【作者】 惠国华；

【导师】 陈裕泉；

【作者基本信息】 浙江大学 ， 生物医学工程， 2008， 博士

【摘要】 六氟化硫（Sulfur Hexafluoride,SF₆）自法国化学家Moissan和Lebeau于1900年首次合成以来,以其独特的物理、化学性质在工农业生产生活中得到了广泛应用。与此同时,SF₆气体泄漏也给人们的生产、生活带来了不容忽视的危害。因此准确、安全、方便、快速的检测SF₆气体泄漏具有极大的理论和实际应用价值。SF₆属于惰性绝缘气体,是已知化学稳定性最好的物质之一,传统的基于吸附-脱附、氧化-还原原理的气敏传感器对SF₆气体检测无能为力。气相色谱/热导池检测器（GC/TCD）法可以检测SF₆气体,但这种检测方法已经基本被更高精度的色谱检测方法取代。SF₆是不可燃烧的无机气体,一些新发展的色谱分析技术（如气相色谱/氢火焰检测器（GC/FID）等）也无法达到检测的目的。气相色谱-质谱联用（GC-MS）设备能够检测微量SF₆气体,但和气相色谱设备一样存在着价格昂贵（单套设备通常在几十万到一百万人民币左右）、体积庞大、单次检测时间长、须在专门实验室检测、运行成本高等不利因素。目前进口SF₆气体便携式检测仪,如DNS200、GasCheck5000is型,检测范围基本在1～1000ppm,但是高达27000～60000人民币的高昂价格让人望而却步。此外,红外光谱吸收法检测SF₆气体可以达到较高的精度,但是检测设备复杂、成本高、加工难度较大。紫外光电离法和强电压电离法检测SF₆气体,在检测过程中破坏SF₆气体分子产生了硫化氢、二氧化硫等有毒物质,对检测人员的人身安全带来极大的威胁。因此,新型的SF₆气体检测技术的研究已经被许多科研机构和电力部门提上了议事日程。结合传感器理论和纳米材料技术,我们设计了基于随机共振理论和定向多壁纳米碳管电离型气敏传感器阵列的SF₆气体检测系统。使用阳极氧化铝（AnodicAluminum Oxide,AAO）模板作为基底,以化学气相沉积（Chemical VaporDeposition,CVD）法在模板上快速生长定向多壁纳米碳管作为一体化电极,与铝电极构成电离型气敏传感器。论述了气敏传感器的工作原理,实验研究了该传感器的放电特性和影响因素。出于优化电场效应、增强纳米碳管气敏传感器气敏性的需要,本文实验研究了纳米碳管直径、间距和高度的可控制备方法,最终传感器采用了符合电场效应优化条件的纳米碳管材料。为了解决单个纳米碳管气敏传感器存在着的“交叉敏感性”的问题,本文的SF₆气体检测系统采用了传感器阵列的设计方法,阵列中的传感器具有不同极性或不同极板间距。使用4传感器阵列结合主成分分析（PCA）的方法,实现了对纯氩气、SF₆、氮气、二氧化碳的定性检测。随机共振理论在微弱信号检测领域正在成为研究的热点。本文中SF₆气体检测系统对ppm级SF₆气体的检测输出信号处于微安级,如何准确、快速的处理实验数据给出结果十分关键。本文首先对双稳态系统的随机共振模型进行了仿真研究,旨在研究双稳态系统的参数对随机共振产生的影响,为随机共振在SF₆检测系统中的应用奠定理论基础。检测系统采用2个纳米碳管气敏传感器构成的传感器阵列结合双稳态随机共振方法,以信噪比极大值标定SF₆气体浓度,实现了轻度（500～1000ppm）、中度（1000～2000ppm）、重度污染（高于2000ppm）三级浓度的监测。实验观察到外加白噪声诱发碳管气敏传感器系统内部1.65MHz周期性分量产生多重随机共振现象,本文采用多稳动力学系统随机共振的数值仿真研究,得到与实验结果相一致的随机共振曲线,为实验所观察到的多重随机共振提出了一种解释。该数值仿真方法使用三势阱折线分段势函数,邻接两个双稳态随机共振以构建多稳态随机共振模型。如果在实验中观察到更多重随机共振现象,该方法也可以发展到更多重的情况,具有良好的扩展性。本文在SF₆气体检测系统的设计基础之上,制备AA0模板,并以其为基底定向生长多壁纳米碳管制作传感器电极,采用主成分分析结合4传感器阵列实现氩气、SF₆、氮气、二氧化碳的定性检测;以双稳态随机共振结合2传感器阵列标定SF₆气体浓度,在中国国家和行业SF₆防护标准的框架下,以系统实时检测实验数据计算信噪比极大值,根据信噪比极大值判断空气中的SF₆气体污染情况,实现轻度、中度、重度三级污染监测。采用了多重动力学模型数值仿真研究了气体放电中的多重随机共振现象。基于随机共振和定向多壁纳米碳管电离型传感器阵列的SF₆气体检测系统具有一定灵敏度、稳定性好、响应速度快、可重复检测、体积小、成本低、常温常压下即可进行检测等优点,具有较好的实际应用价值。本文对优化气敏传感器场效应的纳米碳管可控制备方法进行了论述和实验研究,以多稳态动力学模型数值仿真研究了气体放电中的多重随机共振现象,设计使用了纳米碳管气敏传感器阵列检测SF₆气体,以双稳态随机共振方法进行信号处理,实现SF₆气体的定量检测。更多还原

