基于案例推理的大停电恢复系统设计

设计了一个应用于大停电事故恢复的基于案例推理系统。系统纵向采用分层结构,抽象层体现恢复过程的一般策略,具体层案例包含恢复操作的具体措施。系统横向分成黑启动模块、网架和负荷恢复功能模块。黑启动模块中的抽象层用一个定性进程理论模型表示,模型考虑了黑启动中的主要限制条件和各类机组的启动特性,发电容量调度算法用于具体层的修正。网架、负荷恢复模块的抽象层和具体层由详细程度不同的系统潮流图表示。开发了串并行送电阶段优化算法、系统网架及负荷恢复优化算法,用于具体层的修正。该系统以基于案例推理作为主框架,把定性分析和定量分析、理论研究和电力部门的实际经验很好地结合起来,多种在线优化算法提高了系统的灵活性。     

基于多数据源的电力变压器分层故障诊断系统设计

在现有变压器在线监测条件普遍不足的前提下,介绍了一种基于多数据源的变压器分层故障诊断系统。该系统充分利用在线监测数据和各种电气试验、运行状况以及专家经验等信息,将人工神经网络和基于事例(case)推理的专家系统结合起来,在不同层次的故障诊断模型中采用不同的智能诊断手段,逐步深入、细化诊断结果,并对系统的数据流和知识流进行合理的总体规划。该系统还提供了模拟变电站数据的仿真器,可用于系统的测试,在一定程度上保证了其可靠性。     

基于分层案例推理的黑启动决策支持系统

分析了电力系统恢复的层次特征和基于案例推理法的黑启动任务决策系统中存在的问题,提出将分层案例推理的思想应用于黑启动调度任务决策支持。文中给出了分层案例库的框架结构及分层黑启动案例的推理算法,并结合山东电网实际,设计并实现了黑启动决策支持系统,通过问题分解、分层案例推理和子案例复用提高了系统的应变能力,其分层式图形交互案例平台清晰地表达了多层次的黑启动决策过程,能够起到调度员培训和在线决策支持的作用。     

A restoration aid expert system for distribution substations

This paper presents a practical restoration aid expert system for 154 kV distribution substations. The proposed expert system has a topology-based simple structure and utilizes basic rules suggested by the Korean electric power company including heuristic rules, topology identification rules and search rules to generate switching sequences for feasible restoration processes. In this paper, the general system structure as well as rules and data representations are discussed. The expert system is realized using an HP9000 workstation. X-window and Motif shells are used for the user-friendly GUI. The system has been tested in a practical sub-control center as a part of an intelligent fault diagnosis and restoration aid system to assist the decision making of SCADA operators     

A flexible graphical user interface for embedding heterogeneousneural network simulators

A graphical user interface (GUI) for heterogeneous neural network simulators proposed in this article is intended to be of use both for the novice and for the experienced neural network user. For the novice, it provides an easy-to-use neural network simulation package that insulates the user from the requirements of knowing the simulator implementation details or the configuration file syntax. For the experienced neural network professional it provides an interface that is easily extensible to include any additional neural network simulator in its binary form. To satisfy both academic and personal computer environments, the GUI has been developed by using the free TCL/TK software package, available on workstations running Unix and on PCs running the free Linux operating system. Although the GUI and the embedded simulators have been successfully tested both in neural network research and training programs, a more extensive testing in undergraduate and graduate level classes is in progress     

基于网络分析和智能推理的调度操作票专家系统

在已有通用认知模型的基础上将操作票系统与EMS相结合,将网络分析引入专家系统中,进一步延伸了通用认知模型的范畴,使拓扑分析功能与通用认知模型中的操作对象概念相结合,加强了专家规则的通用性和表达能力。同时在推理过程中引入了潮流分析,能辅助调度员根据电网潮流分析结果进行决策,从而生成安全性、经济性最优的调度指令。所开发的系统已在2个实际地调投入现场运行,取得了良好的效果。     

Application of expert system to power system restoration insub-control center

An expert system is developed for emergency operation of power system restoration in a sub-control center. The proposed method restores the blackout area in a power system by using sensitivity analysis and heuristic rules. The rules are categorized into flow-control rules, energization rules, line switching rules, load shedding rules, and voltage correction rules. The switching (opening, closing) of transmission lines, transformers and shunt capacitors, and the load shedding are used for the emergency operations in this paper. For the real time recognition of the network topology, the power system components are represented in hierarchical data structure     

A case-based construction method for a fault-restoration support system in secondary power systems

