外文翻譯--設計實現電網知識的可視化

1、<p>　　中文3000字,1690單詞，9600英文字符</p><p>　　畢業(yè)設計(論文)外文資料翻譯</p><p>　　學 院： 信息工程學院 </p><p>　　專 業(yè)： 計算機科學與技術 </p><p>　　姓 名：

2、 </p><p>　　學 號： </p><p>　　外文出處 Knowledge Acquisition and Modeling, 2009. KAM'09. </p><p>　　Second International Symposium on. IEEE, 2009, 1:

3、 442-445. </p><p>　　附 件： 1.外文資料翻譯譯文；2.外文原文。 </p><p>　　附件1：外文資料翻譯譯文</p><p>　　設計實現電網知識的可視化</p><p>　　Zhaoyang Qu, Junzhou Liu</p><p>　　摘要：正確高效率的電網工作狀態(tài)是電力系統

4、運行穩(wěn)定性的保障。如今，電網知識學習方式不直觀，不生動，且效率不高。本文提出了電網知識可視化的概念，使用基于PFCM（概率模糊認知地圖）電網可視化技術，設計了電網知識的可視化平臺。在VR（虛擬現實）技術的基礎上，以圖形模式的方法來展示知識，可以很容易理解和分析。</p><p>　　關鍵詞：知識可視化;電網運行狀態(tài)顯示; PFCM，虛擬現實;SOA</p><p><b>　　1

5、．引言</b></p><p>　　電網的數據和運行條件信息是復雜和實時的。如何展示使其能夠有效地被人們學習接受是研究的熱點?！耙粡垐D片勝過千言萬語！”這句話反映了電網的運行狀態(tài)顯示的關鍵是可視化。通過圖形顯示模式，可視化向我們展示了電網的運行數據。人類比起計算機來說，在邏輯對比分析中具有令人難以置信的能力和優(yōu)勢。計算機安全事故的發(fā)生比人類快10000倍[1]，當從一個圖片發(fā)現趨勢或者深刻含義時，人類

6、的直覺往往比計算機更靈敏。然而，目前的電網的運行條件顯示系統面臨的一個普遍問題需要解決：</p><p>　　基于二維或三維的電網運行條件顯示方面有他們的缺點。例如，顯示的數據量太小，缺乏顯示使用和深層次的抽象信息，“看不清楚，看不明白，無法看透！”等缺點導致人們的學習能力降低，不能為潛在的安全威脅做出正確的決策和警告，所以前者的方法不能滿足生動的視覺需求[2]。</p><p>　　鑒于

7、上述問題，提出了基于電網知識可視化平臺的電網運行條件的顯示方法。</p><p>　　2.電網知識可視化平臺架構</p><p>　　電網知識可視化可以被定義為隱性知識，如電網實時數據與知識提取技術的顯性知識轉化過程，知識推理和虛擬現實等；當然也包括數據和信息提取、數據和信息的創(chuàng)新的處理。</p><p>　　基于SOA（面向服務的架構）的平臺架構如圖1所示。鏈接表

8、示邏輯層，業(yè)務邏輯層，客戶端層，底層數據庫與該服務總線技術的交互，以便處理電網的可視化和客戶端層獲得了知識可視化平臺中客戶端和服務器的鏈接。在平臺中它包括所有的系統的客戶端或設備。這些設備可以通過提供的API服務的平臺進行連接;表示邏輯層是對取得的知識可視化的核心處理，提取出電網實時資料和可視化的知識，通過Servlet和JSP單元實現，與此同時封裝了所有知識可視化服務和訪問系統中電網和客戶端的業(yè)務邏輯。業(yè)務邏輯層提供各種所需的平臺客戶

9、端中電網知識可視化的服務，包括平臺運行配置客戶端所有可視化的表示邏輯，并支持插件模式，平臺控制臺、底部渲染功能；通過平臺服務總線數據API的連接和SCADA，DMS的數據庫，底層數據庫提供所需的電網知識可視化的實時數據。</p><p>　　圖1.可視化平臺架構</p><p><b>　　3.電網知識提取</b></p><p>　　電網知識

