• ๋Œ€ํ•œ์ „๊ธฐํ•™ํšŒ
Mobile QR Code QR CODE : The Transactions of the Korean Institute of Electrical Engineers
  • COPE
  • kcse
  • ํ•œ๊ตญ๊ณผํ•™๊ธฐ์ˆ ๋‹จ์ฒด์ด์—ฐํ•ฉํšŒ
  • ํ•œ๊ตญํ•™์ˆ ์ง€์ธ์šฉ์ƒ‰์ธ
  • Scopus
  • crossref
  • orcid

  1. (Dept. of Electrical Engineering, Gangneung-Wonju National University, Korea)



AI, ANN, BP, Fault recovery system, Fault type identification, Learning, SOP, Substation

1. ์„œ ๋ก 

๋ณ€์ „์†Œ๋Š” ๋ฐœ์ „์†Œ์˜ ๊ณ ์ „์•• ์ „๋ ฅ์„ ์ˆ˜์šฉ๊ฐ€์— ์ €์ „์•• ์ „๋ ฅ์œผ๋กœ ๋ณ€ํ™˜์‹œ์ผœ ์•ˆ์ •์ ์œผ๋กœ ๊ณต๊ธ‰ํ•˜๋Š” ์ „๋ ฅ๊ณ„ํ†ต์˜ ๊ตฌ์„ฑ์š”์†Œ ์ค‘์˜ ํ•˜๋‚˜์ด๋‹ค. ์ „๋ ฅ๊ณ„ํ†ต์˜ ๋Œ€๊ทœ๋ชจ ์ •์ „์„ ๋Œ€๋น„ํ•˜์—ฌ ๋ณ€์ „์†Œ๋Š” ์ค‘์•™์—์„œ ์‹ค์‹œ๊ฐ„์œผ๋กœ ๊ณ„ํ†ต์— ๊ด€ํ•œ ์ค‘์š”ํ•œ ์ •๋ณด๋“ค์„ ์–ป๊ณ , ๊ณ„ํ†ต์„ ์ œ์–ดํ•  ์ˆ˜ ์žˆ๋Š” ๋Šฅ๋ ฅ์ด ๋ถ€๊ฐ€๋˜๋ฉด์„œ ์†Œ๋น„์ž์—๊ฒŒ ๋‹ค์–‘ํ•œ ์„œ๋น„์Šค๋ฅผ ํ•ด์ค„ ์ˆ˜ ์žˆ๋„๋ก ๋ณ€ํ™”ํ•˜๊ณ  ์žˆ๋‹ค. ์„ธ๊ณ„์ ์œผ๋กœ IEC 61850 ๊ตญ์ œํ‘œ์ค€๊ทœ๊ฒฉ ๊ธฐ๋ฐ˜์˜ ๋ณ€์ „์†Œ์ž๋™ํ™”์‹œ์Šคํ…œ(SAS : Substation Automation System)์ด ์šด์˜ ์ค‘์— ์žˆ๋Š” ๋ฐ”, ๊ตญ๋‚ด์—์„œ๋„ IED๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ๊ธฐ์กด ๋ณ€์ „์†Œ ๋ฐ ์‹ ์„ค ๋ณ€์ „์†Œ๋ฅผ SAS๋กœ ๊ตฌ์ถ•ํ•˜๊ณ  ์žˆ๋‹ค[1,2].

๋ณ€์ „๋ถ„์•ผ ๊ณ ์žฅ์‚ฌ๋ก€์ง‘์— ๊ทผ๊ฑฐํ•˜๋ฉด 2012๋…„ ์ฃผ์š” ๊ณ ์žฅ์— ๋Œ€ํ•œ 100[MVA] ๊ณ ์žฅ๋ฅ ์€ 0.0066, 271,247[MVA]์— ๋‹ฌํ•˜๊ณ  ์žˆ๋‹ค[3]. ์ „๋ ฅ์„ค๋น„์— ๊ณ ์žฅ์ด ๋ฐœ์ƒํ–ˆ์„ ๊ฒฝ์šฐ ์‹ ์†ํ•œ ๊ณ ์žฅ๋ณต๊ตฌ์™€ ๊ณ ์žฅํŒŒ๊ธ‰๋ฐฉ์ง€๋ฅผ ์œ„ํ•˜์—ฌ ํ•œ์ „์—์„œ๋Š” ํ‘œ์ค€๋ณต๊ตฌ์ ˆ์ฐจ(SOP : Standard Operation Procedure)๋ฅผ ์ˆ˜๋ฆฝ, ์‹œํ–‰ํ•˜๊ณ  ์žˆ๋‹ค[4]. ๊ทธ๋Ÿฌ๋‚˜ ์ƒˆ๋กœ์šด ์ „๋ ฅ์„ค๋น„์˜ ์šด์ „์— ๋”ฐ๋ฅธ ๊ธฐ์กด๊ณผ ๋‹ค๋ฅธ ๊ณ ์žฅ์œ ํ˜•์ด ๋ฐœ์ƒํ–ˆ์„ ๊ฒฝ์šฐ ๊ณ ์žฅ๋ณต๊ตฌ์— ๋งŽ์€ ์–ด๋ ค์›€์ด ๋‚ดํฌํ•˜๊ณ  ์žˆ๋‹ค. ์ด์— ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์€ ์ „๋ ฅ๊ณต๊ธ‰ ์ค‘๋‹จ๊ณผ ๋น„์ƒ์‚ฌํƒœ๋ฅผ ์ผ์œผํ‚ค๊ธฐ ๋•Œ๋ฌธ์— ๊ฐ€๋Šฅํ•œ ์‹ ์†ํ•˜๊ฒŒ ๊ณ ์žฅ์„ ์ฐพ์•„๋‚ด์–ด ์ •์ „๋ณต๊ตฌ์‹œ๊ฐ„๊ณผ ๊ฒฝ์ œ์ ์ธ ์†์‹ค์„ ์ตœ์†Œํ™”ํ•ด์•ผ ํ•œ๋‹ค[5,6].

์ตœ๊ทผ, ์ปดํ“จํ„ฐ ํ•˜๋“œ์›จ์–ด ๋ฐ ์†Œํ”„ํŠธ์›จ์–ด ํ”Œ๋žซํผ์˜ ๋ฐœ์ „์œผ๋กœ ์Šค๋งˆํŠธํ•œ ์žฅ์ ์ด ์žˆ๋Š” ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง(ANN : Artificial Neural Network)์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋ถ€์ƒํ•˜๊ณ  ์žˆ๋‹ค. ํŠนํžˆ, ์•ŒํŒŒ๊ณ  ์ดํ›„, AI ๊ณ ๋„ํ™”, ์ž๊ธฐํ•™์Šต(Self Learning) ๋“ฑ์„ ๋ณ€์ „์†Œ ๊ณ ์žฅํŒ๋‹จ ์‹œ์Šคํ…œ์— ์ ์šฉํ•จ์œผ๋กœ์„œ ์ž๋™๊ณ ์žฅ๋ณต๊ตฌ ์‹œ์Šคํ…œ์œผ๋กœ ์ „ํ™˜ํ•˜๋ ค๋Š” ์›€์ง์ž„์ด ์ œ๊ธฐ๋˜์—ˆ๋‹ค. ๊ทธ๋Ÿฐ๋ฐ, ์‹ ์†ํ•œ ๊ณ ์žฅ๋ณต๊ตฌ๋ฅผ ์œ„ํ•ด์„œ๋Š” ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„๊ณผ ๊ณ ์žฅ์  ํ‘œ์ • ํŒ๋‹จ์ด ์„ ํ–‰๋˜์–ด์•ผ ํ•œ๋‹ค[7,8].

