时序数据挖掘在电站设备监测中的应用研究

发布时间：2018-10-17 11:38

【摘要】：当前,在电力企业中信息化建设发挥着无与伦比的重要作用,在信息化建设中数据库技术已被越来越广泛使用,着眼于数据库技术对设备运行的大规模历史数据的应用也日趋重要。伴随着电力企业运行设备的大数量级历史数据的积累,过往的关系型数据处理方式,很难满足用户的需求。现存保守的数据处理,只能满足对数据进行初级的分析,但是并不能挖掘出海量数据中隐含的极具现实意义和价值的知识,而且更无法从时和空意义上更深入渗透的理解海量历史数据,结果使数据的利用率显然不足。所以,众多拥有“历史数据宝库”的企业,受限于知识的匮乏,不能深层挖掘出“宝库”中有价值的、隐含的信息。论文着重于当前电站企业共同面临的海量历史数据得不到最有价值有效利用的问题,利用数据仓库知识,数据挖掘技术和数据挖掘算法,探索出了一个可以满足电站设备负责管理人对电站设备进行有效合理管理的解决方案。论文在深层次理解电站设备运行原理的前提上,运用数据仓库相关的知识理论,并且运用时序数据挖掘算法对电站的海量历史数据进行全方面的分析。时序数据挖掘的主要技术与方法是ARTXP算法和ARIMA算法,在对时序数据信息进行KDD(知识发现)的过程中为了制定出比较合理的长期或短期预测,进而为合理决策提供有效的支持,通过对这两种挖掘算法的研究与分析,创建ARTXP算法和ARIMA算法的有效混合模型是理想的选择,进而能够提高趋势分析的准确度与时效性发掘出给定的历史数据中存在的规则、知识。最后对电站设备中最核心的设备凝汽器的真空度进行判断分析,发现历史数据显示的影响凝汽器真空度的最主要因素,最后,给出科学合理的检修监测措施。
[Abstract]:At present, the information construction plays an unparalleled important role in the electric power enterprise, and the database technology has been more and more widely used in the information construction. The application of large-scale historical data based on database technology is becoming more and more important. With the accumulation of the large order of historical data of the power enterprise operating equipment, the past relational data processing method is very difficult to meet the needs of users. The existing conservative data processing can only satisfy the primary analysis of the data, but it cannot mine the most practical and valuable knowledge implied in the massive data. Moreover, it is more difficult to understand the massive historical data more deeply in the sense of time and space, the result is that the utilization rate of the data is obviously insufficient. Therefore, many enterprises with "historical data treasury", limited by the lack of knowledge, can not dig out the valuable and hidden information in the "treasure house". This paper focuses on the problem that the massive historical data which the power plant enterprises are facing together can not be used most valuable and effectively, using the knowledge of data warehouse, the technology of data mining and the algorithm of data mining. In this paper, a solution that can satisfy the effective and reasonable management of power plant equipment by the responsible manager of power plant equipment is explored. On the premise of deeply understanding the operation principle of power plant equipment, this paper applies the knowledge theory of data warehouse and uses the algorithm of time series data mining to analyze the massive historical data of power station in all aspects. The main techniques and methods of time series data mining are ARTXP algorithm and ARIMA algorithm. In the process of KDD (knowledge discovery) of time series data information, in order to make reasonable long term or short term prediction, and then provide effective support for reasonable decision making. Through the research and analysis of these two mining algorithms, it is an ideal choice to create an effective mixed model of ARTXP algorithm and ARIMA algorithm, which can improve the accuracy and timeliness of trend analysis and discover the rules existing in the given historical data. Knowledge. Finally, the vacuum degree of condenser, which is the core of power plant equipment, is judged and analyzed, and the main factors that influence the vacuum degree of condenser are found out by historical data. Finally, the scientific and reasonable maintenance monitoring measures are given.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM62;TP311.13

【相似文献】