10、可視化需要三個步驟。第一步：挖掘電網知識和建立可視化過程所需的知識單元;第二步：通過知識單元實現建立知識圖形的推理模型;第三步：通過OPENGL函數模型輸出，最后以餅狀圖和柱狀圖的形式顯示。</p><p>　　例如，我們選取局部電網中的運行情況，以此來詳細的討論怎樣實現電網知識的可視化。假設本地電網設備有8個，其中，我們以x，y和z分別顯示斷路器，過流保護和距離保護的數量。</p><p&g

11、t;　　1.1創(chuàng)建電網屬性列表的運轉狀態(tài)及相關矩陣</p><p>　　首先選擇一個示例，根據電網實時操作信息集，以斷路器，過流保護和距離保護條件屬性中的動作信息，創(chuàng)造條件屬性列表S，條件是屬性列表中包括數N（N= X + Y+ Z）的實時操作。</p><p>　　表1中為信息樣本，我們在其中列出斷路器CB1，CB2... CB（x），距離保護RR1，RR2... RR（y）時，過電流保

12、護器為CO1，CO2... CO（Z），其中所述值的條件屬性是“0”或“1”，“1”表示關閉修改斷路器及保護動作，“0”表示斷路器修改或復位動作的值。</p><p>　　表一 電網屬性列表操作條件</p><p>　　創(chuàng)建等差矩陣，以響應該包括n個樣本中的決策表，可以得到命令-N*N的矩陣所示為圖2：</p><p><b>　　圖2 可辨矩陣</

13、b></p><p>　　2.1創(chuàng)建電網知識操作單元</p><p>　　首先判斷識別矩陣MD的單一屬性，并分別為各矩陣元素建立分離表達式;然后進行計算，結合所有分離表達式，使之成為析取范式，以字母k加入析取表達式，逐個結合;最后，通過判斷結合的結果，發(fā)現知識單元Rmai[3]，過程和詳細說明如圖3。</p><p>　　圖3.電網知識單元建設進程</p

14、><p><b>　　電網知識推理</b></p><p>　　創(chuàng)建電網操作知識單元是建立在數據的基礎上的，然而知識單元是挖掘原始數據密鑰中的信息。并且不能以人類視覺形式為接受的方式來展示。創(chuàng)建出基于VR技術顯示方式的知識單元[4]的認知地圖，那么整個過程中就可以完成實現和應用知識的可視化。</p><p>　　1.1創(chuàng)建基于PFCM的數學模型&l

15、t;/p><p>　　使用PFCM（概率模糊認知圖）概念能夠有效地解決專家知識和不確定推理或問題之間的不確定性所產生的因果關系。因果關系的不確定性通過PFC和PFC來融合。而FCM增強了模擬因果關系FCM的能力[5]。數學模型為：</p><p><b>　　(1)</b></p><p>　　在式（1）中，n代表所有的概念數量，Vcj(t+1)代

16、表概念Cj狀態(tài)值在t+1內的時間，Vci（t）代表狀態(tài)值時間t的推理，Vcm（t）代表相關概念Cm狀態(tài)值在t內的時間。Wij（t）代表在強度t內概念Ci和概念Cj之間的因果關系，r代表這一時刻狀態(tài)值對下一時刻狀態(tài)值反應時間的影響因素，F（x）代表概念Cj的閾值函數。</p><p>　　2.1創(chuàng)建基于PFCM知識推理模型：</p><p>　　基于PFCM的知識推理模型：假設一個知識單元包

17、含n個信息點，那么PFCM模型應該包含n個節(jié)點，然后建立狀態(tài)矩陣C是，建立關系的權重矩陣W是：</p><p><b>　　（2）</b></p><p><b>　?。?）</b></p><p>　　狀態(tài)矩陣：C1，C2，C3，，Cn是n個信息點的狀態(tài)的值。其狀態(tài)轉換函數：</p><p>　　

18、Y=C*W （4）</p><p>　　然后，得到的每一個節(jié)點的狀態(tài)輸出：</p><p><b>　　（5）</b></p><p>　　yj是基于關系權重作用于信息點j和狀態(tài)Cj的知識動作[7]。這樣的信息點

