【作者】 刘晓东；

【导师】 杨宝清；

【作者基本信息】 大连交通大学 ， 机械制造及其自动化， 2012， 博士

【摘要】 由于智能测控系统是涉及多学科、多专业领域,集测量、控制、管理等于一体的多功能、多任务的复杂系统,系统设计至关重要。面对多样化的需求环境和不断丰富的软硬件资源,目前越来越多地形成了针对不同应用的特定智能测控系统,这在一定程度上使智能测控系统通用结构研究在描述和求解时遇到了诸多困难。另一方面,现有的智能测控系统性能评价理论尚未成熟,其评价方案存在着较大的局限性。在这种背景下,本文对通用智能测控系统逻辑结构建模、系统智能性、评价指标体系建立和系统综合性能评价等方面进行了理论研究和实践探索,以期对智能测控系统可持续研究与发展提供依据,主要研究内容和创新点总结如下：首先研究了智能测控系统的内涵。智能测控系统的应用广泛性、设计复杂性、需求多样性等特点决定了智能测控系统结构构成的多样性和不确定性,论文从智能测控系统的本质特征出发,研究智能测控系统的功能,抽取一般规律,探讨了一种合理的广义智能测控系统,为智能测控系统结构构成与性能评价等深入研究奠定基础。提出了基于拟人理论的USR逻辑结构模型并应用到智能测控系统的建模之中。在全面考虑拟人理论特点和现阶段典型智能测控系统实例的基础上,利用拟人理论建立智能测控系统通用逻辑结构USR模型,给出了系统USR模型的定义,分析讨论了结构单元集U、服务集S、关系集R等模型三要素的内涵,并对USR模型的优点及所解决的问题进行了分析。USR模型将智能测控系统看作一个由各个部分构成的有机整体,从系统结构方面和运行机理上进行形式化描述,可作为系统深层次分析与评价的起点。研究了智能测控系统的智能性。智能是智能测控系统与一般测控系统的最大区别,论文剖析了智能测控系统的智能含义,从技术层面上阐述了智能化方法在感知系统、分析处理系统、知识系统及表达系统等结构单元中的体现典型应用,并以智能弧压调高系统为应用实例,具体分析了模糊控制智能算法在应用中的优越性。针对系统智能水平的评价问题,采用拟人的智商算法进行计算,探讨基于改进离差智商算法进行智能测控系统智能水平评价的方法,为深入理解智能特性提供了理论依据和技术手段。构建了智能测控系统综合性能评价指标体系。指标体系按照系统科学性、可操作性、定量指标与定性指标相结合的原则,以系统逻辑结构划分为基础,以智能测控系统综合性能为总目标,共包含6个决策指标和15个基础指标,该指标体系能够较为全面地反映智能测控系统的综合性能状况。论文阐述了各评价指标的测算方法并利用层次分析方法确定了各指标权重。提出了一种利用证据推理理论评价智能测控系统综合性能的方法。由于智能测控系统的性能评价问题属于复杂的多目标决策问题,基于智能测控系统具有评价信息多源性、参数不确定性和模糊性等特点,论文选取综合证据理论、效用理论、模糊理论等优点的证据理论ERA(Evidential Reasoning Approach)评价方法实现对智能测控系统性能的评价,该方法能够有效处理智能测控系统中含有不知道和不确定参数信息的情况。论文分析了评价算法,并建立了智能测控系统综合性能评价ERA模型,而且针对评价信息的多源性给出了相应的信息融合处理方法。最后,在上述逻辑结构分析与综合性能评价方法的指导下,应用两个典型智能测控系统实例分别进行研析和理论验证。选取油井作业智能监测系统实例进行逻辑结构分析,利用拟人理论建立了油井作业监测系统逻辑结构USR模型,验证了基于拟人理论建立的逻辑结构模型的有效性。同时以电脑鼠迷宫寻迹系统为背景,以本团队设计的三种电脑鼠迷宫寻迹系统方案为对象建立了证据理论ERA评价模型,进行了系统性能综合评价。研究表明,评价结果比较合理,证据理论ERA评价方法能够有效解决智能测控系统性能综合评价问题,并对相关领域以及智能测控系统进一步研究具有一定的意义。更多还原

