【作者】 焦晓佑；

【导师】 胡兆光； 周渝慧；

【作者基本信息】 北京交通大学 ， 电力系统及其自动化， 2009， 博士

【摘要】 本文以电力供需平衡复杂性问题为切入点展开了研究工作。由于经济增长与电源建设等因素制约着电力需求与电力供应两者之间的平衡关系,从需求端来看,它的表现之一是宏观经济政策对电力消费影响的复杂性;从供应端来看,它的表现之一是由于随机性和模糊性导致的不确定发电规划复杂性。这两类复杂性问题具有动态性、模糊性和不确定性等特点,依靠传统的数学理论和方法来建立定量的、精确的数学模型已显露出它的局限。上述两类复杂性问题均属于典型或复杂的智能工程B₂问题,因此,本文以电力供需平衡复杂性问题为研究对象,对智能工程第二类问题B₂的求解理论和优化理论进行了研究,扩展了复杂问题求解的智能工程理论框架和方法体系,分析了多层次的智能建模结构,建立了Rule-based Model的问题B₂求解模型,研究了复杂过程的智能推理和路径寻优方法,并给出了智能工程问题B₂求解理论和优化理论在上述两类供需平衡复杂性问题分析中的具体应用。一、理论方面,本文用系统的观点对智能工程进行了分析,论述了从系统工程向智能工程扩展的四元结构体系;给出了问题B₂求解的一般模型描述,从“问题的分解”和“解的合并”两个层次对问题B₂求解进行了过程分析,给出了映射函数f的合成原则与方法;基于α-优越解和β-优越解研究了智能工程问题B₂的优化理论,提出了经过α-优化过程和β-优化状态两个步骤来实现对目标状态解集的优化方法,给出了优化问题B₂的目标解定理、最优解定理和比较定理,并对相关定理进行了证明。二、方法方面,本文给出了求解问题B₂的知识表达方法,针对粗糙集理论应用于知识发现时的不足,在前人的研究基础上给出了一种模糊粗糙知识约简算法,形成了以历史数据为基础的知识建模和规则挖掘方法,通过对挖掘得到的各类规则进行完备性和相容性分析,证明了其有效性;以这些知识规则为基础,设计了一种分段定值的改进型变论域智能推理模型,模拟仿真结果证明了该模型相比于经典的模糊推理模型具有更高的推理精度。三、建模方面,本文基于智能工程从横向和纵向两个角度对电力供需复杂性问题的建模方法进行了研究,通过对复杂性问题的本质信息进行不同层次的知识描述,探讨了多层次的智能建模方法;针对复杂系统要素之间的协调问题,研究了分布式的智能建模方法;建立了复杂性问题分析的智能推理模型Rule-basedModel,给出了其输入-输出模型的具体数学形式阐述,通过算例说明了Rule-basedModel的在复杂性问题建模中的适用性。四、应用方面,本文分别给出了智能工程问题B₂求解理论和优化理论在电力供需平衡复杂性问题分析中的应用:基于问题B₂求解理论,按照多层次、分布式建模思想构建了一个具有智能推理功能的电力经济关系动态分析模型REEDAM,研究了电力供需平衡动态过程的智能推理和模拟仿真,实现了对宏观政策影响电力消费复杂性问题的求解;基于问题B₂优化理论,提出了基于路径和状态的智能规划方法,针对随机性和不确定性造成系统过程很难做到使初态在某个时刻精确地成为某个终态,通过α-优化过程和β-优化状态两个步骤来对目标状态解集进行寻优,给出了不确定规划优化模型,实现了对风力发电不确定规划的优化,并进行了算例分析。更多还原

【Abstract】 Power system is a typical and especial complex system.Besides the energy conversion and power technology,the complexity analysis also includes the balance relationship between power supply and demand:how the macropolicy impacts the balance of power supply and demand in electricity-economic system? how to optimize the distributed generation planning for the uncertain factors such as the randomicity and fuzziness of wind speed? Because of these multi-dimension dynamical and some uncertain characteristic,the classical theories and precise models have become unfit to these situations.So this dissertation focuses on the research of the theory and methodology of Intelligent Engineering Programming and Optimization about problem B₂,Based on the idea of Theory-Methodology-Model-Application,the main contents are shown as follows:1.About theories,this paper analyzes the Intelligent Engineering from the system point of view,attempts to do the research work from System Engineering to Intelligent Engineering,presents a intelligent reasoning model based on rules in programming theories,and researches the optimization theories and methods about problem B₂ based onα-optimized andβ-optimized solution.Describes the Intelligent Programming four-part frame system,presents the synthesize principle and method of mapping function f,presents and proves the aim result theorem, optimum result theorem and comparison theorem of problem B₂.2.About methodologies,this paper presents the state space and knowledge expression method in Intelligent Engineering Programming Theoryies,gives a fuzzy-rough reduction arithmetic,builds a knowledge modeling and rule mining method,and analyzes the completeness and compatibility of rules.Based on these knowledge rules,this paper designs an intelligent reasoning method based on Intelligent Engineering and Variable Universe,the simulation results showed its well effect.3.About modeling,this paper proposes a complex system modeling method in terms of horizontal and vertical orientation:designs a multi-level intelligent modeling method by multi-level essential information knowledge description,and designs a distributed intelligent modeling method aimed at the complex harmony question. An intelligent Rule-based Model is built and its input-output mathematics form is expatiated.Furthermore,a case studies showed this modeling method is good in complex system modeling.4.About applications,on one hand,this paper builds an intelligent system model REEDAM（Rule-based Electricity-Economic Dynamic Analysis Model） based on Intelligent Engineering Programming Theory,and the dynamic course of balance between power supply and demand is simulated by this model.On the other hand, this paper presents an intelligent planning method based on Intelligent Engineering Optimization Theory,aimed at the uncertainty in distributed generation planning, this paper researches the intelligent planning method from two hiberarchy onα-optimized solution of path andβ-optimized solution of state.Case study shows the validity of this method.更多还原