ํ•ด์™ธ์˜ ๊ณ ์žฅ์  ํ‘œ์ •์— ๊ด€ํ•œ ์—ฐ๊ตฌ๋กœ, ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์  ํ‘œ์ •์„ ์œ„ํ•œ ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์ด ์ ์šฉยท์ œ์‹œ๋œ ์ดํ›„, ๊ทผ๋ž˜์—๋Š”, ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์„ ์ด์šฉํ•œ ๋ถ„์‚ฐ์ „์›์˜ ๊ณ ์žฅ์  ํ‘œ์ •๊ณผ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์ธ์‹์„ ์œ„ํ•œ ์›จ์ด๋ธŒ๋ฆฟ๊ธฐ๋ฐ˜ ์‹ ๊ฒฝํšŒ๋กœ๋ง์ด ๋ฐœํ‘œ๋˜์—ˆ๋‹ค[9โˆผ12].

๊ตญ๋‚ด์—์„œ๋Š” ๊ณ ์žฅํ‘œ์ • ์•Œ๊ณ ๋ฆฌ์ฆ˜์— ์ ํ•ฉํ•œ ์›ํ˜• ์›จ์ด๋ธŒ๋ฆฟ ๋ณ€ํ™˜์ด ๋ฐœํ‘œ๋œ ์ดํ›„, ์†ก์ „์„ ๋กœ์šฉ ๋””์ง€ํ„ธ ํ‘œ์ •์žฅ์น˜(Fault Locator)๊ฐ€ ๊ฐœ๋ฐœ๋˜์—ˆ๊ณ , ๊ทผ๋ž˜์—๋Š” Weka ์†Œํ”„ํŠธ์›จ์–ด๋ฅผ ์ด์šฉํ•œ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์  ํ‘œ์ •์ด ์ œ์‹œ๋˜์—ˆ๋‹ค[13~15]. ๋˜ํ•œ, ๋ฐฐ์ „ ๋ณ€์ „์†Œ๋ฅผ ์œ„ํ•œ ์ „๋ฌธ๊ฐ€์‹œ์Šคํ…œ ๊ธฐ๋ฐ˜์˜ ๋ณต๊ตฌ์™€ SCADA ์‹œ์Šคํ…œ์— ์˜ํ•œ GIS ๊ณ ์žฅ์œ„์น˜๊ฒ€์ถœ์— ๊ด€ํ•œ ์—ฐ๊ตฌ๋„ ์ง„ํ–‰๋˜์—ˆ๋‹ค[16,17]. ํ˜„์žฌ AI ๋“ฑ ์ง€๋Šฅ๊ธฐ๋ฒ•์„ ์ด์šฉํ•œ ์ž๋™ํ™” ๋””์ง€ํ„ธ๋ณ€์ „์†Œ ๊ณ ์žฅ๋ณต๊ตฌ๋ฐฉ์•ˆ์˜ ๊ธฐ์ดˆ ์—ฐ๊ตฌ๊ณผ์ œ๊ฐ€ ์ง„ํ–‰ ์ค‘์— ์žˆ๋‹ค[18~21].

๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š”, ์ด ๊ณผ์ œ์˜ ์ผ๋ถ€ ์‚ฐ์ถœ๋ฌผ๋กœ์„œ ANN์„ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„(Fault Type Identification)์— ์ ์šฉํ•˜๊ณ ์ž ํ•œ๋‹ค. ๋จผ์ €, ํ•ด๋‹น ๋ณ€์ „์†Œ์˜ ๊ตฌ์„ฑ์š”์†Œ์ธ CB, DS, IED ๋“ฑ์˜ ๋™์ž‘์ƒํƒœ์™€ SOP์˜ ๊ณ ์žฅ์œ ํ˜•์„ ์ด์šฉํ•˜์—ฌ ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์˜ ๊ตฌ์กฐ๋ฅผ ์„ค๊ณ„ํ•˜์˜€๋‹ค. ํ•™์ŠตํŒจํ„ด์€ ์ •์ƒ์ƒํƒœ์™€ SOP์— ๊ทœ์ •๋œ 15๊ฐ€์ง€์˜ ๊ณ ์žฅ์œ ํ˜•์„ ๊ณ ๋ คํ•˜์—ฌ ํฌํ•จ์‹œ์ผฐ๋‹ค. ์—ญ์ „ํŒŒ (BP : Back Propagation)๋ฅผ ํ†ตํ•ด ์ œ์‹œํ•œ ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์„ ํ•™์Šตํ•œ ํ›„, ์‹œํ—˜ํŒจํ„ด์œผ๋กœ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„ ์—ฌ๋ถ€๋ฅผ ์‹œํ—˜ํ•˜์˜€๋‹ค. ๋์œผ๋กœ ์ œ์‹œํ•œ ๊ธฐ๋ฒ•์„ ๋‹ค์–‘ํ•œ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ์กฐ๊ฑด์—์„œ ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•˜๊ณ ์ž ํ•œ๋‹ค.

2. ๋ณ€์ „์†Œ ํ‘œ์ค€๋ณต๊ตฌ์ ˆ์ฐจ

2.1 ๋ณ€์ „์†Œ

๋ณ€์ „์†Œ๋Š” ์ „์••์˜ ๋ณ€์„ฑ, ์ „๋ ฅ์˜ ์ง‘์ค‘ ๋ฐ ๋ฐฐ๋ถ„, ์กฐ์ƒ์„ค๋น„์™€ ์ „์••์กฐ์ •์žฅ์น˜์— ์˜ํ•œ ์ „์••์กฐ์ •, ์ „๋ ฅ์กฐ๋ฅ˜์˜ ์ œ์–ด์™€ ์†ก๋ฐฐ์ „์„  ๋ฐ ๋ณ€์ „์†Œ๋ฅผ ๋ณดํ˜ธํ•˜๋Š” ๊ธฐ๋Šฅ์„ ํ•œ๋‹ค. IEC 61850์€ ์ฃผ๋กœ ๋ฏธ๊ตญ์—์„œ ์‚ฌ์šฉ๋˜๋˜ UCA 2.0์— ์œ ๋Ÿฝ์˜ ์š”๊ตฌ์‚ฌํ•ญ์„ ํ•จ๊ป˜ ๊ณ ๋ คํ•˜์—ฌ ๊ตญ์ œ ํ‘œ์ค€์œผ๋กœ ๊ฐœ๋ฐœ๋œ ํ”„๋กœํ† ์ฝœ์ด๋‹ค. ๊ทธ๋ฆผ 1์€ IEC 61850 ๊ธฐ๋ฐ˜ SAS์˜ ๊ตฌ์กฐ๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ๊ทธ๋ฆผ 1๊ณผ ๊ฐ™์ด SAS๋Š” 3๋‹จ๊ณ„ ๋ ˆ๋ฒจ(Station/Bay/Process Level)๊ณผ ๊ฐ ๋ ˆ๋ฒจ์„ ์—ฐ๊ณ„ํ•˜๊ธฐ ์œ„ํ•ด ๋‘ ๊ฐ€์ง€์˜ ํ†ต์‹  ๋„คํŠธ์›Œํฌ(Station/Process Bus)๋กœ ๊ตฌ์„ฑ๋œ๋‹ค[1].

