【作者】 黄俊辉；

【导师】 唐丽春； 卢乔；

【作者基本信息】 华南理工大学 ， 工业工程， 2013， 硕士

【摘要】 随着现代钢铁企业的快速发展，其中产品的质量控制占据越来越重要的地位。为确保产品质量，对整个企业生产过程中实行质量的控制和管理显得非常关键。当前信息技术的不断发展，而传统基于经验的产品质量性能分析已经满足不了现代企业的发展需求。因此，如何应用合适的信息技术从企业已积累的大量数据中发现影响钢铁产品质量的因素，并找出生产过程中质量控制中出现的问题，以及对产品质量进行预测分析，已成为现代化企业生产过程质量管理的一个重要发展方向。本文在前人研究基础上，重点开展了基于数据挖掘的钢铁产品质量控制应用研究工作，研究内容归纳如下：（1）针对企业质量管理以及钢铁企业质量管理的发展历程、国内外发展现状进行了综述分析，并对数据挖掘技术的现状进行了简单介绍。对数据挖掘中的数据仓库技术、决策树分析以及相应的算法进行了总结和归纳。（2）针对韶钢企业的实际背景，阐述了韶钢企业集团公司的基本现状，介绍了钢铁质量数据管理现状以及韶钢企业质量数据管理的流程和规范，分析了韶钢产品质量数据管理存在的问题以及由此产生的产品质量数据管理拟采取的措施。（3）以产品质量数据为研究对象，阐述了基于数据挖掘的产品质量管理控制思路；深入分析了韶钢产品的质量特征，建立了一种基于产品质量特征的数据描述模型，最后构建了基于数据挖掘的产品质量控制系统框架。在此框架基础上，简要介绍了在该架构下的数据挖掘过程以及设计了基于产品质量数据的挖掘软件系统结构。（4）通过对韶钢生产过程积累的产品质量数据进行分析，建立了产品质量预测数学模型，提出了一种时序数据挖掘技术，并应用于产品质量的动态控制；在此基础上，运用决策树分类技术，利用对积累的历史数据分析工艺参数对产品质量的影响，发现隐藏在数据中的潜在规律或规则，这些规则展示了隐藏在数据中的质量知识，可为质量控制提供有利的决策参考依据。（5）最后对全文进行了总结和展望。更多还原

【Abstract】 With the rapid development of modern iron and steel enterprises, the quality controloccupies an increasingly important position. It is very critical to implement the quality ofcontrol and management for ensuring the quality of the entire production. Traditionalexperience-based performance analysis of product quality has failed to meet the modernenterprises with the development of information technology. Therefore, how to find thefactors that affect steel product quality in the large amounts of data which the enterprise hasaccumulated, to identify issues that yield in the procedure of quality management, as well asto predictive the product quality, has become an important research direction for the qualitymanagement. In this thesis, on the basis of previous studies, we focus on the quality controlbased on the data mining technology for steel product. The research works of this thesis issummarized as follows:(1) According to the development process of enterprise quality management and thedevelopment of iron and steel enterprise quality management, the development situation ofquality management at home and abroad are introduced and analyzed. Besides, the datamining technology is introduced simply. Especially, the data warehouse technology, decisiontree analysis and the corresponding algorithms are summarized.(2) According to the background of Shaogang enterprise, first, the basic situation ofShaogang and the situation of quality data management are introduced. Second, the problemsof quality data management are analyzed. Finally, the measures are taken into account for theproduct quality data management.(3) First of all, the ideas of product quality management and control are introduced basedon data mining for the data of product quality; second, the features of the quality of productare deeply analyzed, then the data description of characteristics of product quality isestablished based on the frame of data mining. In the framework, the process of data miningis given and the software system structure is designed.(4) A quality prediction model is developed for the product data, and a time-series datamining tech is proposed for the application of the dynamic control of product quality in thethesis. On the basis, applying the decision tree method to analysis the influence of processparameters on product quality and to find the knowledge of quality data which hidden in thedata sets. Finally, the knowledge can be as the favorable decision-making reference. (5) Finally, the summary of the thesis is given.更多还原