《電子技術(shù)應(yīng)用》
您所在的位置:首頁 > 通信與網(wǎng)絡(luò) > 設(shè)計(jì)應(yīng)用 > 面向邊緣計(jì)算的電力通信網(wǎng)告警歸并技術(shù)研究
面向邊緣計(jì)算的電力通信網(wǎng)告警歸并技術(shù)研究
2021年電子技術(shù)應(yīng)用第4期
李霽軒1,,吳子辰1,郭 燾1,,朱鵬宇2,,吳季樺3
1.國網(wǎng)江蘇省電力有限公司信息通信分公司,江蘇 南京210000,; 2.國網(wǎng)電力科學(xué)研究院有限公司,,江蘇 南京210012;3.北京郵電大學(xué)網(wǎng)絡(luò)與交換國家重點(diǎn)實(shí)驗(yàn)室,,北京100876
摘要: 電力通信網(wǎng)的覆蓋范圍及復(fù)雜程度逐漸增大,,為電力通信網(wǎng)帶來巨大的運(yùn)維壓力。通過部署邊緣節(jié)點(diǎn)在邊緣側(cè)完成數(shù)據(jù)采集和信息過濾,,提供計(jì)算支持,,能夠極大程度上緩解電力通信網(wǎng)管理側(cè)壓力。告警分析是運(yùn)維當(dāng)中的重難點(diǎn)問題,,傳統(tǒng)的告警分析先使用規(guī)則對于告警進(jìn)行歸并從而減少后續(xù)處理的工作量,,但是規(guī)則的完備需要大量專家知識和人力資源的投入且存在局限性。將無監(jiān)督聚類引入到電力通信邊緣云部署架構(gòu)下的告警歸并流程當(dāng)中,,提出了一個(gè)新的輕量級算法,,將基于密度的聚類方法(Density-Based Spatial Clustering of Applications with Noise,DBSCAN)與現(xiàn)有的歸并規(guī)則進(jìn)行結(jié)合,,結(jié)果表明加入無監(jiān)督學(xué)習(xí)能夠顯著提高告警歸并的效果,,切實(shí)提高了電力通信網(wǎng)運(yùn)維中缺陷定位的準(zhǔn)確性和完備性。
中圖分類號: TN915
文獻(xiàn)標(biāo)識碼: A
DOI:10.16157/j.issn.0258-7998.201269
中文引用格式: 李霽軒,,吳子辰,,郭燾,等. 面向邊緣計(jì)算的電力通信網(wǎng)告警歸并技術(shù)研究[J].電子技術(shù)應(yīng)用,,2021,,47(4):17-23.
英文引用格式: Li Jixuan,Wu Zichen,,Guo Tao,,et al. Research of alarm correlation technique for edge cloud computing in power communication network[J]. Application of Electronic Technique,2021,,47(4):17-23.
Research of alarm correlation technique for edge cloud computing in power communication network
Li Jixuan1,,Wu Zichen1,,Guo Tao1,Zhu Pengyu2,,Wu Jihua3
1.Information and Communication Branch of State Grid Jiangsu Electric Power Co.,,Ltd,Nanjing 210000,,China,; 2.State Grid Electric Power Research Institute Co.,Ltd,,Nanjing 210012,,China; 3.State Key Laboratory of Networking and Switching Technology,,Beijing University of Posts and Telecommunications,, Beijing 100876,China
Abstract: The demand for electric power and the coverage of power communication network has been gradually expanding, which brings the challenges that need to be addressed in the operation and maintenance of the domain. The deployment of edge nodes provides the availability of data collection, information filtering and computational support, which can heavily alleviate the pressure of management. Alarm analysis is a key and difficult problem in operation and maintenance. The traditional alarm analysis first uses rules to merge alarms, so as to reduce the workload of subsequent processing. However, the completeness of rules requires a lot of expert knowledge and human resources investment, and there are limitations. This paper proposed a novel and lightweight algorithm. In this paper, Unsupervised clustering is introduced into the alarm merging process of power communication network edge cloud computing, and the density based clustering method is combined with the existing merging rules. The experimental results show that the effect of alarm merging can be significantly improved by adding unsupervised learning, which is helpful to improve the accuracy and completeness of subsequent defect location.
Key words : unsupervised learning,;DBSCAN,;alarm correlation;edge computing