๊ทธ๋ฆผ. 1. ๋ณ€์ „์†Œ์ž๋™ํ™” ์‹œ์Šคํ…œ์˜ ๊ตฌ์กฐ

Fig. 1. Structure of SAS

../../Resources/kiee/KIEE.2019.68.9.1039/fig1.png

๊ทธ๋ฆผ 2๋Š” 154kV ํ‘œ์ค€ ๋ณ€์ „์†Œ์˜ ๋‹จ์„ ๋„๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค[1,4,5,18]. ๊ทธ๋ฆผ 2์™€ ๊ฐ™์ด ์†ก์ „์„ ๋กœ(T/L), ๋ฐฐ์ „์„ ๋กœ(D/L), 2๊ฐœ์˜ ์†ก์ „๋ชจ์„ (TBUS), ๋ฐฐ์ „๋ชจ์„ (DBUS), 4๊ฐœ์˜ ์ฃผ๋ณ€์••๊ธฐ(M.Tr), 51๊ฐœ์˜ ์ฐจ๋‹จ๊ธฐ ๋ฐ 96๊ฐœ์˜ ๋‹จ๋กœ๊ธฐ ๋“ฑ์œผ๋กœ ๊ตฌ์„ฑ๋œ๋‹ค. T/L, D/L, TBus, DBUS, M.Tr, ๋“ฑ์œผ๋กœ ๊ตฌํš๋˜์–ด ์žˆ๊ณ  ๊ณ„์ ˆ ๋“ฑ์— ๋”ฐ๋ผ ๋‹ค์–‘ํ•˜๊ฒŒ ์šด์˜๋˜๊ธฐ ๋•Œ๋ฌธ์— ๊ณ ์žฅ์ด ๋ฐœ์ƒํ•  ๊ฒฝ์šฐ ๊ณ ์žฅ์  ํ‘œ์ •๊ณผ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์ด ์šฉ์ดํ•˜์ง€ ์•Š๋‹ค.

๊ทธ๋ฆผ. 2. 154kV ํ‘œ์ค€ ๋ณ€์ „์†Œ์˜ ๋‹จ์„ ๋„

Fig. 2. Single line diagram of 154kV standard substation

../../Resources/kiee/KIEE.2019.68.9.1039/fig2.png

2.2 ํ‘œ์ค€๋ณต๊ตฌ์ ˆ์ฐจ

ํ‘œ์ค€๋ณต๊ตฌ์ ˆ์ฐจ๋Š” ๋ณ€์ „์†Œ์—์„œ ๋ฐœ์ƒํ•˜๋Š” ์ •์ „ ๋ฐ ๋ฌด์ •์ „ ๊ณ ์žฅ์— ๋Œ€ํ•œ SOP๋ฅผ ๊ทœ์ •ํ•˜์—ฌ, ์ •์ „์‹œ๊ฐ„ ์ตœ์†Œํ™” ๋ฐ ์•ˆ์ •์  ์ „๋ ฅ๊ณต๊ธ‰์„ ๋„๋ชจํ•จ์„ ๋ชฉ์ ์œผ๋กœ ํ•œ๋‹ค.

์ด ์ ˆ์ฐจ์„œ์—๋Š” ๋ณ€์ „์†Œ ๊ณ ์žฅ๋ฐœ์ƒ์‹œ ๋Œ€์ฒ˜ํ๋ฆ„๋„์™€ ์•„๋ž˜์™€ ๊ฐ™์€ 15๊ฐ€์ง€ ๊ณ ์žฅ์œ ํ˜•์— ๋Œ€ํ•œ ํ‘œ์ค€๋ณต๊ตฌ์ ˆ์ฐจ๋ฅผ ์ˆ˜๋กํ•˜๊ณ  ์žˆ๋‹ค[1, 4, 5, 18].

$\quad\quad$- ์œ ํ˜•1 : T/L ์ฃผ๋ณดํ˜ธ ๋‹จ(์ง€)๋ฝ, ํ›„๋น„๋ณดํ˜ธ ๋‹จ(์ง€)๋ฝ

$\quad\quad$- ์œ ํ˜•2 : T/L ์†ก์ „์„ ๋‹จ์„ 

$\quad\quad$- ์œ ํ˜•3 : BUSPRO 87B1

$\quad\quad$- ์œ ํ˜•4 : M.Tr CB B/F, M.Tr 87, 96Ry

$\quad\quad$- ์œ ํ˜•5 : T/L CB B/F, ์ฃผ๋ณดํ˜ธ ๋‹จ(์ง€)๋ฝ, ํ›„๋น„๋ณดํ˜ธ ๋‹จ(์ง€)๋ฝ

$\quad\quad$- ์œ ํ˜•6 : BUSPRO 87B1, 87B2

$\quad\quad$- ์œ ํ˜•7 : BUSPRO 87B1, 87B2(154kV ์–‘DS ON์‹œ)

$\quad\quad$- ์œ ํ˜•8 : M.Tr 87, 96P(D,T)

$\quad\quad$- ์œ ํ˜•9 : M.Tr 59GA

$\quad\quad$- ์œ ํ˜•10: M.Tr 59GT

$\quad\quad$- ์œ ํ˜•11: M.Tr 51SN(51S, 51P)

$\quad\quad$- ์œ ํ˜•12: M.Tr 51SN(51S, 51P), โ’ถD/L OC(G)R

$\quad\quad$- ์œ ํ˜•13: #1 M.Tr 51SN(51S, 51P) ๋‹ค์ค‘๋ชจ์„ ์šด์ „์‹œ

$\quad\quad$- ์œ ํ˜•14: #1,2 M.Tr 51SN(51S, 51P) ๋‹ค์ค‘๋ชจ์„ ์šด์ „์‹œ

$\quad\quad$- ์œ ํ˜•15: UFR ๋™์ž‘

3. ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง

์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์€ 1943๋…„ Warren McCulloch๊ณผ Walter Pitts์˜ โ€œArtificial Neural Network ๋ชจ๋ธโ€ ์ œ์•ˆ ์ดํ›„ ์ง€์†์ ์ธ ๊ธฐ์ˆ  ๊ฐœ๋ฐœ์„ ํ†ตํ•˜์—ฌ ๋ฐœ์ „ํ•ด์™”๊ณ , ์ตœ๊ทผ AlphaGo๋ฅผ ํ†ตํ•ด ์ œ2์˜ ์ธ๊ณต์ง€๋Šฅ ๋ถ€ํฅ๊ธฐ๋ฅผ ๋ถˆ๋Ÿฌ์˜จ ๊ธฐ๊ณ„ํ•™์Šต ๋ฐฉ๋ฒ•์ด๋‹ค. ANN์€ BP์ด๋ผ๋Š” ๋‹จ๊ณ„๋ฅผ ํ†ตํ•˜์—ฌ ํ•™์Šต์„ ์ˆ˜ํ–‰ํ•˜๋ฉฐ, ํ•™์Šตํ•  ๋ฐ์ดํ„ฐ๋ฅผ ์ œ๊ณตํ•˜์—ฌ ํ•™์Šต ์‹œํ‚ค๋Š” ์ง€๋„ํ•™์Šต(Supervised Learning) ๋ฐฉ๋ฒ•์„ ์ด์šฉํ•˜์—ฌ ํ•™์Šต์„ ์ˆ˜ํ–‰ํ•œ๋‹ค.