19、狀態(tài)輸出函數可以得到一個結果，就是信息點j的新狀態(tài)。</p><p><b>　　（6）</b></p><p>　　矩陣計算和狀態(tài)輸出函數對知識單元的狀態(tài)轉變中的每一步進行推理[8]。假設n=6，x=3，y=1，z=2，則它可以基于PFCM上獲得知識的推理模型的關系得到的權重矩陣w。矩陣和PFCM如圖4所示：</p><p>　　圖4 關系權

20、重矩陣W和PFCM</p><p><b>　　VR展示</b></p><p>　　知識推理模型計算后可以得到結果值。結果值示于表‖：5次計算，每次間隔時間為1秒。</p><p>　　圖5中，平臺的VR是基于三維坐標所顯示的數據和其他數據的系統。通過旋轉z周圍360度的坐標軸，可以觀察到，信息分別被連接到x和y坐標軸。平臺建立了三維模型和一

21、般的動態(tài)交互系統，為用戶提供交互式窗口和工具。</p><p>　　以SF6斷路器CB1為例：x軸包括屬性有“設備名稱”，“設備規(guī)格”，“設備制造商”，“設備備注”;y軸包括屬性有“運行狀態(tài)”，“工作日志”，“維修記錄”。</p><p>　　移動鼠標到屬性欄中可以查看特定的屬性。Y軸“工作日志”和運行狀態(tài)”向我們展示以圖形形式的知識形式。并且可以選擇看到另外的圖形顯示方式，如“折線圖”，

22、“餅狀圖”，“直方圖”，等等。</p><p>　　當CB1存在問題時，平臺的虛擬現實顯示系統可以給用戶顯示CB1的哪部分有問題。從圖5中，可以看出用紅色標記出了斷路器斷電。平臺的虛擬現實顯示系統向我們展示了故障原因是由于天然氣泄漏率SF6的超標而發(fā)生的。超標能夠通過檢查運行狀態(tài)和工作日志查明原因。當問題出現時，點擊“維修記錄”斷電模式，以檢查和編輯其歷史的維護和維修記錄。然后將鼠標移動到x軸和檢測設備的特定屬性

23、，按照“設備規(guī)格”、“設備制造商”與設備制造商聯系進行維修并做出維修計劃。</p><p><b>　　結論</b></p><p>　　本文提出了電網運行條件顯示的新方法-電網知識可視化。這種方法通過轉變電網操作單元所獲取知識中的信息。通過基于PFCM與VR技術的數學模型，和顯示推理模型建立知識可視化的推理模型。然后通過VR顯示并實現電網知識的可視化。電網運行條件的

24、最終顯示的結果是圖形形式，使用戶可以生動直觀的觀察，理解，做出更有效正確的決策。</p><p><b>　　表2 計算結果值表</b></p><p><b>　　附件2：外文原文</b></p><p>　　Design and Realization of Power Gird Knowledge Visualiza

25、tion</p><p>　　Abstract—The correct, high efficiency power gird operating condition display is guarantee of the power system operating stability.Now, power gird operating condition display is not intuitive,vi

26、vid ,and efficiency is not high.The paper proposes the concept of power grid knowledge visualization,designs power gird visualization platform.Using power grid visualization technology based on PFCM(Probabilistic Fuzzy C

27、ognitive Map),power grid operating fundamental data can be transform into power grid op</p><p>　　Keywords-knowledge visualization; power grid operating condition display; PFCM, virtual reality; SOA</p>

28、<p>　　1. INTRODUCTION</p><p>　　Data and information of power grid operating condition is complex and real-time.How to display it and make it can be accepted effectively is research hotspot." A Pi

29、cture is worth a thousand words!" This sentence reflects that the key of power grid operating condition display is visualization. By the graphic display mode, visualization shows us power grid operating data. Human

30、has an advantage over computer since human have unbelievable ability of logical contrastive analysis. Computer security i</p><p>　　Power grid operating condition display based on two-dimensional or three-dim

31、ensional has disadvantages.For example, display Data Quantity is too small, lack of display usage and high-level abstractions information," can't see clearly, don't understand, can't see through!",

