Mobile QR Code QR CODE

Journal of the Korea Concrete Institute

J Korea Inst. Struct. Maint. Insp.
  • Indexed by
  • Korea Citation Index (KCI)




์ธ๊ณต์‹ ๊ฒฝ๋ง, ์†์ƒํ‰๊ฐ€, ํšŒ์ „๊ฐ•์„ฑ, ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ
Artificial neural network, Damage detection, Rotational stiffness, Steel moment frame

1. ์„œ ๋ก 

๊ฑด์ถ•๋ฌผ์„ ๊ตฌ์„ฑํ•˜๋Š” ๊ตฌ์กฐ ๋ถ€์žฌ ๋ฐ ์š”์†Œ๋Š” ์ค‘๋ ฅ, ์ง€์ง„, ๋ฐ”๋žŒ ๋“ฑ๊ณผ ๊ฐ™์€ ์™ธ๋ถ€ ํ•˜์ค‘์— ์ €ํ•ญํ•˜์—ฌ ๊ฑด์ถ•๋ฌผ์˜ ํ˜•ํƒœ ๋ฐ ๊ณต๊ฐ„์„ ์œ  ์ง€ํ•˜๋Š” ์—ญํ• ์„ ์ˆ˜ํ–‰ํ•œ๋‹ค. ์ด๋ฅผ ํ†ตํ•ด ๊ฑด๋ฌผ์˜ ๊ณ ์œ  ๊ธฐ๋Šฅ์ด ์ œ๊ณต ๋œ๋‹ค. ์ด๋“ค์˜ ์„ฑ๋Šฅ์€ ๊ตฌ์กฐ๋ฌผ์˜ ์•ˆ์ „๊ณผ ๋ฐ€์ ‘ํ•œ ๊ด€๋ จ์ด ์žˆ๊ธฐ ๋•Œ ๋ฌธ์— โ€˜์ดˆ๊ณ ์ธต ๋ฐ ์ง€ํ•˜์—ฐ๊ณ„ ๋ณตํ•ฉ๊ฑด์ถ•๋ฌผ ์žฌ๋‚œ๊ด€๋ฆฌ์— ๊ด€ํ•œ ํŠน๋ณ„ ๋ฒ•โ€™, โ€˜์ง€์ง„ยทํ™”์‚ฐ์žฌํ•ด๋Œ€์ฑ…๋ฒ•โ€™์—์„œ๋Š” ๊ฑด๋ฌผ์˜ ๊ณ„์ธก ๊ด€๋ฆฌ๋ฅผ ๊ฐ•์ œ ํ™”ํ•˜๊ณ  ์žˆ๋‹ค. ์ด๋Š” ๊ฐ€์†๋„์™€ ๊ฐ™์€ ๊ตฌ์กฐ๋ฌผ์˜ ๊ตฌ์กฐ์‘๋‹ต์„ ๊ณ„์ธก ํ•˜์—ฌ ๊ตฌ์กฐ๋ฌผ์˜ ์•ˆ์ „ ๋ฐ ์œ„ํ—˜์„ ํŒ๋‹จํ•˜๊ณ  ์‹ ์†ํ•œ ๋Œ€์ฒ˜๋ฅผ ํ†ตํ•ด ํ”ผํ•ด๋ฅผ ์ตœ์†Œํ™”ํ•˜๋Š”๋ฐ ๋ชฉ์ ์„ ๋‘”๋‹ค.

๊ตฌ์กฐ๋ฌผ์˜ ๋ฐ˜์‘์€ ์ž‘์šฉํ•˜๋Š” ํ•˜์ค‘์˜ ์‹œ๊ฐ„์ , ๊ณต๊ฐ„์  ํŠน์„ฑ์— ๋”ฐ๋ผ ๋‹ฌ๋ผ์ง€๊ธฐ ๋•Œ๋ฌธ์— ๊ตฌ์กฐ๋ฌผ์˜ ๋ฐ˜์‘ ๊ณ„์ธก ๋ฐ์ดํ„ฐ์˜ ๋ถ„์„ ์—† ์ด ๋ฐ˜์‘ ํฌ๊ธฐ๋งŒ์œผ๋กœ ๊ตฌ์กฐ๋ฌผ์˜ ์•ˆ์ „ ๋ฐ ์†์ƒ์„ ํ‰๊ฐ€ํ•˜๊ธฐ์—๋Š” ํ•œ๊ณ„๊ฐ€ ์žˆ๋‹ค. ๊ทธ๋ž˜์„œ ๊ตฌ์กฐ๋ฌผ์˜ ์†์ƒ ์—ฌ๋ถ€๋ฅผ ์˜ˆ์ธกํ•˜๋Š”๋ฐ ๊ตฌ์กฐ ๋ฌผ์˜ ๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ๊ณผ ๊ฐ™์€ ๋™์  ํŠน์„ฑ์ด ์ฃผ๋กœ ํ™œ์šฉ๋œ ๋‹ค(Yoo and Lee, 2013; Yoo, 2014; Kim et al., 2003; Peeters et al., 2001; Doebling et al., 1998; Fan and Qiao, 2011). ์ด๋Ÿฌํ•œ ๋™์  ํŠน์„ฑ์€ ์™ธ๋ถ€ ํ•˜์ค‘์˜ ์‹œ๊ฐ„์ , ๊ณต๊ฐ„์  ํŠน์„ฑ๊ณผ๋Š” ์ƒ๊ด€์—†์ด ๊ตฌ์กฐ๋ฌผ์˜ ๊ฐ•์„ฑ๊ณผ ์งˆ๋Ÿ‰ ๋“ฑ์— ์˜ํ•ด ๊ฒฐ์ •๋˜๋Š” ๊ณ ์œ  ํŠน์„ฑ์ด๊ธฐ ๋•Œ ๋ฌธ์— ์†์ƒ์ด ์—†์œผ๋ฉด ๊ตฌ์กฐ๋ฌผ์˜ ๋™์  ํŠน์„ฑ์€ ๋ณ€ํ™”๊ฐ€ ์—†๋Š” ๋ฐ˜๋ฉด ์—, ์†์ƒ์ด ๋ฐœ์ƒํ•˜๋ฉด ๋™์  ํŠน์„ฑ์€ ๋ณ€ํ•˜๊ฒŒ ๋œ๋‹ค.