๊ทธ๋ฆผ 3์€ ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์˜ ๊ตฌ์กฐ๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ๊ทธ๋ฆผ 3๊ณผ ๊ฐ™์ด ์ž…๋ ฅ์ธต, ์€๋‹‰์ธต, ์ถœ๋ ฅ์ธต์œผ๋กœ ๊ตฌ์„ฑ๋˜๋ฉฐ $x_{1},\:x_{2}\sim x_{n}$์€ ์ž…๋ ฅ์˜ ๊ฐœ์ˆ˜, $y_{1},\:y_{2}\sim y_{m}$๋Š” ์ถœ๋ ฅ์˜ ๊ฐœ์ˆ˜์ด๋ฉฐ $w_{1,\:1}^{1},\:w_{2,\:1}^{2}\sim w_{n,\:m}^{p}$๋Š” ๊ฐ€์ค‘์น˜(Weight)๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ๊ฐ ์ธต์„ ๊ตฌ์„ฑํ•˜๊ณ  ์žˆ๋Š” ๋…ธ๋“œ๋“ค์€ ํ•„์š”์— ๋”ฐ๋ผ ๊ฐœ์ˆ˜๋ฅผ ์กฐ์ ˆ ํ•  ์ˆ˜ ์žˆ๋‹ค. ๋ณต์ˆ˜์˜ ์€๋‹‰์ธต์„ ๊ฐ€์ง„ ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ๋”ฅ๋Ÿฌ๋‹(Deep Learning)์ด๋ผ๊ณ  ๋ถ€๋ฅธ๋‹ค. ๋˜ํ•œ ์ธต๊ณผ ์ธต์„ ์ด์–ด์ฃผ๋Š” ์„ ์€ ๊ฐ€์ค‘์น˜๋กœ์„œ ๋‹ค์Œ ๋…ธ๋“œ๋กœ ์ „๋‹ฌ๋˜๋Š” ๊ฐ’์„ ์กฐ์ ˆํ•˜์—ฌ, ์ถœ๋ ฅ y๊ฐ€ ์ ๋‹นํ•œ ๊ฐ’์œผ๋กœ ์ถœ๋ ฅ ๋  ์ˆ˜ ์žˆ๋„๋ก ํ•œ๋‹ค[19,20].

๊ทธ๋ฆผ. 3. ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์˜ ๊ตฌ์กฐ

Fig. 3. Structure of ANN

../../Resources/kiee/KIEE.2019.68.9.1039/fig3.png

4. ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์— ์˜ํ•œ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„

4.1 ์ธ๊ณต์‹ ๊ฒฝ๋ง์˜ ๊ตฌ์กฐ์™€ ํ•™์ŠตํŒจํ„ด

์ œ์•ˆํ•˜๋Š” ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์„ ์œ„ํ•œ ์ธ๊ณต์‹ ๊ฒฝํšŒ๋กœ๋ง์˜ ๊ตฌ์กฐ๋Š” 154kV ํ‘œ์ค€ ๋ณ€์ „์†Œ์˜ CB, DS, IED ์ˆ˜์˜ ํ•ฉ์ธ 253๊ฐœ๋กœ ์„ค์ •ํ•˜์˜€์œผ๋ฉฐ, ์ถœ๋ ฅ ์ธต์˜ ๋‰ด๋Ÿฐ์ˆ˜๋Š” SOP์˜ 15๊ฐ€์ง€์˜ ๊ณ ์žฅ์œ ํ˜•์„ ํ‘œ์‹œํ•  ์ˆ˜ ์žˆ๋„๋ก 15๊ฐœ๋กœ ์„ค์ •ํ•˜์˜€๋‹ค[18].

ANN์„ ์œ„ํ•œ ํ•™์ŠตํŒจํ„ด์€ Excel์„ ์ด์šฉํ•ด CSV(Comma Separated Value) ํ˜•์‹์œผ๋กœ ์ž‘์„ฑ๋˜์—ˆ๋‹ค. ๊ทธ๋ฆผ 4๋Š” $72\times 268$ ํ–‰๋ ฌ๋กœ ๊ตฌ์„ฑ๋œ ํ•™์ŠตํŒจํ„ด์˜ ์ผ๋ถ€๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ํ•™์ŠตํŒจํ„ด์€ ์ •์ƒ์ƒํƒœ์™€ SOP์— ๊ทœ์ •๋œ 15๊ฐ€์ง€์˜ ๊ณ ์žฅ์œ ํ˜•์œผ๋กœ ํ•˜์˜€๋‹ค. ์ด ํ•™์ŠตํŒจํ„ด์€ BP๋ฅผ ํ†ตํ•ด ํ•™์Šต๋œ๋‹ค. ๊ทธ๋ฆผ 4์™€ ๊ฐ™์ด CB์™€ IED์˜ ๋™์ž‘์ƒํƒœ(๊ฐœ:0, ํ:1)๋ฅผ Excel์„ ์ด์šฉํ•˜์—ฌ ๋ณด์—ฌ์ฃผ๋Š” ๋ฐ์ดํ„ฐ ํŒŒ์ผ์„ ์ƒ์„ฑํ•˜์—ฌ ์ฝค๋งˆ๋กœ ๊ตฌ๋ถ„๋œ ํ…์ŠคํŠธ ํŒŒ์ผ๋กœ ์ €์žฅํ•˜์˜€๋‹ค. ํ•™์ŠตํŒจํ„ด์—์„œ ์—ด์˜ ์ˆ˜๋Š” 268๊ฐœ์ด๋ฉฐ, ์ด๋Š” 50๊ฐœ์˜ CB, 96๊ฐœ์˜ DS, 107๊ฐœ์˜ IED์˜ ์ด ์ˆ˜์ธ 253๊ฐœ์— ๊ณ ์žฅ์œ ํ˜• 15๊ฐ€์ง€๋ฅผ ํฌํ•จํ•œ ๊ฐœ์ˆ˜๊ฐ€ ๋œ๋‹ค. ์‹ค์ œ๋กœ ํ•™์ŠตํŒจํ„ด์—์„œ ๊ณ ์žฅ์œ ํ˜• 1์€ 6๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 2๋Š” 6๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 3์€ 2๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 4๋Š” 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 5๋Š” 6๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 6์€ 1๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 7์€ ๊ณ ์žฅ์œ ํ˜• 8๊ณผ ๋™์ผํ•œ ํŒจํ„ด์„ ๊ฐ€์ง€๋ฏ€๋กœ ์ƒ๋žตํ•˜์˜€๊ณ , ๊ณ ์žฅ์œ ํ˜• 8์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 9๋Š” 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 10์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 11์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 12๋Š” 22๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 13์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 14๋Š” 3๊ฐ€์ง€, ๊ทธ๋ฆฌ๊ณ  ๊ณ ์žฅ์œ ํ˜• 15๋Š” 1๊ฐ€์ง€๋กœ ์ •ํ•˜์˜€๋‹ค.