32、Disadvantages lead to human's perception ability low, and can't make correct decisions and warning for potential security threats, so the former methods can't satisfy the vivid and visual requirements [2].<

33、;/p><p>　　In view of the above problems, this paper proposes a power grid operating condition display method based on power grid knowledge visualization platform. </p><p>　　2. POWER GRID KNOWLEDGE

34、VISUALIZATION PLATFORM ARCHITECTURE </p><p>　　Power grid knowledge visualization can be defined as the process that transform the tacit knowledge such as power grid real-time data to explicit knowledge with

35、the technology of knowledge extraction, knowledge reasoning and virtual reality, ect;Also the process of data and information extraction and data and information innovation.</p><p>　　Platform architecture as

36、 shown in Figure 1 which is based on SOA (Service Oriented Architecture). Link the presentation logic layer, business logic layer, the client layer, the bottom database to each other with the service bus technology, in o

37、rder to deal with the visualization of the power grid the client layer achieved the connection of the knowledge visualization Platform Client and server.It includes all the system client or equipment of the platform.Thes

38、e devices can be connected through t</p><p>　　processing it extract the power grid real-time datas as knowledge,than visualize the knowledge,achieve the element by Servlet and JSP,At the same time, it encaps

39、ulates all the the presentation logic of visualization of the power grid and client of service and access system.The business logic layer provides kinds of the power grid knowledge visualization business services needed

40、by platform client,Including the platform running configuration, the model supporting plug-ins, the platform console fe</p><p>　　3. THE POWER GRID KNOWLEDGE EXTRACTION</p><p>　　The power grid kn

41、owledge visualization requires three steps. Step one is to extract in the power grid knowledge mining and the knowledge unit required to establish the visualization process; Step two : the knowledge unit is to implement

42、the knowledge inferring model by establishing knowledge graph; Step three: the model output through the OPENGL function, can display in the form of pie charts and histograms.</p><p>　　Taken the running situa

43、tion of local power grid equipment for example, the detailed realization of the power grid knowledge visualization was discussed.</p><p>　　Suppose the local power grid equipment consists of 8, the form in wh

44、ich we show the number of breaker, overcurrent protection and distance protection are x,y and z respectively.</p><p>　　1.1 Create the operating condition of power grid attribute list and related identifiable

45、 matrix </p><p>　　It firstly selects the sample set according to power grid real-time operation information, then taking the action message of breaker, overcurrent protection and distance protection as condi

46、tion attributes, and creates the condition attributes list S. The condition attributes list consists of the number of n(n=x+y+z) real-time operation </p><p>　　Figure 1. Visualization platform architecture &

47、lt;/p><p>　　information samples as shown as table 1, the attribute in which we show the breaker CB1 CB2…CB(x), distance protection RR1 RR2…RR(y) overcurrent protection CO1 CO2…CO(z) the value of the condition a

48、ttributes is "0" or "1", the value of "1" express breaker from shut modification to cut and protection action, and the value of "0" express breaker unmodification or reset action.

49、</p><p>　　TABLE I. THE OPERATING CONDITION OF POWER GRID ATTRIBUTE LIST </p><p>　　Create the decision table of discernible matrix , in response to the decision table consists of n samples, can

50、be get order-N*N discernible matrix as shown as Figure 2: </p><p>　　Figure 2. Discernible matrix </p><p>　　2.1 Create the power grid knowledge operating unit </p><p>　　It firstly ma

51、ke a single attribute judgment of the discernible matrix MD, and build up the disjunctive expression for all matrix element respectively; then carry out conjunction calculations for all disjunctive expression and turn it

52、 into disjunctive normal form, the elements in set K join the disjunctive expression in order to conjunction one by one;finally, by judging the conjunction result and so the knowledge unit Rmai was found[3]. The processe

53、s are described in detail within </p><p><b>　　Figure 3.</b></p><p>　　4. THE POWER GRID KNOWLEDGE REASONING </p><p>　　Creation of the power grid operating knowledge unit

54、is for establish a data basis but the effect of knowledge unit is original data key information mining.Knowledge unit is not be displayed in manner of visual forms that can be accepted by human.Creat the cognitive map us

55、ing knowledge unit[4],and show the cognitive map in VR technology,then the whole realization and application of knowledge visualization process is completed. </p><p>　　1.1 Create the mathematical model based

56、 on PFCM</p><p>　　Using PFCM (probabilistic fuzzy cognitive map) can effective resolve uncertainty of experts' knowledge and uncertain reasoning or issue of the causal relationship between the concepts.

