外文翻译-煤矿安全生产应急决策的数据集成研究系统

英文原文Data Integration Research of Coal Mine Safety ProductionSystem for Emergency Decision-makingPan Qi-dong；Zhang Rui-xin；Duan Dong-sheng ；School of Resources and Safety Engineering, China University of Mining and Technology(Beijing), Beijing 100083;School of Energy Science and Engineering, Henan Polytechnic University, Henan, JiaoZuo, 454003.ABSTRACT: Today the serious casualties of Chinese coal mines happened frequently and caused great casualties and property losses. This situation is extremely associated with scarcity of advanced emergency management(EM) methods for emergences forecastemergency decision-making;data integration;multi-source heterogeneousI. INTRODUCTIONNowadays, the situation of CMSP is still very serious. The EM functions implement incompletely and lack of advanced EM methods are the significant causes. Relying on information technology（IT） to solve CMSP problems, and establishing intelligent coal mine emergency decision support system, are extremely vital significance to CMSP. CMSP system is a huge system. It includes a series platforms of softwares and hardwares such as communication network system, ascension and adhesive tape transportation machine system, gas drainage system, drainage and pressing air system, power monitoring and controlling system, safety production monitoring system, and so on. It also includes command and organization system, legal rules and regulations, publicity, training, personnel and supporting materials, etc.Coal mine EM is an important segment to CMSP. It contains four processes, that is prevention and emergency preparedness, monitoring and early warning, emergency handling and rescue, restoration and reconstruction. It is a whole management process in coal mine operation. Data integration of CMSP is a prerequisite to EM system for emergences forecast and precaution and decision-making. Due to different construction time, using dissimilar technical standard and different subordinate departments, the data of information systems in CMSP simultaneously have all characteristics of system heterogeneous, model heterogeneous and sources heterogeneous. Actually, data integration of CMSP is a typical multi-source heterogeneous data integration problem.II. MODE AND OPTION OF DATAINTEGRATIONA. Concept of Data IntegrationData integration is a process of organizing the fixed data and heterogeneous data in different database systems, which aims to make different applications access to the shared data, and then establish more application functions. Data integration is a basic integration. It always is start of the application integration and information system integration. Fig.1 is the abstract model of data integration.CustomerSyt fDa IntgiDat ource1t ource2at Source3plxFigure1. A abstract model of data integrationB. Comparison and Selection of the Ways of Data IntegrationUsually, multi-source heterogeneous data integration has two modes. One is to transplant the original data into the new database management system, which includes federated database and data warehouse regularly. Another is the middleware . And the modes of data warehouse and middleware are popular and get more applicationData warehouse utilizes metadata models to identify and catalogue these scattered and independent databases. The data sources integrate into a global model after ETL (Data Extraction, Transformation and Loading) process, and then stored in the data warehouse. The data warehouse will wash those mass useless data during above process， and