๊ทธ๋ฆผ. 4. ํ•™์ŠตํŒจํ„ด์˜ ์ผ๋ถ€

Fig. 4. Part of learning pattern

../../Resources/kiee/KIEE.2019.68.9.1039/fig4.png

4.2 ์‹œํ—˜ํŒจํ„ด

๊ทธ๋ฆผ 5๋Š” $66\times 268$ ํ–‰๋ ฌ๋กœ ๊ตฌ์„ฑ๋œ ์‹œํ—˜ํŒจํ„ด์˜ ์ผ๋ถ€๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ์‹œํ—˜ํŒจํ„ด์€ ์ •์ƒ์ƒํƒœ์™€ SOP์— ๊ทœ์ •๋œ 15๊ฐ€์ง€์˜ ๊ณ ์žฅ์œ ํ˜•์œผ๋กœ ๊ตฌ์„ฑํ•˜์˜€๋‹ค. ์‹ค์ œ๋กœ ์‹œํ—˜ํŒจํ„ด์—์„œ ๊ณ ์žฅ์œ ํ˜• 1์€ 5๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 2๋Š” 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 3์€ 2๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 4๋Š” 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 5๋Š” 6๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 6์€ 1๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 7์€ ์—†๊ณ , ๊ณ ์žฅ์œ ํ˜• 8์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 9๋Š” 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 10์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 11์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 12๋Š” 17๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 13์€ 4๊ฐ€์ง€, ๊ณ ์žฅ์œ ํ˜• 14๋Š” 3๊ฐ€์ง€, ๊ทธ๋ฆฌ๊ณ  ๊ณ ์žฅ์œ ํ˜• 15๋Š” 3๊ฐ€์ง€๋กœ ์ •ํ•˜์˜€๋‹ค.

๊ทธ๋ฆผ. 5. ์‹œํ—˜ํŒจํ„ด์˜ ์ผ๋ถ€

Fig. 5. Part of test pattern

../../Resources/kiee/KIEE.2019.68.9.1039/fig5.png

4.3 ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๋ฐ ๊ฒฐ๊ณผ

4.3.1 C ์–ธ์–ด์— ์˜ํ•œ ๊ตฌํ˜„

C ์–ธ์–ด๋ฅผ ์ด์šฉํ•˜์—ฌ ๋ณ€์ „์†Œ์—์„œ ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ๋Š” 15๊ฐ€์ง€ ๊ณ ์žฅ์œ ํ˜•๊ณผ 1๊ฐ€์ง€์˜ ์ •์ƒ์ƒํƒœ๋ฅผ ํŒ๋ณ„ํ•  ์ˆ˜ ์žˆ๋Š” ANN์„ ์„ค๊ณ„ยท๊ตฌํ˜„ํ•˜์˜€๋‹ค. createNet ํ•จ์ˆ˜๋ฅผ ํ†ตํ•˜์—ฌ ANN์„ ๊ตฌ์„ฑํ•˜๊ณ , ์ž…๋ ฅ ๋ฐ ์ถœ๋ ฅ ๋ฐ์ดํ„ฐ๋ฅผ ๋ถˆ๋Ÿฌ์˜จ ํ›„, for๋ฌธ์„ ํ†ตํ•ด ์ž…๋ ฅ๋ฐ์ดํ„ฐ์˜ ๊ฐ€์ค‘์น˜ ์—…๋ฐ์ดํŠธ๋ฅผ ํ†ตํ•ด ํ•™์Šต์„ ์ˆ˜ํ–‰ํ•˜๋„๋ก ํ•˜์˜€๋‹ค.

๊ทธ๋ฆผ 6์€ C ์–ธ์–ด๋ฅผ ๊ตฌํ˜„ํ•œ ํ”„๋กœ๊ทธ๋žจ์˜ ์ผ๋ถ€๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ๊ทธ๋ฆผ 6์—์„œ ๋ณผ ์ˆ˜ ์žˆ๋“ฏ์ด ANN์˜ ํ•™์ŠตํŒจํ„ด์˜ ๊ณผ์ • ๋ฐ ๊ฒฐ๊ณผ ๋“ฑ์„ ์„ค์ •ํ•  ์ˆ˜ ์žˆ๋‹ค. ์ „์ฒด์ ์œผ๋กœ C ์–ธ์–ด ์ฝ”๋“œ๋Š” ํ•จ์ˆ˜์˜ ์ •์˜ ๋ฐ ์„ ์–ธ, if๋ฌธ๊ณผ for๋ฌธ์„ ํ†ตํ•ด ๋ฐ˜๋ณตํ•™์Šต์„ ์ˆ˜ํ–‰ํ•˜๋ฉฐ ํ•™์Šต์„ ์ง„ํ–‰ํ•˜๋„๋ก ํ•˜์˜€๋‹ค[18].

๊ทธ๋ฆผ. 6. C ์–ธ์–ด๋กœ ๊ตฌํ˜„ํ•œ ํ”„๋กœ๊ทธ๋žจ์˜ ์ผ๋ถ€

Fig. 6. Part of a program implemented in C language

../../Resources/kiee/KIEE.2019.68.9.1039/fig6.png

4.3.2 ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒฐ๊ณผ

๋ณธ ๋…ผ๋ฌธ์—์„œ ์ œ์‹œํ•œ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์„ ์œ„ํ•œ ANN ์ž…๋ ฅ์ธต์˜ ๋‰ด๋Ÿฐ ์ˆ˜๋Š” 253๊ฐœ, ์ถœ๋ ฅ์ธต์˜ ๋‰ด๋Ÿฐ ์ˆ˜๋Š” 15๊ฐœ, ์€๋‹‰์ธต์˜ ์ธต์ˆ˜ 1๊ฐœ, ์€๋‹‰์ธต์˜ ๋‰ด๋Ÿฐ ์ˆ˜๋Š” ์ž…๋ ฅ์˜ ๊ฐœ์ˆ˜์™€ ๋™์ผํ•œ 253๊ฐœ์ด๋‹ค. ์‹œ๋ฎฌ๋ ˆ์ด์…˜์˜ ์กฐ๊ฑด์€, ๋‘ ๊ฐ€์ง€์˜ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜(1,000,000ํšŒ, 2,000,000ํšŒ)์— ๋Œ€ํ•˜์—ฌ ํ•™์Šต๋ฅ (Learning Rate)์„ 0.01, 0.1๋กœ ๊ฐ€๋ณ€ํ•˜์˜€๊ณ  ๊ด€์„ฑํ•ญ(Momentum)์„ 0.1๋ถ€ํ„ฐ 0.9๊นŒ์ง€ 0.2๊ฐ„๊ฒฉ์œผ๋กœ ์„ค์ •ํ•˜์˜€๋‹ค. $72\times 268$์˜ ํ•™์ŠตํŒจํ„ด์œผ๋กœ ํ•™์Šต์‹œ์ผฐ์œผ๋ฉฐ, $66\times 268$์˜ ์‹œํ—˜ํŒจํ„ด์œผ๋กœ ์‹œํ—˜ํ•˜์˜€๋‹ค.

