【作者】 张炜；

【导师】 张勇传； 周建中；

【作者基本信息】 华中科技大学 ， 系统分析与集成， 2017， 博士

【摘要】 水是人类赖以生存的重要资源。如何实现水资源的有效管理与高效利用一直是水行业中诸多专家学者重点研究的问题。面对日益严重的水资源危机、日趋复杂的水管理难题、以及水信息“数据大爆炸”时代的到来,水资源管理领域对水信息系统软件提出了“新技术、大平台、互共享、广应用”前瞻性的应用需求。然而,传统面向过程、面向对象、面向构件的软件结构体系已无法支撑如今水资源管理体系多元化、多维化、多层化的发展模式。因此,亟待探究新的水利信息化理论架构与技术体系来匹配日益复杂的水资源管理模式,从而解决信息资源共享、多元业务耦合、多目标寻优决策等一系列管理难题。另一方面,随着计算机网络的高速发展与广泛应用,围绕网络服务为核心的分布式软件体系在金融、电商、营销、工管、通信及教育等诸多领域发挥了重要作用,并取得了长足的发展。但是,在水资源管理领域,分布式网络环境下的水信息系统架构研究尚处于起步阶段,支撑理论和应用技术尚不成熟。为此,基于水利信息化应用的跨时空、跨区域和分散性特点,结合水资源管理中的业务服务化、信息共享化理念,以及水资源领域对地理信息技术的强大依附性,在综合考虑行业多元业务运作模式与集成需求的基础上,本文首次研究并提出了具有普适性意义的分布式水信息系统软件架构,并对架构的相关理论、通信方法、作业机理、集成模式进行了系统阐述。本文的主要研究成果和创新点如下:(1)深入探讨了分布式体系下的概念性底层架构——面向服务架构(SOA)和基于Web Service中间件的通信服务实现方法,提出了常规模式和非常规模式下分布式体系中的软件元结构。进一步,围绕以上有关架构概念、技术主题、软件元体系,并结合水资源领域水文学、水力学、优化调度科学等业务作业模式的综合需求,研究并提出了水资源管理驱动下的分布式系统架构。(2)针对水资源管理领域数据所具有的海量多源异构特性,提出了多类型水信息数据的分区分库存储方法,建立了基于多存储构件的分布式数据存储平台,为分布式网络环境下的数据访问、传输及管理提供统一数据支持。同时,为满足分布式事务对复杂软件系统的一致性、可用性、容错性要求,结合水信息领域业务模型和数据特征,提出了适用于分布式软件环境的模型和数据集成方法。模型集成包括组件式集成和服务式集成,数据集成涉及数据融合、数据建模以及模型的应用与实践。(3)针对水资源管理分布式系统中多业务、多模型串并行作业特征,提出了多模型串并行耦合作业模式。结合水资源领域分布式栅格新安江模型、水动力-水质耦合模型、复杂紧急避险模型、空间分析模型的计算作业特点,实现了多元模型在分布式系统中的高效耦合与集成。(4)最后,基于SOA概念与Web Service服务技术,综合上述分布式水信息系统架构、数据存储平台、多模型集成方法及串并行作业调度模式,结合实际工程应用需求,进行了分布式水信息系统的设计、开发与集成的研究和实践,并在有关工程中获得应用,研究成果为水资源管理多模型驱动下的分布式系统集成开发和应用提供了理论基础与技术支撑。更多还原

【Abstract】 Water is critical resource for human survival,therefore,how to achieve efficient and proficient water resource management is the major studying area for scholars and specialist in the water industry.Facing the deteriorating water crisis,more and more complex problems occur in the context of water resources management,this is especially urgent,because data explosion make the hydroinformatics to be a critical and complex managing environment.Under the circumstances,new and provoking managing requirements are proposed for developing hydroinformatics software systems,such as,using new technology,designing big platform,creating sharing function,expanding multiple applications.However,the traditional software designing system that emphasizing facing process,objectives,structures cannot satisfy the updated requirements for development of new systems.Therefore,in the area of hydroinformatics,it is urgent to develop new framework and technology to match the requirements proposed by updating software development.By doing these,we hope to fix the following managerial problems during the software developing:sharing of information and resources,coupling of multiple services,optimal decision by searching multiple objectives.In addition to above-mentioned situation,as the Web Service technique developed,distributed software system designing which takes the Web Service as its core technique is significantly developing and playing an important role in the area of finance,electric commercial,marketing,industrial management,telecom and education.However,in the area of hydroinformatics,the applying of distributed software system is only at the initial stage,the relevant studies is inadequate.Given that various hydro-junctions and the corresponding managerial departments distribute in different locations,and the basic principles of water resources management are business servitization and information sharing,it can be easily assumed that the applying distributed software designing to the hydroinformatics studies could be pretty promising and useful.Considering that studying water resource management heavily relies on the geographic information technology,it is necessary to couple multiple business and integrate different modules.Based on this,the present study proposed a distributed hydroinformatics system designing framework hoping to provide general application of distributed software designing in the area.Moreover,this thesis introduced following content in details:the relevant theory for the framework or system,the method for communication,the mechanism of functioning,and the patterns of system integration.After the accomplishing the work,we hope to make several contributions and provide some novel understandings as following.(1)This study thoroughly discussed the conceptual bottom layer of the distributed framework-SOA and the communicational method of Web Service mid-ware.Based on this,we proposed the distributed software framework for regular patterns and irregular patterns of water resource management.Building on the conceptual frameworks,technique topic and meta-system for the software and combining the business functioning patterns for relevant hydrological,hydrodynamic and water regulating studies,the present thesis proposed a distributed software framework driven by water resource management issues.(2)Dealing with massive heterogeneous data is a major problems for successful water resource management,as a result,we build a distributed data storing platform to provide technical support for storage,remote access,transferring,and managing of high quality,multi-component massive data.In line with the distributed system designing principles,consistency,utility,fault tolerance,we put the stability of the functioning system as first priority for our system designing.In addition to that,we also take the various attributes of models and data into consideration,based on this,we proposed an adequate data integration and model integration which are applicable for distributed software designing environment.Model integration can be classified as two types,component-oriented integration and service-oriented integration.Data integration also includes two parts,data fusion and practical content of data modeling.(3)Based on that the multi-models work under serial connected patterns,we proposed "Coupling and serial running patterns for multi-models".After comprehensive consideration of the consumption of time,money and resources for running the jobs and demand of the users,we classified the running patterns of the models into two categories:one-step serial job dispatching and multi-steps asynchronous job dispatching.In turn,by considering the attributes of Xinan jiang grid model and coupling of hydrodynamic model and water quality model,we finally proposed the coupling integrating method of above mentioned models in our system.In the distributed system,a requirement for parallel job running will be raised driven by multi-models.Opposing to the serial job running,we also proposed the "Coupling and parallel running patterns for multi-models".Combining the analytical requirements for emergent evacuation during the flood,we achieved the parallel and coupling running of emergent evacuation models and spatial analytical models in the distributed system.(4)To sum up our works,the whole designing process of this distributed systems involve following parts:usage of SOA concept,Web service techniques,distributed hydroinformatic system,distributed data storage platform,distributed job dispatching.The system designing and developing nicely satisfy the requirements arise in the practical project,which would provide technical support for designing and developing distributed software system that serves for practical water resource management project.更多还原