can be preferable used to carry on data mining and knowledge discovery through historical data analysis. On contrast，Middleware mode use a unified data logical view, which provided by a software located between a heterogeneous database system (data layer) and the application procedures (application layer), to hide the details of bottom data. Through this way, users can regard the data integration as a unified whole. It can provide general data interface upward to accessing the applications of integrated data. Through comparison，the later has advantage of the former on deal with the real-time data.Emergency decision data root in CMSP system, which need to know the real-time running state, while the historical mass data also should be made in-depth study and utilizing. In order to satisfy the need of EM, this paper builds a data center(data warehouse + middleware)as data integration mode of CMSP system. The middleware provide flexible independent data extraction and data integration for the data warehouse. It has been proven effective to integrate the historical, real-time and heterogeneous data.III. METAMODEL STRUCTURE OF COAL EMERGENCY MANAGEMENT SYSTEMIn order to clearly reflect the relationship between CMSP and EM, and grasp the panorama of EM system, it is necessary to establish metamodel structure of coal mine EM system before EM system is established. This metamodel is the most basic model, and it gives decision-makers and system designers a global guidance and support to get decision problems navigation.The causes of coal mine accidents and disasters are instability or loss of control of all kinds of dangerous sources. These dangerous sources always include 7 major categories of common hazards in coal mine, and also include floods, landslides and other natural factors outside. All kinds of dangerous sources are monitoring focus of CMSP, and also are data sources of EM system. Based on domain ontology and life cycle of disaster and accident, emergency decision process can be divided into three periods, namely beforehand preparation, immediate response and post-processing.During beforehand preparation period, EM system should support obtaining relevant data from all kinds of monitoring systems and production operation systems, and can research and analyze various risk factors and their development. This period includes 3 stages, that is monitoring and controlling, forecast and precaution, and emergency preparation. Monitoring and controlling stage should achieve real-time monitoring data for all kinds of dangerous sources and natural environment; emergency preparation stage includes planning and allocation of emergency resources, and accumulation of comprehensive knowledge of emergency; and the preparation stage must identify the risk signal using tools of data mining, simulation technology or experts judgments.The most important goal during immediate response period is to find effective ways and means to reduce casualties and property losses, which includes 2 stages, that is accident analysis and accident handling. During accident analysis stage, the system should provide as much information as possible for decision makers, including the occurrence of similar incidents in the past, the results of simulation, the current state and emergency preparations, and so on. During accident handling stage, the system should generate emergency preparation plans and rescue plans intelligently, and adjust the rescue plans based on real-time trace and assessment of the development trends of