๊ทธ๋ฆผ 7์€ ํ•™์Šต๋ฅ ๊ณผ ๊ด€์„ฑํ•ญ์ด 0.1 ์ผ๋•Œ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜์— ๋”ฐ ๋ฅธ ์ •์ƒ์ƒํƒœ ๋ช‡ ๊ฐ€์ง€ ๊ณ ์žฅ์œ ํ˜•์˜ ํ•™์Šต๊ฒฐ๊ณผ๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. ๊ทธ๋ฆผ 7(a)๊ณผ ๊ฐ™์ด ์ •์ƒ์ƒํƒœ์˜ ๊ฒฝ์šฐ ๋ชจ๋“  ์˜ˆ์ธก๊ฐ’์€ 0์— ๊ฐ€๊น๊ฒŒ ๋‚˜ํƒ€๋‚ฌ์œผ๋ฉฐ, ๊ณ ์žฅ์œ ํ˜• 2๋ฅผ ๋‚˜ํƒ€๋‚ด๋Š” ๋‘ ๋ฒˆ์งธ ํ–‰์˜ ์˜ˆ์ธก๊ฐ’์€ 0.2079 ๋ฐ 0.1446๋กœ ํฌ๊ธฐ๊ฐ€ ์ž‘๊ธฐ ๋•Œ๋ฌธ์— ๊ณ ์žฅ์œผ๋กœ ํŒ๋ณ„ํ•  ์ˆ˜ ์—†์œผ๋ฉฐ, ๊ฐ๊ฐ์˜ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜์—์„œ MSE๋Š” 0.003095์™€ 0.001453 ์ž„์„ ๋‚˜ํƒ€๋ƒˆ๋‹ค. ๊ทธ๋ฆผ 7(b)์™€ ๊ฐ™์ด ๊ณ ์žฅ์œ ํ˜• 1์„ ๋‚˜ํƒ€๋‚ด๋Š” ์ฒซ ๋ฒˆ์งธ ํ–‰์˜ ์˜ˆ์ธก๊ฐ’์ด ๊ฐ๊ฐ 0.9458 ๋ฐ 0.9638๋กœ ๋‚˜ํƒ€๋‚ฌ์œผ๋ฉฐ, ์˜ˆ์ธก๊ฐ’์ด ํฌ๊ธฐ ๋•Œ๋ฌธ์— ๊ณ ์žฅ์œ ํ˜• 1๋กœ ํŒ๋‹จ ํ•  ์ˆ˜ ์žˆ๋‹ค. ๋‘ ๊ฒฐ๊ณผ๋ฅผ ๋น„๊ตํ•˜๋ฉด ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜๊ฐ€ 2,000,000ํšŒ์ผ ๋•Œ ์˜ˆ์ธก๊ฐ’์ด 0.018๋งŒํผ ํฌ๊ธฐ ๋•Œ๋ฌธ์— ์ข€ ๋” ์•ˆ์ •์ ์ด๋ผ๊ณ  ํ•  ์ˆ˜ ์žˆ์œผ๋ฉฐ, ๊ฐ๊ฐ์˜ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜์—์„œ MSE๋Š” 0.000272์™€ 0.000117 ์ž„์„ ๋‚˜ํƒ€๋ƒˆ๋‹ค. ๊ทธ๋ฆผ 7(c)์™€ ๊ฐ™์ด ๊ณ ์žฅ์œ ํ˜• 15๋ฅผ ๋‚˜ํƒ€๋‚ด๋Š” ์—ด๋‹ค์„ฏ ๋ฒˆ์งธ ํ–‰์˜ ์˜ˆ์ธก๊ฐ’์ด ๊ฐ๊ฐ 0.8304 ๋ฐ 0.8997๋กœ ๋‚˜ํƒ€๋‚ฌ์œผ๋ฉฐ, ์˜ˆ์ธก๊ฐ’์ด ํฌ๊ธฐ ๋•Œ๋ฌธ์— ๊ณ ์žฅ์œ ํ˜• 15๋กœ ํŒ๋‹จ ํ•  ์ˆ˜ ์žˆ๋‹ค. ๋‘ ๊ฒฐ๊ณผ๋ฅผ ๋น„๊ตํ•˜๋ฉด ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜๊ฐ€ 2,000,000ํšŒ์ผ ๋•Œ ์˜ˆ์ธก๊ฐ’์ด 0.693๋งŒํผ ํฌ๊ธฐ ๋•Œ๋ฌธ์— ์ข€ ๋” ์•ˆ์ •์ ์ด๋ผ๊ณ  ํ•  ์ˆ˜ ์žˆ์œผ๋ฉฐ, ๊ฐ๊ฐ์˜ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜์—์„œ MSE๋Š” 0.001979์™€ 0.000699 ์ž„์„ ๋‚˜ํƒ€๋ƒˆ๋‹ค.

ํ‘œ 1์€ ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜๊ฐ€ 1,000,000๊ณผ 2,000,000์ผ ๋•Œ ์‹œ๋ฎฌ๋ ˆ์ด์…˜์˜ ์„ฑ๋Šฅ์„ ๋‚˜ํƒ€๋‚ธ๋‹ค. ํ‘œ 1๋กœ๋ถ€ํ„ฐ 15๊ฐ€์ง€ ๊ณ ์žฅ ์œ ํ˜•์— ๋Œ€ํ•˜์—ฌ ํ•™์Šต๋ฐ˜๋ณต ํšŸ์ˆ˜๊ฐ€ 2,000,000๋ฒˆ, ํ•™์Šต๋ฅ ์ด 0.1, ๊ด€์„ฑํ•ญ์ด 0.7 ์ผ๋•Œ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„๊ฐœ์ˆ˜๊ฐ€ 65๊ฐœ๋กœ์„œ ๊ณ ์žฅ์œ ํ˜• 13์ธ ๊ฒฝ์šฐ์˜ 1๊ฐ€์ง€ ํ…Œ์ŠคํŠธ ํŒจํ„ด์„ ์ œ์™ธํ•˜๊ณ  ๋ชจ๋“  ๊ณ ์žฅ์  ํ‘œ์ •์„ ์ •ํ™•ํ•˜๊ฒŒ ์ฐพ์•˜์Œ์„ ์•Œ ์ˆ˜ ์žˆ์—ˆ๋‹ค. ์ „์ฒด์ ์œผ๋กœ ํ•™์Šต๋ฐ˜๋ณต ํšŸ์ˆ˜๋ฅผ ๋Š˜๋ฆด์ˆ˜๋ก ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์˜ ์„ฑ๋Šฅ์ด ๋” ์ข‹์•„์กŒ์œผ๋ฉฐ, ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜๊ฐ€ ์ ์„ ๋•Œ๋Š”, ํ•™์Šต๋ฅ ์€ 0.1 ์ผ๋•Œ ๊ฐ€์žฅ ์ข‹์€ ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค. ๋˜ํ•œ, ๊ด€์„ฑํ•ญ์ด ์ปค์งˆ์ˆ˜๋ก ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์˜ ์ •ํ™•๋„๊ฐ€ ๋†’์•„์กŒ์œผ๋‚˜ ๋„ˆ๋ฌด ํฐ ๊ด€์„ฑํ•ญ์€ ์˜ค์ฐจ๊ฐ€ ์‹ฌํ•˜์—ฌ ์ •ํ™•๋„๊ฐ€ ๋–จ์–ด์ง„๋‹ค๋Š” ๊ฒƒ์„ ์•Œ ์ˆ˜ ์žˆ์—ˆ๋‹ค. ํ•™์Šต๋ฐ˜๋ณต ํšŸ์ˆ˜๋ฅผ ๋Š˜๋ฆฌ๊ณ , ํ•™์Šต๋ฅ ๊ณผ ๊ด€์„ฑํ•ญ์„ ๊ฐ๊ฐ 0.1, 0.7๋กœ ์„ค์ •ํ•˜๋ฉด ์ข€ ๋” ์ •ํ™•ํ•˜๊ฒŒ ๋ชจ๋“  ๊ณ ์žฅ์  ํ‘œ์ •์„ ํ•  ์ˆ˜ ์žˆ์„ ๊ฒƒ์œผ๋กœ ์˜ˆ์ƒ๋œ๋‹ค.