57、Uncertainty of causality is blent into FCM by PFCM.PFCM enhance FCM's abilities that simulate causality[5].The mathematical model is: </p><p><b>　　（1）</b></p><p>　　In the formula

58、(1),n is quantity of all concepts, Vcj(t+1) stand for state value of result concept cj at the time of t+1, Vci(t) stand for state value of reason concept ci at the time of t, Vcm(t) stand for state value of relevant con

59、cept cm at the time of t[6]. Wij(t) stand for causal relation Intensity between concept ci and concept cj at the time of t,r stand for impact factors that state value of this moment react on the next moment, f(x) stand f

60、or threshold function of concept Cj. </p><p>　　Figure 3. Power grid knowledge unit building process </p><p>　　2.1 Create the knowledge reasoning model based on PFCM:</p><p>　　The k

61、nowledge reasoning model based on PFCM is:suppose a knowledge unit contains n information points,so the PFCM model should contains n nodes,then establish state matrix c that is and establish relationship weight matrix W

62、 that is : </p><p><b>　?。?）</b></p><p>　　In the state matrix: c1,c2,c3, ,cn are state value of n information points.Its state transition function is: </p><p>　　Y=C*W

63、 (4)</p><p>　　Then it gets status output of every node:</p><p><b>　?。?）</b></p><p>　　yj is knowledge acti

64、on union of all forward information points based on relationship weights act on state Cj of information point j[7].So a state output function of information point can obtain a result that is the new state of information

65、point j.</p><p><b>　　（6）</b></p><p>　　State transition of knowledge unit is reasoning,matrix calculation and state output function carry out each step of reasoning[8].Suppose n=6,x=3

66、,y=1,z=2,then it can get a result of knowledge reasoning model's relationship weight matrix W based on PFCM. Matrix and PFCM is shown in Figure 4: </p><p>　　Figure 4. Relationship weight matrix W and PF

67、CM </p><p>　　VR DISPLAY</p><p>　　It can get result value after calculation of knowledge reasoning model.The reslut values is shown in TableII: 5 times calculation, 1 second distance every time.

68、</p><p>　　As shown in Figure 5,Platform's VR display system is based on three-dimensional coordinate, knowledge value that is displaied and other data need to be classified. The classification is connect

69、ed to the X and y coordinate axis.It can be observed for around 360 degrees through rotate z coordinate axis.Platform build three-dimensional model and general dynamic interactive system, providing interactive windows an

70、d tools for users. </p><p>　　Take a SF6 breaker CB1 for example: x axis includes attributes" equipment name"," equipment specifications","equipment manufacturer"," equipmen

71、t remarks"; y axis includes attributes" operating state"," operating log"," maintenance records". </p><p>　　Move mouse to attributes column can check specific attributes. y

72、 axis's "operating log"and" operating state"show us knowledge value in graphic form.It can be choosen to see in many forms,like" line charts"," pie chart","histogram"

73、,etc. </p><p>　　When there is a problem with CB1, platform's VR display system can show users which part of CB1 has a problem.From the Figure 5, it can be seen arc extinguish </p><p>　　chamb

74、er that broken down is marked with red color. Platform's VR display system show us the fault reason that is gas-leakage-rate of SF6 exceeded standards.The reason that exceeded standards can be deduced by checking ope

75、rating state and operating log. When the problem appears, click the "maintenance records" of breaker model to check its historical maintenance and edit its maintenance log.Also move the mouse to x axis's an

76、d check equipment specific attributes, make maintenance plan according to</p><p>　　6. CONCLUSION </p><p>　　This paper proposes a new method of power grid operating condition display - power grid

77、 knowledge visualization. This method fetches power grid information with the knowledge, and transform the information into power grid operating unit. It establish reasoning model after establish the mathematical model o