The authors have already clarified that case-based reasoning (CBR) is very useful for the fault-restoration support system in electrical power networks. This paper describes the case-based construction method, adaptation and modification method of a case which is a fundamental process of CBR, and the results of evaluation. The following points were clarified through the simulations of fault restoration under various conditions with electrical power networks of practical scale: (1) a case-based construction method and the guide to the selection of cases which should be included in the case were made clear. Restoration plans which include many cases are registered in the base. Cases which would require long computation time are also registered in the case base. In the application which requires high-speed processing, such as fault restoration, there exists an optimum value in the total numbers of cases in the case base; (2) the function of load switching which is essential to perform adaptation and modification of a case flexibly was studied and embedded in the system herein. Inference power was augmented by the implemented of this function, and optimum for suboptimum solutions were obtained. Processing time with CBR is considerably shorter than other methods, and this feature is remarkable with complicated cases.     

Reinforcement learning in case-based systems

Case-based reasoning (CBR) systems compare a new problem to a library of cases and adapt a similar library case to the problem, producing a preliminary solution. Since CBR systems require only a record of library cases with successful solutions, they are often used in areas lacking a strong theoretical domain model, such as medicine, economics and law. The problem is that many CBR systems use expert knowledge to determine how to build indexes for the case library so that cases that match the current situation can be identified. Human experts - whose time is valuable and scarce - often find it difficult to explain precisely their reasoning. Thus, a knowledge elicitation bottleneck occurs for many knowledge-based applications. Reinforcement-trained case-based reasoning (RETCBR) uses feedback from the user or some external process to learn how to match cases. RETCBR expands the domains in which CBR techniques can be applied, because it requires knowledge only for case recognition, not for determining indexing strategies. We are currently applying these techniques to human-computer interaction problems, such as user modeling and collaborative filtering.     

基于多智能体的全面防误策略研究及智能操作票生成系统实现

为满足值班调度员对于实时生成操作票的要求,通过对国内十几种主要接线方式进行动态接线技术分析,提出了一种全面的基于多智能体(Multi-agent System,MAS)技术的防误策略,并基于模糊认知图(Fuzzy Cognitive Map,FCM) 的智能体推理模型研究出一种实时操作票生成专家系统。该系统在注重智能性、通用性的前提下,充分考虑了可靠性、可维护性及人机界面友好等实用化要求。     

PowerGraf: an educational software package for power systemsanalysis and design

The purpose of PowerGraf is to present a graphical user interface (GUI) and a convenient tool for building one-line diagrams displaying data. The GUI is open and friendly for power systems design, analysis and control, because a picture provides a better illustration of cause and effect relationships than a tabular printout does. Basic elements such as buses, lines and transformers are given such that the power system topology is easily drawn with this tool, and the system parameters are entered on the picture and checked visually. This enables the power system topology and the system data to be shown on one picture page. Advantages of these graphics are that users can view overloads, low voltage problems, and outages on overall (total) system. In addition, it is possible to see the impact of contingencies on the system. To date, PowerGraf 3.0 together with an instructional demo has been used to teach power systems courses in the Electrical Engineering Department at the University of Missouri-Rolla (UM-Rolla). It is also available on the Computer Learning Center network to help students design their own optimal power systems     

新型智能电网调度操作票自动生成与管理系统

针对现有调度操作票自动生成系统大多为离线的单一智能体系统,在智能化、安全校核、流程化审核执行等方面需要提高的现状,提出建立操作规则模板及用户可维护机制解决操作规则描述、学习及推理的通用化问题,并提出利用多智能体系统,开发和实现了与SCADA系统共享实时数据、与调度管理信息系统集成的集自动开票、手工开票、模拟预演、流程化审核、执行与管理等多种实用功能于一体以及用户可维护的新型智能化电网调度操作票系统.研究成果已经在某省电网投入使用,实践证明能大大提高电网调度操作的智能性和安全性,并且在通用性和用户可维护性方面显出优势.     

基于事例和规则混合推理的变电站故障诊断系统

故障诊断是保证电力系统安全运行的重要手段,目前多采用基于规则推理的专家系统,但其知识获取困难,自学习能力差,很难适应电网发展的要求.文中介绍了基于事例推理(CBR)和基于规则推理(RBR)的混合推理的变电站故障诊断专家系统.该系统采用CBR方法确定故障情况,再利用规则评价继电保护和开关的动作情况.事例库包括用RBR系统自动生成的基本事例库及无确定规则的特殊事例,在使用中还可不断地增加新事例以提高系统判断复杂故障的能力.由于采用了混合推理,系统故障诊断快速可靠,动作评价准确,自学习能力也得到很大提高,可减轻运行人员的工作量.     

