基于纳米修饰石墨烯气敏传感器的油浸变压器故障特征气体检测研究

The detection and analysis of dissolved gases in oil is one of the best known and most effective methods for the running state assessment and incipient faults diagnosis of oil-immersed power transformer. The fabrication of highly sensitive and selective sensor, associating with the inhibition of cross-sensitive reponse of sensor to multi-gas, not only are two bottlenecks in the detection of dissolved gases in transformer oil, but also are the key common scientific issues in rapid micro-determination of multi-component gases. Aiming at developing high-sensitive graphene-based gas sensors and their application, this work will further investigate the in-situ nano-modified graphene by means of theatrical simulation and experimental exploration, and fabricate the graphene-based film sensors using MEMS and self-assembly technologies, and then identify the graphene-based gas sensors which are sensitive to dissolved characteristic gases such as acetylene, methane, ethane, polytene, hydrogen and carbon monoxide. We will explore the gas sensing-enhancement mechanism of nano-modified graphene and gas response characteristics. This work will use the blind source separation (BSS) method to eliminate the cross-sensitive and drift reponse of sensors, along with the least square-support vector regression (LS-SVR) method to realize the high-selective identifition of gas component and high-precision quantitative detection. The result of this work may provide a new method for the running state assessment and incipient faults diagnosis of power transformer.

变压器油中溶解气体的检测与分析是当前油浸式变压器状态监测和早期潜伏故障诊断的重要手段。研制高灵敏度高选择性的传感器与抑制不同气体的交叉敏感性是变压器油中溶解气体检测的瓶颈，也是多组分气体微量快速检测中的关键共性科学问题。本项目以高灵敏度石墨烯气敏检测技术及其应用研究为目标，采用理论模拟与实验研究相结合的方法对原位纳米修饰石墨烯进一步深入研究，采用MEMS结合自组装技术制作纳米修饰石墨烯薄膜气敏器件，筛选对变压器油中溶解的乙炔、甲烷、乙烷、乙烯、氢气、一氧化碳等故障特征气体具有高灵敏度和选择性的石墨烯气敏传感器，揭示石墨烯气敏传感器的修饰增敏机理与气体敏感特性。研究盲源分离方法抑制传感器交叉敏感和漂移特性，协同采用最小二乘支持向量回归模型实现混合气体的高选择性识别和高精度定量检测，为电力变压器运行状态监测和早期故障诊断提供一种新方法。

变压器油中溶解气体的检测与分析是当前油浸式变压器状态监测和早期潜伏故障诊断的重要手段。研制高灵敏度高选择性的传感器与抑制不同气体的交叉敏感性是变压器油中溶解气体检测的瓶颈问题。本项目以高灵敏度石墨烯气敏检测技术及其应用研究为目标，采用理论模拟与实验研究相结合的方法对原位纳米修饰石墨烯开展了深入研究，采用MEMS结合自组装技术制作了纳米修饰石墨烯薄膜气敏器件，对变压器油中溶解故障特征气体开展气敏特性实验，对具有高灵敏度和选择性的石墨烯气敏传感器进行了筛选，揭示了石墨烯气敏传感器的修饰增敏机理与气体敏感特性。通过选择特异性强的石墨烯传感器单元制作阵列式传感器，提出了有效的智能数据挖掘方法进一步抑制了传感器交叉敏感和漂移特性，实现变压器故障气体的高选择性识别和高精度定量检测。在此基础上，建立了基于变压器油中溶解气体分析的绝缘故障诊断模型，取得较好的故障诊断正确率。本项目以“新材料-新器件-新算法”为手段，研究了金属氧化物修饰石墨烯薄膜传感器的制备及其气体敏感特性，并融合智能算法搭建预测和诊断模型，为纳米传感技术融合智能算法在变压器故障诊断领域提供了新应用，为电力变压器运行状态监测和早期故障诊断提供一种新方法。在本项目资助下，发表SCI论文三十余篇，申请专利6项，荣获山东省高等学校优秀科研成果奖一等奖，入选全国高校石油矿业与安全工程领域优秀青年科技人才奖。