【作者】 李杰；

【导师】 周利军；

【作者基本信息】 西南交通大学 ， 电气工程， 2024， 硕士

【摘要】 提高铁路运输的安全性和可靠性是当前铁路领域技术与创新发展的工作重心之一。车顶绝缘子作为电气化铁路列车顶部高压系统的支撑与绝缘部件,是保障列车电力系统绝缘安全的核心设备,其服役性能决定了电力机车能否可靠运行。由于列车顶部空间有限,车顶绝缘子的结构高度和爬电距离较小,是电力机车外绝缘安全结构中的薄弱环节。在列车运行过程中,车顶绝缘子表面不可避免地出现积污,在大雾或小雨等湿润条件下污秽层将受潮电解,进而导致车顶绝缘子的绝缘水平下降,严重时引发污闪事故。目前铁路系统主要采取定期清洁绝缘子的措施来预防污闪事故,这种按计划的检修方式不仅耗时耗力,也无法及时有效地掌握车顶绝缘子的运行状态,存在较大的行车安全风险。因此,开展电力机车车顶绝缘子污秽状态的在线监测技术研究,对于保障列车安全可靠运行具有重要意义。鉴于上述原因及意义,本文以车顶绝缘子为研究对象,依据相关试验标准搭建了绝缘子人工污秽模拟与泄漏电流测试平台,设计试验方案研究了车顶绝缘子污秽状态与其泄漏电流的映射关系,基于试验数据基础,提出了一种改进的组合算法模型用于绝缘子污秽状态的回归预测,并研发了一套车顶绝缘子在线监测装置及系统,将上述理论和算法研究成果落实到工程应用。主要的研究工作如下:首先,设计搭建了绝缘子人工污秽模拟与泄漏电流测试平台,建立了车顶绝缘子“污秽程度-湿润程度-电压大小-泄漏电流”的试验数据库,量化分析了绝缘子泄漏电流、湿润程度及电压大小三个影响因素对污秽程度的映射关系。其次,提出了一种改进哈里斯鹰寻优算法优化BP神经网络模型(IHHO-BP)用于车顶绝缘子污秽程度的回归预测,将车顶绝缘子的泄漏电流试验数据库作为模型数据集开展了训练和测试,并与主流的优化算法进行精度对比与分析,验证本文组合模型的有效性和准确性。最后,开展了车顶绝缘子泄漏电流监测技术的实用化设计,研发了包括信号采集、无线传输、电源管理等功能模块和抗干扰及封装设计的车顶绝缘子在线监测硬件装置,并开发了包括泄漏电流实时显示、绝缘子污秽状态评估、清洗指导和周期记录等功能的车顶绝缘子在线监测上位机系统,实现了车顶绝缘子运行状态的地面或车内无线监测。基于实车开展了硬件装置与软件系统的测试及验证,结果表明本文研发的车顶绝缘子在线监测装置及系统的泄漏电流采集与污秽状态评估等功能能够在实车环境下正常运行,且测试数据与实验数据的误差均在5%以内,为本文理论、算法及装置的研究成果应用于实际工程提供了有效支撑。更多还原

【Abstract】 Enhancing the safety and reliability of railway transportation stands as a focal point in the current technological and innovative developments within the railway sector.The roof insulator,serving as a support and insulating component for the high-voltage system atop electrified railway trains,constitutes core equipment ensuring the insulation safety of the train’s electrical system.Its operational performance directly influences the reliability of electric locomotive operation.Due to the limited space atop the train roof,the structural height and creepage distance of the roof insulator are relatively small,rendering it a vulnerable component within the external insulation safety structure of electric locomotives.During train operation,fouling inevitably accumulates on the surface of the roof insulator.Under humid conditions such as heavy fog or light rain,the fouling layer absorbs moisture and undergoes electrolysis,leading to a decrease in the insulation level of the roof insulator.In severe cases,this deterioration can result in flashover accidents.Presently,railway systems primarily employ periodic cleaning measures for insulators to prevent flashover accidents.However,this scheduled maintenance approach is not only time-consuming and labor-intensive but also fails to promptly and effectively assess the operational status of the roof insulator,posing significant risks to operational safety.Therefore,conducting research on online monitoring technology for the fouling status of roof insulators on electric locomotives holds paramount significance in ensuring the safe and reliable operation of trains.Given the aforementioned reasons and significance,this paper focuses on roof insulators as the research object.Based on relevant test standards,an artificial pollution simulation and leakage current testing platform for insulators was constructed.An experimental plan was designed to study the mapping relationship between the pollution state of roof insulators and their leakage current.Building upon experimental data,an improved combination algorithm model for predicting the pollution state of insulators was proposed.Additionally,an online monitoring device and system for roof insulators were developed,translating the theoretical and algorithmic research results into practical engineering applications.The main research work is outlined as follows:Firstly,an artificial pollution simulation and leakage current testing platform for insulators was designed and constructed.An experimental database for roof insulators,involving the mapping relationship between"pollution degree-wetness degree-voltage-leakage current,"was established.The study quantitatively analyzed the mapping relationship of three influencing factors,namely leakage current,wetness degree,and voltage,on the degree of pollution.Secondly,an improved Harris hawk optimization algorithm was proposed to optimize the BP neural network model(IHHO-BP)for regression prediction of the pollution degree of roof insulators.The roof insulator leakage current experimental database was used as the model’s dataset for training and testing.The reliability and accuracy of the proposed model were validated through precision comparisons and analyses with mainstream optimization algorithms.Finally,practical design of roof insulator leakage current monitoring technology was conducted.The development included hardware components such as signal acquisition,wireless transmission,power management,anti-interference,and encapsulation.An online monitoring hardware device for roof insulators was created,incorporating functions such as real-time display of leakage current,assessment of insulator pollution state,cleaning guidance,and periodic recording.This facilitated wireless monitoring of the roof insulator operating status on the ground or inside the vehicle.Testing and validation of the hardware device and software system were carried out based on actual vehicles.The results indicated that the developed roof insulator online monitoring device and system operated normally in real vehicle environments.The error between test data and experimental data was within 5%,providing effective support for the application of the theoretical,algorithmic,and device research results in practical engineering.更多还原