accident. During this stage, the rescue situation also should be bulletined in order to avoid chaos and appease the families of trapped people.The last post-processing period includes restorations or reconstructions and assessments stages. Among them, the assessments include assessments of casualties, property losses, derivatives and secondary disasters caused by disasters and accidents, and also include assessments of effectiveness of EM system, EM level, emergency prevention ability, emergency response capacity, emergency sources allocation level, and emergency rescue capability and so on.IV. DATA INTEGRATION SYSTEM DESIGN FOR COAL MINE EMERGENCY DECISION-MAKINGA. Data Integration Model for Emergency Decision-makingCoal mine emergency decision-making requires not only the accurate analysis of historic data but also should reflect the changes of the real-time information. Based on the advantages and disadvantages of data warehouse and middleware, and combined with current situation of CMSP system, the data integration system architecture based on data center is presented , including data source layer, data center layer and data application layer. As shown in Fig2.Source data layerIt mainly includes many database systems which be used in coal mine manufacture or monitoring operations. Because of different construction time, different technical standards and different subordinated management departments, these database systems are established in different database management system, typically SQL Server, Oracle, Sybase, etc.Stae AsmntDecison AalyisForecast AnlyiRTM(ODS) EDWDSMOMDTADEKDESDLHDDplicoLyr pat CenryTLxchangeamic dtLyer taic dTat oure EDDW: Emergency Decision-making Data Warehouse;RTM: Real-time Middleware;DSMD: Dangerous Sources Monitoring Database;MOMD: Manufacturer and Operation Monitoring Database;TAD: Targeted Accidents Database;EKD: Emergency Knowledge Database ;ESD: Emergency Sources Database;Figure2. Data integration system for coal mine emergence decision-making Data center layerdata center layerThis layer is a data warehouse which adds a real-time middleware. The data integration of data warehouse is the data loaded timely from OLTP（On Line Transaction Processing）system to data warehouse. Through this way, the decision-makers can timely access to the customer information and give a correct decision. On contrast, to data center, its data storage is divided into dynamic data storage and static data storage. Dynamic data is stored in the ODS (Operational Data Store) layer, and static data is stored in data warehouse. The data warehouse face to subject-oriented, stable, integrated, asynchrony data, and reflect the historical data. And its data load time always lag behind that of the business systems. After introducing real-time middleware, the dynamic data of the underlying data source can be in real-time loaded and update on the ODS layer directly and then stored in the data warehouse. At the same time, the static data is loaded into the data warehouse through the ETL tool, so that the research and analysis requirements of decision-makers for the real-time and historical data can be met.In addition, ODS is the operational data store, the same as the data warehouse, which has the characteristics of subject-oriented and integrated, but its data have variability that means the data can be current or near to current. According to the demand for coal mine EM, we choose real-time messaging middleware to achieve synchronization business systems data through real-time freshing