๊ทธ๋ฆผ. 7. ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒฐ๊ณผ

Fig. 7. Simulation result

../../Resources/kiee/KIEE.2019.68.9.1039/fig7_1.png

$\quad\quad\quad\quad\quad$(a) ์ •์ƒ์ƒํƒœ $\quad$ (a) Steady state

../../Resources/kiee/KIEE.2019.68.9.1039/fig7_2.png

$\quad\quad\quad\quad\quad$(b) ๊ณ ์žฅ์œ ํ˜• 1 $\quad$ (b) Fault type 1

../../Resources/kiee/KIEE.2019.68.9.1039/fig7_3.png

$\quad\quad\quad\quad\quad$(c) ๊ณ ์žฅ์œ ํ˜• 15 $\quad$ (c) Fault type 15

ํ‘œ 1. ์‹œ๋ฎฌ๋ ˆ์ด์…˜์˜ ์„ฑ๋Šฅ

Table 1. Simulation performance

ํ•™์Šต๋ฐ˜๋ณตํšŸ์ˆ˜

ํ•™์Šต๋ฅ 

๊ด€์„ฑํ•ญ

์˜ค์ฐจ (MSE)

๊ณ ์žฅ์œ ํ˜•

์‹๋ณ„๊ฐœ์ˆ˜

์˜ˆ์ธก๊ฐ’

1,000,000

0.01

0.1

0.010830

57

0.774372

0.3

0.009486

58

0.798095

0.5

0.008669

58

0.816468

0.7

0.008490

58

0.829183

0.9

0.002635

64

0.914612

0.1

0.1

0.004326

62

0.903698

0.3

0.004293

62

0.905465

0.5

0.003266

63

0.921961

0.7

0.008295

58

0.849640

0.9

0.055557

10

0.150283

2,000,000

0.01

0.1

0.008611

58

0.819974

0.3

0.008519

58

0.826386

0.5

0.008346

58

0.834084

0.7

0.004461

62

0.894233

0.9

0.002206

64

0.937626

0.1

0.1

0.004234

62

0.910383

0.3

0.003239

63

0.926333

0.5

0.003162

63

0.928122

0.7

0.001153

65

0.952971

0.9

0.038486

28

0.403201

5. ๊ฒฐ ๋ก 

๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์€ ์ „๋ ฅ๊ณต๊ธ‰ ์ค‘๋‹จ๊ณผ ๋น„์ƒ์‚ฌํƒœ๋ฅผ ์ผ์œผํ‚ค๊ธฐ ๋•Œ๋ฌธ์— ๊ฐ€๋Šฅํ•œ ์‹ ์†ํ•˜๊ฒŒ ๊ณ ์žฅ์„ ์ฐพ์•„๋‚ด์–ด ์ •์ „๋ณต๊ตฌ์‹œ๊ฐ„๊ณผ ๊ฒฝ์ œ์ ์ธ ์†์‹ค์„ ์ตœ์†Œํ™”ํ•ด์•ผ ํ•œ๋‹ค. ์ตœ๊ทผ, ์Šค๋งˆํŠธํ•œ ์žฅ์ ์ด ์žˆ๋Š” ANN์˜ ๊ณ ์žฅ๋ณต๊ตฌ์™€ ๊ณ ์žฅ์  ํ‘œ์ •์˜ ์ ์šฉ์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋ถ€์ƒํ•˜๊ณ  ์žˆ๋‹ค.

๋ณธ ๋…ผ๋ฌธ์€ ์ˆ˜ํ–‰ ์ค‘์ธ ์ง€๋Šฅ๊ธฐ๋ฒ•์„ ์ด์šฉํ•œ ์ž๋™ํ™” ๋””์ง€ํ„ธ๋ณ€์ „์†Œ ๊ณ ์žฅ๋ณต๊ตฌ๋ฐฉ์•ˆ์˜ ๊ธฐ์ดˆ ์—ฐ๊ตฌ๊ณผ์ œ์˜ ์ผ๋ถ€ ์‚ฐ์ถœ๋ฌผ๋กœ์„œ ์‹ ์†ํ•œ ๊ณ ์žฅ๋ณต๊ตฌ๋ฅผ ์œ„ํ•˜์—ฌ ์„ ํ–‰๋˜์–ด์•ผ ํ•˜๋Š” ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„์„ ์œ„ํ•ด SOP๋ฅผ ๊ณ ๋ คํ•œ ANN์˜ ์ ์šฉ์ด๋‹ค. ๋จผ์ €, ๋ณ€์ „์†Œ์˜ CB, DS, IED ๋“ฑ ๊ตฌ์„ฑ์š”์†Œ์˜ ๋™์ž‘์ƒํƒœ์™€ SOP์˜ ๊ณ ์žฅ์œ ํ˜•์„ ์ด์šฉํ•˜์—ฌ ANN์„ ์„ค๊ณ„ํ•˜์˜€๋‹ค. ๊ตฌ์„ฑํ•œ ํ•™์ŠตํŒจํ„ด์„ BP๋กœ ํ†ตํ•ด ํ•™์Šตํ•œ ํ›„, ์‹œํ—˜ํŒจํ„ด์œผ๋กœ ๋‹ค์–‘ํ•œ ์กฐ๊ฑด์—์„œ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„ ์—ฌ๋ถ€๋ฅผ ์‹œํ—˜ํ•˜์˜€๋‹ค. ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒฐ๊ณผ 2,000,000์˜ ํ•™์Šต๋ฐ˜๋ณต ํšŸ์ˆ˜, 0.1์˜ ํ•™์Šต๋ฅ , 0.7์˜ ๊ด€์„ฑํ•ญ์ผ๋•Œ 65๊ฐœ์˜ ๊ณ ์žฅ์œ ํ˜• ์‹๋ณ„๋กœ 100%์˜ ๊ณ ์žฅ์  ํ‘œ์ •์„ ๋‚˜ํƒ€๋ƒ„์œผ๋กœ์„œ ์ˆ˜ํ–‰ํ•œ ์กฐ๊ฑด ์ค‘์—์„œ์˜ ์ตœ์ ์˜ ANN ์ž„์„ ์ฐพ์„ ์ˆ˜ ์žˆ์—ˆ๋‹ค.

ํ˜„์žฌ, ๋ณ€์ „์†Œ ๊ตฌ์„ฑ์š”์†Œ์˜ ๋™์ž‘์‹œ๊ฐ„๊ณผ ๊ณ ์žฅ๋ฅ  ๋“ฑ์„ ๊ณ ๋ คํ•œ ๊ณ ์žฅ์  ํ‘œ์ •์„ ๊ธฐ๋ฐ˜์œผ๋กœ ์ง€๋Šฅ๊ธฐ๋ฒ•์— ์˜ํ•œ ๋ณ€์ „์†Œ์˜ ๊ณ ์žฅ๋ณต๊ตฌ๋ฐฉ์•ˆ์„ ๋ชจ์ƒ‰ํ•˜๊ณ  ์žˆ๋‹ค.