0 引言

    我國電力行業(yè)的高效平穩(wěn)發(fā)展是保證經(jīng)濟(jì)安全,、快速,、穩(wěn)定發(fā)展的能源保障。人工智能時(shí)代對電力通信領(lǐng)域提出了新的要求,,也為電力通信管理系統(tǒng)(Telecom Management System,,TMS)的發(fā)展提供了新方向[1]。TMS作為電力領(lǐng)域信息化產(chǎn)物,,為整個(gè)電力系統(tǒng)中的電網(wǎng)調(diào)度,、自動(dòng)化、繼電保護(hù),、安全自動(dòng)控制,、電力市場交易以及企業(yè)信息化等工作提供了堅(jiān)實(shí)的基礎(chǔ),,同時(shí)也為電力通信中的異常檢測,、路由優(yōu)選等智能化應(yīng)用提供支撐。

    隨著特高壓電網(wǎng),、各級電網(wǎng)協(xié)調(diào)的統(tǒng)一發(fā)展,,智能網(wǎng)的建設(shè)的需求也逐漸加強(qiáng),對支撐電網(wǎng)信息化基礎(chǔ)TMS系統(tǒng)提出了更高要求,。在電力通信信息化,、智能化建設(shè)和應(yīng)用實(shí)踐過程中,電力公司積累了海量的實(shí)時(shí)數(shù)據(jù)和運(yùn)行數(shù)據(jù),,傳統(tǒng)基于規(guī)則的缺陷處置方法難以滿足智能化的需求,,尤其缺乏一種對拓?fù)鋸?fù)雜,、設(shè)備類型繁多的缺陷數(shù)據(jù)進(jìn)行智能分析的方法[2]

    電力通信網(wǎng)在信息化過程中產(chǎn)生了大量的數(shù)據(jù),,然而這些數(shù)據(jù)的海量增長,,促使了數(shù)據(jù)歸并技術(shù)(即告警歸并技術(shù))的發(fā)展。目前國內(nèi)外主要使用基于規(guī)則匹配的方法進(jìn)行告警歸并[3],。具體而言,,就是操作員根據(jù)系統(tǒng)實(shí)時(shí)情況結(jié)合專家知識動(dòng)態(tài)地調(diào)整告警歸并規(guī)則。同時(shí),,也有基于規(guī)則匹配方法上的改進(jìn),。例如,加入數(shù)據(jù)預(yù)處理和數(shù)據(jù)過濾等方法輔助告警歸并[4],。上述方法在告警數(shù)據(jù)規(guī)模較小,、告警延遲低、告警類別固定等情況下,,能達(dá)到很好的歸并效果,。但隨著告警數(shù)據(jù)的海量增長,上述方法及其相關(guān)改進(jìn)方法難以適應(yīng)當(dāng)前的數(shù)據(jù)環(huán)境,。MADZIARZ A在移動(dòng)通信網(wǎng)領(lǐng)域提出了基于K-MEANS聚類的告警聚類方法[5],,嘗試引入無監(jiān)督聚類以擺脫對規(guī)則的依賴。雖然該方法無須大量人力資源的投入,,但實(shí)際歸并效果差強(qiáng)人意,,且需要業(yè)務(wù)專家參與預(yù)測缺陷的數(shù)量,有著極大的局限性,。

    5G技術(shù),、邊緣計(jì)算、人工智能新技術(shù)的到來給電力通信領(lǐng)域帶來了新鮮血液,。新技術(shù)與電力通信領(lǐng)域的有機(jī)結(jié)合,,對于構(gòu)造電力通信新生態(tài),解決遺留問題,,節(jié)約人力資源,,面對新的挑戰(zhàn)至關(guān)重要。

    本文介紹了一種基于密度聚類(Density-Based Spatial Clustering of Applications with Noise,,DBSCAN)[6]結(jié)合人工規(guī)則進(jìn)行告警歸并協(xié)助通信缺陷診斷的無監(jiān)督學(xué)習(xí)算法,。該算法具有良好的魯棒性、輕量性,,支持邊緣云部署,,將算法在TMS系統(tǒng)提供的數(shù)據(jù)中進(jìn)行實(shí)驗(yàn),結(jié)果顯示算法達(dá)到了較好的效果。




本文詳細(xì)內(nèi)容請下載:http://forexkbc.com/resource/share/2000003457




作者信息:

李霽軒1,,吳子辰1,,郭  燾1,朱鵬宇2,,吳季樺3

(1.國網(wǎng)江蘇省電力有限公司信息通信分公司,,江蘇 南京210000;

2.國網(wǎng)電力科學(xué)研究院有限公司,,江蘇 南京210012,;3.北京郵電大學(xué)網(wǎng)絡(luò)與交換國家重點(diǎn)實(shí)驗(yàn)室,北京100876)

此內(nèi)容為AET網(wǎng)站原創(chuàng),,未經(jīng)授權(quán)禁止轉(zhuǎn)載,。