基于大数据的运检培训技术研究

发布时间：2018-11-06 15:12

【摘要】：随着电网建设的快速发展,电网系统的规模与复杂程度也日益增大,各信息系统已经积累了大量输变电设备状态信息数据。这给运检培训带来新的问题与挑战。而传统的人工教学培训具有成本昂贵、培训效果不高的特点。因此推动运检培训向标准化、高效化、网络化、形象化发展已经是未来的趋势。随着电网设备运行状态检测技术的发展,运检决策人员所掌握的设备状态信息量迅速增长,另一方面,信息的组成从以文字和数字等结构化数据为主发展到图像和视频等非结构化数据占有数据量的绝大部分。随着机器人和无人机等自动检测平台在输变电巡检中的应用,非结构化数据的分析需求将大大增加,这些数据的特征和规模导致依靠专家的人工分析已经无法及时处理,制约了上述先进检测手段在运检决策中发挥作用。本文通过研究基于输变电设备在线监测、带电检测等多源异构数据的检测、状态评价与检修技能的培训与考核技术,实现培训系统与状态评价系统的数据共享和知识共享,提高培训水平和效率。本文针对运检多源异构检测数据,研究故障与检测数据间关联性,提取故障特征信息,并通过对变压器、断路器典型故障案例库建模与分析,红外影像库和超声超高频检测图谱库的建模与分析,建立典型故障库和红外检测图谱库在线培训系统,供学员学习。最后基于多源异构数据、故障案例库建立基于大数据的输变电运检培训系统,并开发手机应用,方便学员学习。总体研究表明,本文所研究的基于大数据的运检培训技术,为运检培训提供了一种新方法,可有效提高设备故障诊断和检修决策水平,能够产生显著的经济和安全效益。
[Abstract]:With the rapid development of power grid construction, the scale and complexity of power grid system is increasing day by day. The information system has accumulated a large number of information data of transmission and transformation equipment. This brings new problems and challenges to operation and inspection training. The traditional manual teaching and training has the characteristics of high cost and low training effect. Therefore, it is the future trend to promote the standardization, high efficiency, network and visualization of operation and inspection training. With the development of power network equipment operating state detection technology, the amount of equipment state information grasped by operation inspection decision makers increases rapidly, on the other hand, From structured data such as text and number to unstructured data such as image and video, the information consists of most of the data. With the application of automatic detection platform such as robot and UAV in power transmission and transformation inspection, the demand for analysis of unstructured data will be greatly increased. The characteristics and scale of these data make it impossible to deal with them in time depending on the manual analysis of experts. It restricts the above advanced detection means to play a role in the operation and inspection decision. In this paper, the data sharing and knowledge sharing between the training system and the condition evaluation system are realized through the research of the multi-source heterogeneous data detection, such as on-line monitoring of transmission and transformation equipment, live detection and so on, as well as the training and examination technology of condition evaluation and maintenance skills. Improve training level and efficiency. In this paper, we study the relationship between fault and detection data, extract fault characteristic information, and model and analyze the typical fault database of transformer and circuit breaker. The modeling and analysis of infrared image library and ultrasonic ultra-high frequency test atlas library are carried out, and the online training system of typical fault database and infrared detection atlas library is established for students to learn. Finally, based on the multi-source and heterogeneous data, the fault case database is used to establish the training system of transmission and substation operation based on big data, and the mobile phone application is developed to facilitate the students to learn. The overall research shows that the training technology based on big data provides a new method for operation and inspection training, which can effectively improve the level of equipment fault diagnosis and maintenance decision, and can produce significant economic and safety benefits.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13;TM76

【参考文献】