【Abstract】 SF₆ （Sulfur Hexafluoride） has been widely used in industry, agriculture and daily life due to its unique physical and chemical characteristics since it was firstly prepared by French chemists Moissan and Lebeau in 1900. At the same time, the leakage of SF₆ does great harm to people’s work condition and life security. Thus how to detect SF₆ with accuracy, convenience, and efficiency has been a serious subject from theoretical and practical aspects.SF₆ is an inactive insulated gas and it presents excellent chemical stability. Gas sensors in some mechanisms （such as adsorption-desorption, oxidation-deoxidation, etc）, could not detect SF₆. GC/TCD method presents high accuracy in SF₆ detection, but this method has been almost replaced by other new method, such as GC/FID. GC-MS method also could detect microdosage SF₆ but its equipments have bulky architectures, expensive price and long detecting period, which limits the field usages of GC/MS equipments. Foreign portable SF₆ detectors, such as GasCheck5000is and DNS200, own 1～1000ppm detection range. But their prices （from 27000 to 60000 RMB） are also expensive. Other sensing methods, such as infra-red spectrum absorption method could analyze SF₆ gas in satisfying accuracy, but the detecting system is complicated, high-cost and hard to machine. UV-light ionization method and high voltage ionization method own SF₆ detecting capability, but some hazardous chemical compounds, such as H₂S and SO₂, which could do great harm to the people who are operating the equipments. So a novel SF₆ gas detecting method presents great value.Combines with modern sensing theory and developing nano-material techniques, we proposed a novel SF₆ detecting system based stochastic resonance theory and aligned multi-walled carbon nanotubes ionization gas sensor array. We utilized Anodic Aluminum Oxide （AAO） template as substrate, and Chemical Vapor Deposition （CVD） method to fast grow aligned multi-walled carbon nanotubes on substrate to fabricate substrate-carbon nanotubes integrated electrode. We used the carbon nanotubes electrode and Al electrode to assemble ionization gas sensor. Meanwhile, we explored discharge characteristics and influence factors （interelectrode distance, temperature, atmospheric pressure, relative humidity, etc） through experiments. To optimize the characteristics of ionization sensor, we studied the diameter, distance and height of carbon nanotubes when field effects of carbon nanotubes were in optimized conditions. Finally, we concluded the correlated treatments during the preparation procedure of carbon nanotubes.In order to solve the ’cross-sensitivity’ problem lies in single carbon nanotubes gas sensor, we designed a carbon nanotubes gas sensor array in our SF₆ gas detecting system. Utilizing Principle Component Analysis （PCA） method and a four-sensor array, we could detect pure Ar, SF₆, N₂, and CO₂. Stochastic resonance theory has been widely studied in the last twenty years. When we detected SF₆ /Air mixtures in ppm level, the output signals were in microampere level. The key lied in how to determine SF₆ concentration according to output signals accurately and fast. Based on the research of bistable stochastic resonance, we proposed a two-sensor array combined with bistable stochastic resonance method to determine SF₆ concentration successfully. Furthermore, we set up a three-level SF₆ concentration alarm: low level （500-1000ppm）, medium level （1000～2000ppm）, and high level （over 2000ppm）.The stochastic multi-resonance phenomenon was observed within the gas sensor system. A periodical signal （1.65MHz） presented stochastic multi-resonance induced by adding white noise to the sensor system. Here we adopted a multi-stable dynamic system to simulate the stochastic multi-resonance phenomenon observed in the sensor system. The simulation results agreed with experimental results well.Based on design of SF₆ gas detecting system, we prepared AAO template, and grew aligned multi-walled carbon nanotubes on the template to assemble ionization gas sensor. PCA and a four-sensor array were used to classify pure Ar, SF₆, N₂, and CO₂. Bistable stochastic resonance and a two-sensor array were utilized to detect microdosage SF₆ gas in low concentration. A multi-stable dynamic system was used to study the stochastic multi-resonance phenomenon of internal periodical signal within the sensor system. The SF₆ detecting system owns the characteristics of small size, high sensitivity, high stability, quick response, and low cost. It presents good theoretical and practical value.更多还原