ODS data.Data Application LayerThis layer views data-center as its platform. With the help of the real-time middleware, many applications systems, such as the situation analysis system, decision analysis system and the analysis and forecasting systems, etc, can be developed in this layer. And many corresponding application data are also generated in this layer.B. Conversion of the Heterogeneous Source DataThe methods of data conversion between heterogeneous data sources mainly includes several methods based on some tools, such as software, intermediate database and database component, etc. In order to achieving the purposes of putting the data into the data center, all conversion methods must find a suitable data model and exchange norms to solve the differences of data sources caused by the different storage structure and platforms. Our system chooses XML data model, which is a good information exchange language and uses a semi-structured data model. XML data is a self-describing semi-structured data, which has advantages especially in data model and data exchange standard, and can easily realize the packaging of resources and releasing of integration. In order to achieve the data transformation between a relational database and application programs, the key is to realize the mapping between XML and databases. There are two general mapping methods, that is template-driven mapping and model-driven mapping. In practice, considering that CMSP system comprises of numerous complex information systems and EM integration will involve many database table, we choose model-driven as our mapping methods, and utilize the dominant or recessive data structure of database to map into a model structure of the XML document.V. CONCLUSIONToday, the rapid development of information technology（IT）gives the informationization of coal mine a good opportunity, and also provides coal mine a strong support to establish EM system to keep away various accidents and disasters. On the basis of management characteristics and its actual requirements of coal mine emergency, this paper proposes data center model to integrate the multi-source heterogeneous data of CMSPS, then design and research the coal mine emergence management system and data integration system. It is a good reference to system integration of coal mine safety and its establishment of emergence system.REFERENCES:1 Widom J, “Research Problems in Data WareHousing”. In Proceedings of the 4th Int'L Conference on Information and Knowledge Management(CIKM), November 1995；2 Xue Hui-zhong, Zhuang Xiao-qing, Dong Yi-sheng. Data Transformation in the Data Warehouse J. Modern Computer, 2003；3 Wang Yu-ming. Design of a B2B Enterprise Information System Based on EAIJ, Application Research of Computer, 2005；4 Xu Fang, Zhou Jun, Lu Hao,Research on Data Integration Architecture in the Early Warning Surveillance Information System J. Journal of Air Force Radar Academy, 2008；5 Wang Ni-hong, Liu Mei-ling. ODS Operational Data Store-data warehouse new technology J. Information Technology，2004；中文译文煤矿安全生产应急决策的数据集成研究系统潘祺东;张瑞昕;段东生;中国矿业大学资源与安全工程学院（北京） ，北京 100083;河南理工大学能源科学与工程学院，河南，焦作，454003摘要：今天的中国煤矿重大人员伤亡频繁发生，造成重大人员伤亡和财产损失。这种情况是因为预测和预防灾害和决策的应急管理缺乏相关的先进应急决策方法。煤矿安全生产系统（CMSP）是解决上述问题的基础。根据多源异构系统和CMSP的特点，分析了几种方法相结合的数据，并得出结论认为，数据仓库和中间件组合（名为数据中心）可以有效地解决问题。在此基础上，建立了煤矿应急决策防盗系统的元框架模型，设计出了煤矿应急决策数据集成系统。最后，对异构源数据转换问题进行了分析。关键词：煤矿；应急决策；数据集成；多源1 引言如今，煤矿安全生产系统（CMSP）的形势依然十分严重。执行功能不完全的应急决策和缺乏先进的应急决策方法是重要原因。依托信息技术（IT）来解决 CMSP问题，建立智能的煤矿应急决策支持系统，对于煤矿安全生产系统（CMSP）来说具有极其重要的意义。煤矿安全生产系统（CMSP）系统是一个庞大的系统。