์ผ๋ฐ˜์ ์œผ๋กœ ๊ฑด์ถ•๋ฌผ์— ์†์ƒ์ด ๋ฐœ์ƒํ•˜๋Š” ๊ฒฝ์šฐ๋Š” ์ง€์ง„, ํƒœํ’ ๊ณผ ๊ฐ™์ด ๊ฑด์ถ•๋ฌผ์— ๊ณผ๋„ํ•œ ํ•˜์ค‘์ด ์ž‘์šฉํ•˜๊ฑฐ๋‚˜, ์‹œ๊ฐ„ ๊ฒฝ๊ณผ์— ๋”ฐ ๋ฅธ ๋…ธํ›„ํ™”๊ฐ€ ์ง„ํ–‰๋œ ๊ฒฝ์šฐ๊ฐ€ ์žˆ๋‹ค. ๊ตฌ์กฐ์  ์†์ƒ์€ ์™ธ๋ถ€ ํž˜์— ๋Œ€ ํ•œ ์ €ํ•ญ ์„ฑ๋Šฅ์ด ์ €ํ•˜๋œ ๊ฒƒ์„ ์˜๋ฏธํ•˜๋ฉฐ, ์ด๋Š” ๊ตฌ์กฐ ์š”์†Œ์˜ ๊ฐ•์„ฑ ์ €ํ•˜๋กœ ์ •์˜๋  ์ˆ˜ ์žˆ๋‹ค. ๊ฐ•์„ฑ ์ €ํ•˜๊ฐ€ ๋ฐœ์ƒํ•˜๋ฉด ๊ตฌ์กฐ๋ฌผ์˜ ๋ณ€ํ˜• ๋Ÿ‰ ๋ฐ ๊ณ ์œ ์ง„๋™์ˆ˜ ๋“ฑ์ด ์ฆ๊ฐ€ํ•˜๊ฒŒ ๋œ๋‹ค. ๋”ฐ๋ผ์„œ ๊ฐ€์†๋„ ์‘๋‹ต์„ ์ด์šฉํ•ด ๊ตฌ์กฐ๋ฌผ์˜ ๋™ํŠน์„ฑ์„ ๋ถ„์„ํ•˜๋ฉด ๊ตฌ์กฐ๋ฌผ์˜ ์†์ƒ ์—ฌ๋ถ€๋ฅผ ํŒ๋‹จํ•  ์ˆ˜ ์žˆ๋‹ค.

