【正文】 有功損耗和無功損耗得到了下降，從而使線損降低，優(yōu)化了該地區(qū)的網(wǎng)絡(luò)結(jié)構(gòu)。第5章 結(jié)論 本文結(jié)合了配電網(wǎng)重構(gòu)的思想，研究了目標(biāo)配電網(wǎng)優(yōu)化，確定將以線路損耗的目標(biāo)優(yōu)化作為配電網(wǎng)優(yōu)化的目標(biāo)函數(shù)。通過對(duì)配電網(wǎng)優(yōu)化算法的研究，本文采用遺傳算法作為配網(wǎng)優(yōu)化的算法。同時(shí)本文結(jié)合配電網(wǎng)絡(luò)拓?fù)浣Y(jié)構(gòu)特征和遺傳算法特點(diǎn)具體探討了如何實(shí)現(xiàn)遺傳算法的網(wǎng)絡(luò)重構(gòu)，基于配電網(wǎng)絡(luò)拓?fù)涞母倪M(jìn)方法，提高了算法的收斂速度和收斂性。 算例結(jié)果表明，該算法通過對(duì)城市中壓配電網(wǎng)進(jìn)行優(yōu)化重構(gòu)，能夠降低系統(tǒng)的網(wǎng)損。參考文獻(xiàn)[l]王守相，張鵬，兼顧供電可靠性和經(jīng)濟(jì)性的多目標(biāo)網(wǎng)絡(luò)重構(gòu)，繼電器，2005，33(7)，1620.[2]陳禮義等，1999，11(4)，6165.[3]劉健，畢鵬翔，董海鵬，復(fù)雜配電網(wǎng)簡化分析與優(yōu)化，北京：中國電力出版社，2002.[4]徐其迎，李日隆，陳樹挺，一種基于VB編程的配電網(wǎng)可靠性評(píng)估算法[J]，電網(wǎng)技術(shù)，2004，28(3)，4850.[5]畢鵬翔，復(fù)雜配電網(wǎng)簡化分析及配電網(wǎng)重構(gòu)，西安交通大學(xué)博士學(xué)位論文[M}，2002.[6]Felix . Chan， . Chung Multicriterion genetic optimization for due date assigned distribution network problems. Decision Support Systems，2005(39)，661 675.[7]王峻峰，周家啟，謝開貴，中壓配電網(wǎng)可靠性的模糊評(píng)估，重慶大學(xué)學(xué)報(bào)(自然科學(xué)版)，2006，29(2)，4549.[8]Billin R，Allan R N，Reliability Evaluation of Power System. Boston， Pistman，1984.[9]別朝紅，王錫凡，配電系統(tǒng)的可靠性分析，中國電力，1997，30(5)，1013.[10]張焰，電網(wǎng)規(guī)劃中的模糊可靠性評(píng)估方法，中國電機(jī)工程學(xué)報(bào)，2000，20(11)，7780.[l1]劉莉，陳學(xué)允，基于模糊遺傳算法的配電網(wǎng)絡(luò)重構(gòu)[J]，中國電機(jī)工程學(xué)報(bào)，2000，20(2)，6669.致 謝通過這一階段的努力，我的論文終于完成了，我的論文是在武曉冬老師的親切關(guān)懷和細(xì)心指導(dǎo)下完成的。非常感謝武曉冬老師在我大學(xué)的最后學(xué)習(xí)階段—畢業(yè)設(shè)計(jì)階段給我的指導(dǎo)，從最初的論文定題，到資料收集，到寫作、修改，到論文定稿，他給了我耐心的指導(dǎo)和無私的幫助。我不是您最出色的學(xué)生，但您是我最尊敬的老師。武老師認(rèn)真的工作態(tài)度，誠信寬厚的為人處世態(tài)度，給我留下了，深刻的印象。同時(shí)我還要感謝在我學(xué)習(xí)期間給我極大關(guān)心和支持的各位老師以及關(guān)心我的同學(xué)和朋友。 由于我的水平有限，所寫論文難免有不足之處，懇請(qǐng)各位老師批評(píng)和指正。 李淑芳 英文原文Cost/worth assessment of reliability improvement in distribution networks by means of artificial intelligenceAbstractA major challenge for the power utilities today is to ensure a high level of reliability of supply to customers. Two main factors determine the feasibility of a project that improves the reliability of supply: the project cost (investment and operational) and the benefits that result from the implementation of the project. This paper examines the implementation of an Artificial Intelligence System in an urban distribution network, capable to locate and isolate short circuit faults in the feeder, thus acplishing immediate restoration of electric supply to the customers. The paper describes the benefits of the project, which are supply reliability improvement and distribution network loss reduction through network reconfigurations. By parison of the project benefits and costs the economic feasibility of such a project for an underground distribution feeder in Greece is demonstrated.Keywords:power restoration 。 reliability improvement 。loss reduction1. Introduction Distribution system protection and restoration along with loss reduction are some of the most important topics concerning contemporary power engineering research. Two main parameters lead power utilities nowadays to redefine their policies regarding investments on power networks. The first parameter is the demand for increased reliability considering the fact that the way of living today (both at personal and professional level) tends to bee fully dependent on electricity. On the other hand the energy problem and its consequences to the environmental pollution render the need for energy saving more imperative than ever. It is estimated that the largest proportion of losses in power networks corresponds to distribution networks for a typical system in a developing country, distribution losses account for approximately 8% of the total electrical energy produced [1]. Considerable research has been acplished so far for systems and methods that contribute in loss reduction across distribution networks and reliability improvement.Automation has been applied to the distribution network in order to achieve significant service reliability improvement for electricity customers [2–4]. Other approaches investigate reliability improvement and interruption cost minimization based on appropriate switch location or relocation across a distribution feeder [5,6]. Finally, significant research has been conducted on loss reduction in distribution systems via network reconfiguration. These applications are based on the development of algorithms for switching operations utilizing heuristic, fuzzy logic and other approaches [7–9]. The undertaking of investments in such novel systems from the utility’s perspective constitutes a plex procedure which depends on many factors. The investment decisions are based on project economic feasibility studies through which the most beneficial investment alternative is determined. For an investment to beeconomically viable, utility’s cost to improve reliability should be less that the customer’s cost. In this paper, the implementation of a Multiagent System (MAS) [10–12] in urban distribution networks is examined as an investment proposal. The MAS is capable of locating and isolating sustained faults to the shortest possible segment of the feeder, and of achieving restoration of supply to the maximum possible number of consumers (within approximately min). Additionally, the system deals with the loss reduction problem by transferring loads to adjacent buses. The problem of fault detection and power restoration is a multiobjective dynamic binatorial problem with topology constraints [13]. In practice considering that the plexity of the problem has been classified as NPplete it bees obvious that for real time fault restoration the problem cannot be solved exactly. The proposed MAS on the other hand implements heuristic approaches in order to deal with the problem, thus it bees suitablefor online implementation. Modifying the implementation approach of the proposed MAS, either a centralized management of the network or local supervision and control of adjacent topologies,the required time for optimal or nearoptimal solution could also be improved. Furthermore, once the ontology of the software agents is constructed, additional functionalities for the MAS, such as network reconfiguration for loss reduction or real time ma