它包括一系列软件和硬件平台如通信网络系统、提升系统、胶带输送机系统、瓦斯抽放系统、压风系统和排水系统、电力监控系统、安全生产监控系统等等。它还包括指挥和组织体系、法律法规、宣传、培训、人员和辅助材料等。煤矿应急决策是煤矿安全生产系统（CMSP）的一个重要组成部分。它包含四个过程，分别是预防和应急准备、监测和预警、应急处置和救援、恢复和重建。这些贯穿在煤矿经营管理的全过程。对于应急决策防盗系统来说 CMSP数据融合是预报及防范灾害决策的先决条件。由于不同的施工时间，采用不同的技术标准和不同的下属部门。在煤矿安全生产系统（CMSP）信息系统的数据同时拥有所有异构系统的特点，模式和来源的异质异构。其实，煤矿安全生产系统（CMSP）数据一体化是一个典型的多源异构数据集成问题。2 模式及数据集成方案2.1 数据集成的概念数据集成是一个固定的组织数据和不同的数据库系统，其目的是使不同的应用访问共享数据的异构数据处理，然后建立更多的应用功能。数据整合是一个基本的集成。它始终是应用集成和信息系统集成的开始。图 1是抽象的数据集成模型。 中中12中3图 1   一个抽象的数据集成模型2.2 比较和数据整合的途径选择通常情况下，多源异构数据集成有两种模式。一种是移植原始数据到新的数据库管理系统，其中包括联邦数据库和数据仓库定期的原始数据。另一种是中间件。数据仓库的模式和中间件都很受欢迎，得到更多的应用。利用数据仓库元数据模型，以确定和编目这些分散的，独立的数据库。数据来源通过 ETL（数据抽取，转换和加载）等过程整合成一个全球模式，然后存储在数据仓库中。数据仓库将在上述过程中清洗大多数无用数据，通过分析历史数据，可较好的用于进行数据挖掘和知识发现。相反，中间件模式使用一个统一的数据逻辑视图，它提供了一个异构数据库系统（数据层）和应用程序（应用层）之间，以隐藏底层的细节资料位于一个软件。通过这种方式，用户可以看成是一个统一整体的数据集成。它可提供通用数据接口向上访问数据的综合应用。通过比较，后者与实时数据处理相比具有优势。应急决策系统数据的根在煤矿安全生产系统（CMSP） ，其需要知道的实时运行状态，而大量的历史数据也应在深入研究和利用。为了满足应急决策系统的需要，本文建立了一个数据中心（数据仓库+中间件）作为煤矿安全生产系统（CMSP）系统集成模式。 。该中间件提供了灵活的数据仓库独立的数据提取和数据集成。它已被证明是有效整合的历史、实时和异构数据。3 煤矿应急管理体系的元模型结构为了清楚地反映了煤矿安全生产系统（CMSP）和应急决策的关系，把握应急决策防盗系统全景，有必要在建立应急决策系统之前建立煤矿应急决策防盗系统元模型。这个元模型是最基本的模式，这给决策者和系统设计师提供了全球性的指导和支持以便得到决策问题的导航。煤矿事故和灾害的原因是各种危险源的不稳定或失控。这些危险源总是包括 7个主要类别的常见煤矿危害，还包括洪水，山体滑坡和其他外在自然因素。各类危险源是煤矿安全生产系统（CMSP）监控的焦点，也是应急决策系统的数据来源。基于领域本身和生命周期的灾难和事故，应急决策过程可分为三个阶段，即事前准备，立即作出反应和事后处理。在事先准备期，应急决策系统应支持从监测系统获得的各种相关数据和生产运作系统的各种相关数据，可以研究和分析各种风险因素及其发展。这个时期包括 3个阶段，即监测和控制，预测和预防，应急准备。监测和控制阶段应实现对危险源和自然环境的各种实时监测数据，应急准备阶段，包括规划和应急资源配置以及全面的应急知识的积累，准备阶段必须确定使用的危险信号数据挖掘工具，仿真技术或专家的判断。在即时的反应期间最重要的目标是要找到有效的方法和手段，以减少人员伤亡和财产损失，其中包括 2个阶段，即事故分析阶段和事故处理阶段。在事故分析阶段，系统应该为决策者提供尽可能多的信息，包括在过去类似事件的发生，模拟的结果，当前状态和应急准备工作等等。在事故处理阶段，该系统应该智能地产生应急救援预案编制计划，并在调整救援计划的基础上实时跟踪和评估事故的发展趋势。在这个阶段，救援形势也应透明，以避免混乱局面的出现和安抚被困人员的家属。最后的后处理期间包括修复或重建和评估阶段。其中，评估范围包括评估的人员伤亡，财产损失，因灾害和意外事故造成的衍生及次生灾害，也包括对应急决策防盗系统，应急决策水平，应急预防能力，应急反应能力，应急资源配置水平和紧急救援能力等。4 煤矿应急决策的数据集成系统设计4.1 应急决策的数据集成模型煤矿应急决策不仅需要准确地分析历史数据，而且还应该反映实时信息的变化。鉴于数据仓库和中间件的优点和缺点，并与煤矿安全生产系统（CMSP）制度的现状相结合，提出了基于数据中心的数据集成系统建设，包括数据源层，数据中心层和数据应用层。如图 2所示。数据源层它主要包括应用在煤矿生产或监测业务的许多数据库系统，由于不同的施工时间，不同的技术标准和不同服从管理部门，这些数据库系统都建立了不同的数据库管理系统，通常的 SQL服务器，Oracle，Sybase 等。中中中 实 时 中 间 件(运 营 数 据 存 储 层 )ETLETLEDDW：紧急决策的数据仓库;RTM：实时中间件;DSMD：危险源监控数据库;MOMD：制造商和运行监测数据库;TAD：有针对性的事故数据库;EKD：应急知识数据库;ESD：应急资源数据库;图 2  数据煤矿出现决策层的数据中心集成系统数据中心层这一层是数据仓库增加了一个实时中间件。数据集成的数据仓库是数据加载及时的 OLTP（在线交易处理）数据仓库。通过这种方式，决策者能够及时获得客户信息，并给予正确的决定。对比数据中心，其数据存储分为存储到动态数据存储器和静态数据存储器。动态数据存储在 ODS（运营数据存储）层，静态数据存储在数据仓库中。数据仓库面向主体，稳定，完整，不同步的数据，反映了历史数据。其数据加载时间总是落后于业务系统。在介绍实时中间件，基础数据源的动态数据可以在 ODS层实时加载和更新，然后存储在数据仓库中。同时，静态数据通过 ETL工具加载到数据仓库，以便满足决策者对实时数据和历史数据研究和分析的需要。此外，ODS 是运营数据存储层，作为数据仓库，它的特点面向主题和集成，但其数据有变化，这意味着数据可以是当前或接近当前的。根据煤矿应急决策的需求，我们选择实时消息中间件通过实时数据保鲜 ODS数据来实现同步业务系统。数据应用层这一层的意见作为其平台的数据中心。在实时中间件的帮助下，许多应用系统，如形势分析系统，决策分析系统和预报系统等，可开发在这一层。和许多相应的应用程序数据也产生在这一层。4.2 异构数据源的转换数据与异构数据源之间的转换方法主要基于一些以工具为基础德方法，如软件，数据库和数据库的中间部分等，为了实现将数据放入数据中心，所有的转换方法必须找到一个合适的数据模型和交换规范来解决不同的存储结构和平台造成的数据源的差异。我们的系统选择 XML数据模型，这是一个良好的信息交流的语言，并使用半结构化数据模型。 XML数据是自描述的半结构化数据，这在数据模型和数据交换标准方面具有明显的优势，可以很容易地实现资源整合包装和释放一体化。为了实现关系数据库和应用程序之间的数据转换，关键是要实现 XML和数据库之间的映射。一般有两种映射方式，即由模板驱动的映射和模型驱动的映射。在实践中，考虑到煤矿安全生产系统（CMSP）成员包括许多复杂的信息系统和应急决策将涉及许多数据库表，我们选择的模型驱动作为我们的测绘方法，并利用显性或隐性的数据库的数据结构映射到一个 XML文档的模型结构。5 结论今天，信息技术的快速发展给了煤矿信息化一个很好的机会，并提供煤矿一个强有力的支持建设应急决策系统以防范各类事故和灾害。基于管理的特点和煤矿的实际紧急需要的基础上，本文提出的数据中心模式，整合煤矿安全生产系统（CMSP）多源异构数据，然后设计和研究煤矿出现的管理系统和数据集成系统。这对煤矿安全集成系统及其建立是一个很好的参考。参考文献：1 Widom J，“仓储”数据研究的问题，第四届信息和知识管理（CIKM）国际会议论文集，1995 年；2薛惠忠，庄小庆，董易生。数据转换在数据仓库的研究，现代计算机，2003；3王育明。基于 EAI研究系统的 B2B企业信息设计，计算机应用，2005；4许芳，周军，卢昊，数据集成体系结构研究中的预警监测信息系统J。空军雷达学院，2008 年； 5汪逆红，刘美灵，操作数据存储 ODS的数据仓库新技术，信息技术，2004；