,SOF₄,SOF₂ and SO₂F₂,which is beneficial to the effective detection of these gas molecules.It is worth noting that the short desorption time of SOF₂ and SO₂F₂ on Ag₂O-In N surface makes Ag₂O-In N has excellent recovery performance.Finally,the trainglm function is selected to optimize and train the BP neural network,and the GIS partial discharge severity evaluation based on SF₆-N₂decomposition components is preliminarily realized.According to the gas concentration of SF₆-N₂ decomposition components and the severity level of partial discharge,a sample library was established.The model of BP neural network was optimized by 500 groups of samples,and the transfer function and hidden layer nodes were determined.The optimized model successfully predicted the partial discharge severity of 20 groups of samples.In this thesis,the structure and electronic properties of In N monolayer are studied.By introducing cobalt（Co）atom and silver oxide（Ag₂O）molecule,the conductivity and gas-sensing properties of In N are improved,which expands its potential in the gas-sensing field.