Acknowledgements

๋ณธ ์—ฐ๊ตฌ๋Š” ํ•œ๊ตญ์ „๋ ฅ๊ณต์‚ฌ์˜ 2016๋…„ ์„ ์ • ๊ธฐ์ดˆ์—ฐ๊ตฌ๊ฐœ๋ฐœ๊ณผ์ œ ์—ฐ๊ตฌ๋น„์— ์˜ํ•ด ์ง€์›๋˜์—ˆ์Œ (๊ณผ์ œ๋ฒˆํ˜ธ : R17XA05-27)

References

1 
C. W. Park, T. W. Kang, et. al., July 2018, Fundamental Study of Fault Restoration Plan using Intelligence Technique for IEC 61850-based Digital Substation, KEPCO 1st Year Report, pp. 1-155Google Search
2 
Y. H. Ahn, et. al., 2011, IED (Intelligent Electronic Device) Development, MOTIE Final Report, pp. 1-155Google Search
3 
KEPCO, Transmission Substation Operation Department,Substation Operation Team, August 2013, Breakdown cases of substation Vol. 2, Human Resources Development Center, Engineering education team, pp. 1-517Google Search
4 
KEPCO, 2016, Transmission & Substation Standard Operation Procedure, KEPCO Transmission & Transformation, pp. 1-77Google Search
5 
P. H. Cho, J. Y. Kim, B. H. Lee, S. D. Jeon, 2015, A Study of 154 kV Transmission & Substation Standard Operating Procedures, in Proc. of KIEE Summer Conference, pp. 443-444Google Search
6 
P. H. Cho, H. G. Kim, W. J. Park, 2017, A Study of major fault causes based on analysis of substation fault, in Proc. of KIEE Summer Conference, pp. 435-436Google Search
7 
KEPCO, 2016, Plans to Build Automatic Fault Restoration System for Substation, KEPCO Transmission & Transformation, pp. 1-5Google Search
8 
J. B. Ahn, K. M. Lee, C. W. Park, et. al., 2016, A Measures of Substation Automatic Restoration Support System using Artificial Intelligence, in Proc. of KIEE PES Autumn Conference, pp. 195-196Google Search
9 
A. P. Alves da Silva, A. H. F. Insfran, P. M. da Silveira, G. Lambert-Torres, 1996, Neural networks for fault location in substations, IEEE Transactions on Power Delivery, Vol. 11, No. 1, pp. 234-239DOI
10 
Farzad Dehghani, Hamid Nezami, 2013, A new fault location technique on radial distribution systems using artificial neural network, in Proc. of the 22nd International Conference and Exhibition on Electricity Distribution (CIRED 2013), pp. 375-379DOI
11 
Farzad Dehghani, Fereydoun Khodnia, Esfandiar Dehghan, June 2017, Fault location of unbalanced power distribution feeder with distributed generation using neural networks, in Proc. of the 24th International Conference & Exhibition on Electricity Distribution (CIRED 2017), pp. 1134-1137DOI
12 
Bon Nhan Nguyen, Anh Huy Quyen, Phuc Huu Nguyen, Trieu Ngoc Ton, 2017, Wavelet-based Neural Network for recognition of faults at NHABE power substation of the Vietnam power system, in Proc. of the 2017 International Conference on System Science and Engineering, pp. 165-168DOI
13 
I. D. Park, S. H. Lee, S. K. Kim, July 2009, Mother Wavelet Transform Suitable to Fault Method Algorithm, in Proc. of KIEE Summer Conference, pp. 62-63Google Search
14 
C. W. Park, K. M. Lee, 2015, A Study on Digital Fault Locator for Transmission Line, KIEE, Vol. 64p, No. 4, pp. 291-296DOI
15 
C. W. Park, T. W. Kang, et. al., July 2018, Neural Network Using Weka for Fault Location of Substation, in Proc. of KIEE Summer Conference, pp. 1-58Google Search
16 
H. J. Lee, 1996, A Restoration Aid Expert System for Distribution Substations, IEEE Transactions on Power Delivery, Vol. 11, No. 4, pp. 596-603DOI
17 
J. H. Lee, B. H. Lee, S. D. Jeon, D. W. Kim, July 2010, A Study of GIS Fault Area Detection Algorithm by SCADA System, in Proc. of KIEE Summer Conference, pp. 1687-1688Google Search
18 
C. W. Park, T. W. Kang, et. al., 2018, Fundamental Study of Fault Restoration Plan using Intelligence Technique for IEC 61850-based Digital Substation, KEPRI 2nd Year Report, pp. 1-155Google Search
19 
Aaron Courville, et. al., 2016, Deep Learning, Google Books, pp. 1-785Google Search
20 
Ian H. Witten, Eibe Frank, Mark A. Hall, Christopher J. Pal, 2017, Data Mining, Practical Machine Learning Tools and Techniques, 4th Ed, ELSEVIER, MORGAN KAUFMANN, pp. 1-656Google Search
21 
I. S. Oh, January 2019, Machine Learning, Hanbit Academy, Inc., pp. 1-664Google Search

์ €์ž์†Œ๊ฐœ

์ด๊ฒฝ๋ฏผ(Kyung-Min Lee)
../../Resources/kiee/KIEE.2019.68.9.1039/au1.png

He was born in Korea in 1990.

He received his B.S. and M.S. degrees in Electrical Engineering from Gangneung- Wonju National University, Wonju, Korea, in 2014 and 2017.

At present, he is working on his Ph.D in the Department of Electrical Engineering at Gangneung-Wonju National University.

He is a teaching assistant at Gangneung-Wonju National University, since 2018.

His research interests include Smartgrid, LVDC, Microgrid, RES, PMU, AI application of power system, power system modeling & control, and power system protection. He is a member of the KIEE, KIIEE, and IEEE.

Tel : 033-760-8796

Fax : 033-760-8781

E-mail : point2529@naver.com

๋ฐ•์ฒ ์›(Chul-Won Park )
../../Resources/kiee/KIEE.2019.68.9.1039/au2.png

He was born in Korea in 1961.

He received his B.S., M.S. and Ph.D. degrees in Electrical Engineering from Sungkyunkwan University, Seoul, Korea, in 1988, 1990, and 1996, respectively.

From 1989 to 1993 he was an associate researcher at Lucky GoldStar Industrial Systems.

From 1993 to 1996, he was a senior researcher at PROCOM system and lecturer at S.K.K. University.

At present, he is a professor in the Department of Electrical Engineering at Gangneung-Wonju National University, since 1997.

His research interests include IED, SAS, Hybrid AC-DC power grid, RES, PMU, AI application to power grid, power grid modeling & control, and computer application in power grid.

He is a member of the KIEE, KIIEE, KIPE, and IEEE. He is president of PSPES since 2018.

Dr. Park was awarded the Paper Prize of KIEE in 2010, the Paper Prize of the KOFST in 2017, and an Academic Prize of KIIEE in 2018.

Tel : 033-760-8786

Lab : 033-760-8796

Fax : 033-760-8781

E-mail : cwpark1@gwnu.ac.kr