矿山企业数据仓库的应用研究

【作者】 李学锋；

【导师】 段希祥；

【作者基本信息】 昆明理工大学 ， 矿物加工工程， 2005， 博士

【摘要】 矿业是国民经济的基础,矿业的可持续发展是国民经济可持续发展的重要保证;而信息化是提高矿业生产管理水平的重要措施之一。“以信息化带动工业化,以工业化促进信息化”是加快实现矿业现代化的必然选择。本文在全面分析我国有色矿山企业信息化现状、所处阶段和存在的问题的基础上,结合矿山企业的生产经营特点,对矿山数据仓库建设的理论与方法进行了系统研究,设计出大型矿山企业数据仓库的整体解决方案。论文主要内容如下: 1、针对矿山企业的生产经营特点和信息流动特点,开展矿山数据仓库的框架结构设计。矿山数据仓库的框架结构由数据仓库访问层、数据仓库功能层、数据仓库管理层和数据仓库环境支持层组成;该框架结构为矿山数据仓库建设提供了一种通用的“蓝图”,有利于矿山管理者、系统设计者和实现者的沟通和协作,用以指导矿山数据仓库的建设。 2、根据矿山数据仓库的框架结构和数据仓库的相关理论,通过对玉溪矿业公司的业务需求分析,确定了玉溪矿业公司数据仓库建设的具体目标和任务,设计了数据仓库的6个主题域,即资源管理、原矿管理、精矿管理、财务管理、安全管理和设备物资管理,这6个主题域基本涵盖了玉溪矿业公司的生产经营的各个方面。并以系统工程思想为指导,对玉溪矿业公司数据仓库的数据采集系统、中心数据库系统、应用分析系统、数据访问层、数据仓库环境层进行了规划设计。 3、数据采集是矿山数据仓库建设中的重要一环,根据数据仓库的数据要求及玉溪矿业公司信息化建设状况,论文设计了两种采集方式,即基于C/S架构的数据自动采集传输方式和基于B/S架构的数据填报方式,并规划了数据采集的范围和内容,研究了数据过滤机制、数据整合机制,以解决企业各管理层之间、各业务部门之间、各应用系统之间数据独立和相互隔离的问题。 4、建立数据仓库的主要任务是建立企业全局性、基础性的中心数据库,中心数据库包括基础代码库和业务数据库。论文进行了矿山基础代码整理、统计口径整理和矿山指标体系设计,着重进行了关键绩效指标(KPI)的设计;最终使矿山数据达到指标规范、口径一致、数据字典标准。矿山中心数据库按BKR模型设计成三个层次,以实现数据的稳定性与应用的灵活性相统一。多维数据存储采用菱形降维技术对维度进行分解,以提高查询分析的响应速度。 5、设计了集数据仓库、OLAP、数据挖掘于一体的矿山生产经营分析系统并应用于生产实际中。用户可以通过多种手段从矿山数据仓库中获取有用的信息,如信息查更多还原

【Abstract】 Mining industry is the national economic basis, and its sustainable development is an important guarantee of the national economy. Informationization is one of the most powerful measures to improve mining management level. In order to realize mining modernization quickly, "Applying informationization to accelerating industrialization and applying industrialization to promoting informationization" should be executed certainly. After a complete analysis of the current situation,development phase and existing problems of informationization in nonferrous mines, the theories and methods of the Mine Data Warehouse(MDW) are studied systematically according to the characteristics of the nonferrous mines, and the overall solutions for constructing enterprise level Data Warehouse(DW) in large-scale mine are put forward. The main contents of this dissertation can be summarized as follows:1. In accordance with the characteristics of the management and data-flow in mines, the MDW framework is designed, which consists of DW access layer, DW function layer, DW administration layer and DW environment support layer.The MDW framework serves as a blueprint of MDW construction, on which the mine managers, DW designers and executers can communicate and cooperate with each others, and it also provides the guidance for users to construct the MDW application system.2. Based on the above-studied results and related theories of DW, the specific construction task of the Yuxi Mining Corporation DW (YXDW) is defined after the business requirement analysis is completed. Six subject domains of YXDW are defined as well, that is, mineral resource management, ore management, concentration management, finance management, safety management and equipment & material management, which cover main aspects of production and management in Yuxi Mining Corporation. Subsequently, data acquisition system, center database system, analysis application system, data access system and DW environments in YXDW are planned under the guidance of system engineering.3. Data acquisition is the key step for MDW.According to data requirement of MDW and information system situation in Yuxi Mining Corporation, two data acquisition models are designed, that is, data acquisition and transmission automatically based on C/S architecture, data reporting based on B/S architecture. At the same time, data acquisitionextent and content are planned. Then, data cleaning mechanism and data conformity mechanism are discussed in detail for purpose of solving the problem of conflicted and isolated data not only in enterprise management organizations and departments but also in computer application systems.4. The major task of MDW construction is to build a complete, basic enterprise center database, which contains basic-code database and business database. So basic-code and statistic caliber are sorted out and mine index system, especially Key Performance Indicators (KPI) is designed. A normative data index, consistent statistic caliber and standard data dictionary will be the final pattern of the mine data. Center database is arranged as 3 hierarchies according to BKR model so as to meet the requirements of DW management. Rhombus-dimension-reduction technology is adopted to decompose the dimensions in multi-dimension data storage so as to raise the computer answer speed of the search analysis.5. Yuxi Mining Corporation Production & Management Information System (YXPMIS) is built successfully and put into service, which is based on the integration of MDW, OLAP and Data Mining (DM). Mine managers can obtain useful information from the MDW to monitor the production totally and accurately, and to make production & management decision in time and scientifically by various means, such as data query, OLAP and DM.The significance of this paper for mine informationization is that it provides a mining-object oriented architecture and method for building a MDW in response to data acquiring, storage, analysis and decision supporting. The problem of "information islands" in mines can be solved efficiently by taking advantage of data centralized-management, data exchange and data sharing in the MDW. The MDW provides an integrated data analysis environment for mine managers to master total, accurate and timely enterprise data and information so as to make decision in time and scientifically, and it improves and standardizes the information criterion in mines, even in the whole nonferrous metal industry.The contents of the paper, especially the presented methods and solution techniques, can provide useful reference to build data warehouse in other large mine enterprises.更多还原