A Windows-based interactive and graphic package for the educationand training of power system analysis and operation

This paper presents a Windows-based interactive graphic package developed by the authors for the education and training of the power system modeling, analysis and operation with a user-friendly graphical user interface (GUI) and visual database management system (VDBMS). To achieve this goal, the package is designed to have following features: a graphic editor to visually edit the power system diagram; Windows-based power system database to interactively manage the data; graphic representations; animation; and interrupting schemes. The application programs in this package include the power flow calculation, transient stability analysis, fault analysis, economic dispatch and automatic load-frequency control. This package may be useful for educational and training purposes     

基于开关逻辑的电力系统图形和操作票自动生成系统

提出了一种用于开关逻辑表示电力系统的网络拓扑结构,系统运行状态和设备操作的方法,并将该方法应用于电力系统主接线图的绘制和调度操作票的自动生成,开发出一种基于开关逻辑的电力系统图形和操作票自动生成系统。该系统具有图形生成直观,知识表示统一,规则修改容易,推理过程简单和通用性强等特点。     

Expert system for voltage control in restoration of EHV power system following complete blackout

This paper considers voltage control in restoration of EHV power systems. Restoration of a power system is performed by energizing transmission lines sequentially from initial power sources. These operations have already been programmed into an expert system developed by the authors. However, energizing no-load transmission lines causes some bus voltages to rise 20 percent more than rated voltages, and transformers can tolerate a 20 percent voltage rise only for 1 min. Here, the expert system is supplemented with a function for checking bus voltages and reactive powers of initial sources, and which automatically switches on reactors. Conditions for switching reactors and selecting appropriate reactors are written by rules, hence they are easy to modify. The target of voltage control is to keep bus voltages with in ±5 percent, and to keep reactive powers of generators within 30 percent of generator capacity. Applying this method to three basic restoration patterns of a model power system shows that small adjustment of switching conditions is sufficient for two patterns, and some modification of the restoration procedure is necessary for the other pattern to obtain good voltage profiles through the restoration process.     

Group decision support system for backbone-network reconfiguration

Modern power system is unprecedentedly vulnerable and the society is exposed to higher blackout risks. Therefore restoration is a key issue to power systems. Reasonable backbone-network reconfiguration is necessary for re-establishing the network and restoring loads quickly. In order to speed up network reconfiguration under security constraints, a group decision support system (GDSS) is developed. A unified data platform based on data warehouse is constructed to resolve inconsistencies of data, information and models. A three-stage restoration strategy and a two-layer restoration path searching method are proposed to establish restoration schemes. Multiple attribute decision-making is implemented to evaluate alternatives and consider various attributes about restoration speed and system security comprehensively. Group decision-making provides an aggregated cardinal ranking of candidate restoration schemes. The GDSS uses expert rules for qualitative analysis and numerical computation programs for quantitative analysis, and it can resolve the semi-structured decision problem of network reconfiguration efficiently. Restoration plan of Shandong power grid is developed with the proposed method and it has been adopted by State Grid Shandong Electric Power Company. Performance results of Shandong power grid of China demonstrate the feasibility and flexibility of the method.     

Application of genetic algorithms to service restoration in distribution systems

Service restoration in distribution systems can be formulated as a combinatorial optimization problem to determine power sources for each load considering radial network constraints and power source limits. Until now, the problem has been considered using conventional methods, e.g., the branch-and-bounds method, expert system, neural networks, and fuzzy reasoning. Recently, Genetic Algorithms (GA) have been recognized as one of the efficient methods for solving large combinatorial optimization problems. The method can perform parallel search, and can more easily search optimal solutions. This paper presents an application of GA to service restoration in distribution systems. The feasibility of the proposed method is demonstrated on a typical distribution system model. The result shows that the method can solve the problem efficiently, and this tendency becomes dominant by increasing problem dimensions.     

配电网故障恢复系统

将配电网的故障恢复模式分为早期模式、DA模式和DMS模式3种,分析和比较了DMS与DA模式的功能和特点,认为DMS模式将获得更广泛的应用。介绍了DMS故障恢复系统的数据库设计,通过引入区段的概念,将原始的网络数据库中的结点与元件关联表按层次划分为结点与区段、区段与元件2类关联表,为故障恢复系统的研制与开发打下了基础;说明了故障恢复的推理模型和推理方法,通过把故障恢复问题分解成故障检测、状态分析、方案形成、方案选择、操作序列等若干子问题,使问题求解过程易于处理和控制;最后介绍了故障恢复系统的软件实现。     

配电网故障恢复专家系统的一种实现

介绍了一种用于配电网故障后恢复供电的专家系统及其软件。根据DMS系统采集的故障类型、网络状态、潮流等信息,启动依据规则的专家系统进行快速故障恢复。介绍了该系统的基本原理、算法、结构、推理机制及其软件实现方式等。