Research On Information Aggregation Of Power Equipment Condition Monitoring For Large Data

Modern power system, with its higher voltage and capacity than ever before, poses higher standards in electrical equipment security and reliability. Therefore, condition monitoring technology has increasingly drawn more attention. With the in-depth development of power system informatization and intellectualization, there are more condition monitoring types and more advanced monitoring methods for electrical equipment, helping to assessing electrical equipment and detecting faults to a certain degree. But, along with the exponential growth of stored monitoring data as time goes by, those traditional data analysis methods have met difficulty:first, data collected in long and constant monitoring are amazingly large in electrical equipment monitoring system; second, along with the increasingly abundant monitoring means arise a large number of non-traditional power monitoring methods, such as sound signals and video signals, and monitoring data types become more diversified; third, the constantly higher data processing speed required in modern power system, plus the defects of condition monitoring mechanism, including low measurement accuracy and poor reliability, make it one of the most important research subjects in current condition-based monitoring that how to analyze data quickly and effectively. Yet the appearance of big data just provides a new idea and method to solve the problems in the subject.At the beginning, this paper summarized the current research status of dig data and power big data, and focused on analyzing the characteristics of power big data, the data management, analysis and processing techniques. Then this paper, starting from the difficulty existing in current electrical equipment condition monitoring, elaborated the application idea of big data in condition monitoring, thus established a condition method, i.e. to take "prediction" as the core target, correlation analysis and information aggregation as main means, to weaken monitoring data accuracy and attach importance to data analysis speed. After that, it accordingly put forward a multidimensional information aggregation method architecture of electrical equipment for big data, based on the basic architecture of multiple information aggregation. A multidimensional support vector based on the various state quantities of electrical equipment was established, historical data were used for training, and a constantly growing electrical equipment support vector collection was formed, to achieve the assessment and decision making of electrical equipment operation, the results of which were oriented to the lifecycle management system of substation equipment, through the correlation analysis of condition monitoring data and support vector collection. In the end, this paper, taking the overheating fault in transformers as an example, explicated the specific aggregation method of electrical equipment, and studied the condition monitoring platforms for big data via the CyberControl software, realizing data access and analysis of information aggregation.