【作者】 李鹏飞；

【导师】 丁常富； 赵明；

【作者基本信息】 华北电力大学 ， 动力工程， 2015， 硕士

【摘要】 如何利用煤耗实时在线监测系统积累的调峰运行数据,缓解云南火电机组长期低负荷运行经济性差的问题,而使机组在调峰运行时同样具有当前条件下最优经济性,则确定此时运行参数规则是关键。本文针对云南电网的某台300MW CFB机组调峰运行数据,探讨了数据处理的方法和原则。其次,为进一步找到大数据中隐藏的所需信息,本文引进属性模糊聚类和多维关联规则算法对数据进行分析。通过属性模糊聚类分析和多维关联规则算法对该机组的低负荷运行数据分析后,得到105MW~135MW、135MW~165MW、165MW~195MW、195MW~225MW负荷时机组热耗率最优的运行方案。结果表明属性模糊聚类分析能够在无专家指导的情况下简化影响机组经济性指标参数的构成,而多维关联算法可得到运行参数基准值。说明基于煤耗在线监测系统的数据挖掘算法是提高机组运行经济性的一个有效途径,可以为云南电网火电机组给予更加科学的运行指导。更多还原

【Abstract】 How to ulilize peaking operation data accumulated by coal consumption real-time monitoring system and resolve the problem of inefficiency due to long-term underload operation of Yunnan Thermal Power Unit to make the unit more economic under current condition? The Key is to determine the present operating parameter.This paper focuses on underload operation data of a 300 MW CFB Unit of Yunnan Power Grid, discussed the method of data processing and principle. Then, in order to find the needed information hidden in large data, this paper introduced the properties of fuzzy clustering and multidimensional association rules algorithm to analyze the data.Through the analysis of the unit underload data by property fuzzy clustering and multidimensional association rules, when loading 105MW-135 MW, 135MW-165 MW, 165MW-195 MW and 195MW-225 MW, unit heat rate difference reach optimization respectively, we get the corresponding parameter operating rules. The results that property fuzzy clustering analysis can simplify and influence the constitution of the unit economic indicator parameter and multidimensional association rules can obtain the operating parameter reference values, indicate that data mining algorithm based on coal consumption online monitoring system is an effective way to improve efficiency of the unit and can provide more scientific guidance for Thermal Power Unit of Yunnan Power Grid.更多还原