【Abstract】 As a large complex system with multi-function and multi-task, the intelligent measurement and control system is involved in a number of disciplines and professional fields, so the system design is very important. With the diversified requirements and the increasingly enriched software and hardware resources, more and more specific intelligent measurement and control systems have been formed for different applications, which makes the related theory research of general structure about the intelligent measurement and control system very difficult. On the other hand, the exiting performance evaluation theory of intelligent measurement and control system is still not mature, and the exiting evaluation solutions have some limitations. In this background, this paper studies the theory and practices exploration of the logical structure modeling, the intelligent performance, the establishment of evaluation index system and the comprehensive performance evaluation of intelligent measurement and control system, in order to provide basis for sustainable research and development of system. The main research contents and abstracted innovative points are as follows:Firstly, the paper has discussed the connotation of the intelligent measurement and control system. Its universal applications, complex designs, diverse demands and such characteristics have decided the diversity and uncertainty of system structure composition. From the essential characteristics, this paper studies the system function, extracts the general rules, and expounds a kind of reasonable definition for generalized intelligent measurement and control system, which lays a foundation for the profound research of its structure composition and performance evaluation.This paper proposes a general logical structure USR model of intelligent measurement and control system based on the humanized theory. Considering the characteristics of humanized theory and the present typically examples of the intelligent measurement and control system, this paper establishes the USR model by the humanized theory, gives the definition of USR system model, and discusses the connotation of the three model elements including structural unit set U, service set S, relationship set R, etc. The merits and the problems to be solved by this model are analyzed. It takes the intelligent measurement and control system as an organic integrity constituted by each part, and makes the formalization descriptions by the system structure and operation mechanism, which can be used as a starting point of the high-level analysis and evaluation of system.This paper studies the intelligence characteristic. An intelligent measurement and control system can be distinguished from a general measurement and control system by the intelligence characteristic. This paper analyzes the intelligence connotation and expounds the typical applications of the intelligent methods to structure units such as the sense unit, the analysis processing unit, the knowledge system and the expression unit, etc. With the intelligent arc pressure adjustable high system as an application example, this paper explains the application of intelligent fuzzy control algorithm. With respect to the evaluation of system intelligent level, this paper carries on the humanized algorithm, and discusses the intelligent level evaluation method for intelligent measurement and control system based on the improved deviation intelligence quotient algorithm, which provides the theory basis and technology method for thoroughly understanding of intelligent characteristic.This paper builds up a new comprehensive performance evaluation index system for the intelligent measurement and control system. According to the principle of system scientificity, operationality, combination of quantitative index and qualitative index, this index system contains6decision indexes and15basic indexes on the basis of division of system logical structure. The index system can fully reflect its comprehensive performance. Then this paper expounds the calculation methods of each evaluation index and determines the weight of each index by AHP method.This paper proposes a system performance evaluation method by the D-S evidence reasoning theory. This performance evaluation of intelligent measurement and control system belongs to a complex multi-objective decision problem. Based on the characteristics of system such as assessment information polyphyletism, parameter uncertainty and parameter fuzziness etc, this paper takes ERA theory to evaluate, which has the advantage of D-S theory, utility theory, and fuzzy theory etc. This evaluation method can effectively deal with the problem with the unknown and uncertain parameter information of intelligent measurement and control system. This paper analyses the evaluation algorithm, and establishes the ERA comprehensive performance evaluation model. The response information fusion processing method is given with respect to the assessment information polyphyletism.Finally, in the direction of above logic structure analysis and comprehensive performance evaluation method, this paper selects two typical examples of the intelligent measurement and control system for the analysis and theoretical verification. The well operation intelligent monitoring system is selected as an example to analyze the logic structure, and the logic structure USR model is established by the humanized theory, which verifies the effectiveness of the logic structure model based on humanized theory. At the same time, taking computer mouse labyrinth tracing system as the background, this paper selects three kinds of design plans designed by our team as examples to establish the evidence theory ERA evaluation model and finish system comprehensive evaluation. Research shows that the evaluation result is reasonable, and this ERA evaluation method can effectively solve the problem of the performance comprehensive evaluation of intelligent measurement and control system, which has certain significance for the further research of intelligent measurement and control system and related fields.更多还原