【作者】 冉鹏；

【导师】 王松岭； 张树芳；

【作者基本信息】 华北电力大学 ， 热能工程， 2012， 博士

【摘要】 针对动态数据流的数据挖掘研究是一个具有挑战性的新兴领域,动态数据流挖掘在金融行业、电信行业、电力行业、冶金行业等领域有着极其广泛的应用前景。传统针对静态历史数据的数据挖掘技术不能有效地分析和处理动态数据流这类大规模动态增长的数据。本文结合电站热力系统运行优化研究的热点问题,在总结经验、发现问题和不足的基础上,从电站机组热力系统节能分析方法出发,结合仿真技术,构建了电站热力系统实时性能分析与运行优化仿真平台,并将动态数据挖掘技术引入到电站热力系统运行优化研究,主要内容如下：将图论思想引入到电站热力系统节能分析研究中,确定了基于图论的电站热力系统抽象原则,规定了电站热力系统的划分原则及基于图的表达方法。在比较不同电站机组结构特点的基础上,按照基于图的电站热力系统表示方法,确定了一次再热火电机组、二次再热火电机组及核电机组热力系统的有向图带权邻接矩阵填写规则,推导出基于图论的通用电站机组热力系统节能分析模型,将不同类型机组热力系统节能分析纳入到统一的框架下,提高了对不同类型机组进行热力系统分析的通用性。将基于图论的通用电站机组热力系统节能分析模型与电站仿真过程有机的结合起来,结合构件复用技术开发了电站热力系统实时分析与运行优化仿真平台。研究了数据挖掘技术中的关联规则挖掘技术,结合当前动态数据流广泛存在的现实与动态数据挖掘技术的广泛应用前景,以及高速、动态变化的数据流的特点,将模糊集理论引入到动态数据流的数值型数据关联规则挖掘中,提出了一种适合动态数据流环境的模糊关联规则挖掘方法。以挖掘典型负荷状态下锅炉烟气含氧量与其它机组运行参数的关系为例,说明了动态数据模糊关联规则的挖掘过程。分析了当前电站热力系统运行优化目标值获取方法的优缺点,概述了基于动态数据挖掘的电站运行优化系统的体系结构,提出了基于动态数据挖掘的电站热力系统运行优化目标值的获取方法。对电站热力系统运行优化中的最经济煤种决策方法进行了研究,分析了煤质变化对火电厂运营成本的影响,提出了发电企业复合燃煤使用成本的计算模型,基于该模型提出了最经济煤种的决策方法,并应用动态数据挖掘技术与静态历史数据挖掘技术确定了最经济煤种。更多还原

【Abstract】 Data mining for dynamic data stream is a challenging newly emerging field, dynamic data mining has wide application prospects in the field of financial industry, telecommunications industry, power industry, metallurgical industry, etc. Traditional mining algorithms are difficult to cope with data stream due to its characteristics, such as consecution, disorder and real-time.This paper mainly summarized the hot issues of power plant operation optimization, including the method of power plant thermal economy analysis and the existent problems of power plant operation optimization, established the Real-time Power Plant Energy-saving Analysis&Operation Optimization Simulation Platform, introduced dynamic data mining technology into the research of power plant operation optimization. Mainly contents as following:Based on the analysis of the structure features of power plant, graph theory are introduced in the thermal economy analysis, a new unified rule for analyzing the power plant thermal system are established. Combined with first thermodynamics law and mass conservation law, different types of power plant unit’s Weighted Diagraph Adjacency Matrix is deducted. The Weighted Diagraph Adjacency Matrix is of standard and concise form, whose physical meaning is well-expressed. A unified Real-time Power Plant Energy-saving Analysis&Operation Optimization Simulation Platform is established by the power plant unit’s Weighted Diagraph Adjacency Matrix.Reviewed and analyzed the association rule mining techniques and the characteristics of the dynamic data stream, this paper proposed a new association rule mining method based on fuzzy association rule. An example is given to illustrate the usage of the method.Analyzed the advantages and disadvantages of the decision method of power plant operation optimization target value, a new deciding power plant operation optimization target value method based on dynamic data mining technology was proposed. Based on the association characteristic of the power plant operating data, this paper proposed a convenient, low cost method for calculating the most economical coal. Four kinds of pre-purchase of coal are given to illustrate the validity of the method, and the most economical coal is determined.更多还原