๊ฐ€์†๋„๊ณ„๋Š” ๊ตฌ์กฐ๋ฌผ์˜ ๊ฐ€์†๋„ ์‘๋‹ต์„ ํŽธ๋ฆฌํ•˜๊ณ  ์‹ ๋ขฐ์„ฑ ์žˆ๊ฒŒ ์–ป์„ ์ˆ˜ ์žˆ๋Š” ๊ณ„์ธก๊ธฐ ์ค‘์— ํ•˜๋‚˜์ด๋‹ค. ๊ตฌ์กฐ๋ฌผ ๋‚ด ์—ฌ๋Ÿฌ ์ธต์— ์ด๋ฅผ ์„ค์น˜ํ•˜์—ฌ ๊ตฌ์กฐ๋ฌผ์˜ ๊ฐ€์†๋„ ์‘๋‹ต์„ ์ˆ˜์ง‘ํ•˜์—ฌ ๊ตฌ์กฐ๋ฌผ์˜ ๊ณ ์œ ์ง„ ๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ, ๊ฐ์‡ ๋น„ ๋“ฑ์„ ์–ป์„ ์ˆ˜ ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ด๋“ค์€ ๊ตฌ ์กฐ๋ฌผ ๋ ˆ๋ฒจ์˜ ํŠน์„ฑ์น˜๋ฅผ ๋‚˜ํƒ€๋‚ด๋Š”๋ฐ ์œ ์šฉํ•˜์ง€๋งŒ, ๊ตญ๋ถ€์ ์ธ ์† ์ƒ์„ ํƒ์ง€ํ•˜๊ธฐ์—๋Š” ํ•œ๊ณ„๊ฐ€ ์žˆ๋‹ค. ์ฆ‰, ์ ‘ํ•ฉ๋ถ€์™€ ๊ฐ™์€ ๊ตญ๋ถ€์ ์ธ ์†์ƒ์€ ๊ตฌ์กฐ๋ฌผ์˜ ๋™ํŠน์„ฑ์— ๋Œ€ํ•œ ์˜ํ–ฅ์ด ์ ๊ธฐ ๋•Œ๋ฌธ์— ์ ‘ํ•ฉ๋ถ€ ์†์ƒ์— ๋”ฐ๋ฅธ ๊ตฌ์กฐ๋ฌผ์˜ ๋™ํŠน์„ฑ ๋ณ€ํ™”๋งŒ์œผ๋กœ ์†์ƒ์˜ ์œ„์น˜์™€ ์ • ๋„๋ฅผ ์˜ˆ์ธกํ•˜๊ธฐ์—๋Š” ํ•œ๊ณ„๊ฐ€ ์žˆ๋‹ค(Wang and Deng, 1999; Kim and Melhem, 2004). ์ด๋Ÿฌํ•œ ํ•œ๊ณ„์ ์„ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด Curvature Mode Shapes, Flexibility, Modal Strain Energy (MSE)๋“ฑ๊ณผ ๊ฐ™ ์€ ๋‹ค์–‘ํ•œ ์ง€ํ‘œ์˜ ๋ณ€ํ™”๋ฅผ ๋ถ„์„ํ•˜์—ฌ ์†์ƒ์˜ ์œ„์น˜ ๋ฐ ํฌ๊ธฐ๋ฅผ ์˜ˆ ์ธกํ•˜๋Š” ๊ธฐ์ˆ ์ด ์ œ์‹œ๋˜์—ˆ๋‹ค(Pandey et al., 1991; Shi et al., 2000; Pandey and Biswas, 1994; Yan et al., 2010).

๊ตฌ์กฐ๋ฌผ์— ์„ผ์„œ๋ฅผ ๋งŽ์ด ์„ค์น˜ํ• ์ˆ˜๋ก ๋งŽ์€ ์ •๋ณด๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ ๊ธฐ ๋•Œ๋ฌธ์— ๋ณด๋‹ค ๋” ์ •ํ™•ํ•œ ๋ถ„์„์ด ๊ฐ€๋Šฅํ•˜์—ฌ ์†์ƒํ‰๊ฐ€์˜ ์‹ ๋ขฐ ๋„๋ฅผ ๋†’์ผ ์ˆ˜ ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์„ผ์„œ๋ฅผ ๋งŽ์ด ์„ค์น˜ํ• ์ˆ˜๋ก ๋น„์šฉ์ด ์ฆ ๊ฐ€ํ•˜๊ณ  ์œ ์ง€๊ด€๋ฆฌ๊ฐ€ ์–ด๋ ค์šด ๋‹จ์ ์ด ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ ํ˜„์‹ค์—์„œ๋Š” ํ•œ์ •๋œ ์„ผ์„œ๋งŒ ์„ค์น˜ํ•  ์ˆ˜ ๋ฐ–์—์„œ ์—†์œผ๋ฉฐ, ์ด๋Š” ์†์ƒํ‰๊ฐ€์˜ ์‹  ๋ขฐ์„ฑ์„ ์ €ํ•˜์‹œํ‚ค๋Š” ์›์ธ์ด ๋œ๋‹ค(Li et al., 2008; Law et al., 1998; Shi et al., 2000; Park and Park, 2003).

์†์ƒํ‰๊ฐ€๊ธฐ๋ฒ•์˜ ์˜ˆ์ธก ์ •ํ™•๋ฅ ์„ ๋†’์ด๊ธฐ ์œ„ํ•ด ์ตœ์ ํ™” ์•Œ๊ณ ๋ฆฌ ์ฆ˜์„ ํ™œ์šฉํ•œ ์—ฐ๊ตฌ๊ฐ€ ์ˆ˜ํ–‰๋˜์—ˆ๋‹ค. Cha and Buyukozturk (2015) ๋Š” ์ตœ์ ํ™” ์•Œ๊ณ ๋ฆฌ์ฆ˜์œผ๋กœ Implicit Redundant Representration Genetic Algorithm (IRR GA)์„ ์‚ฌ์šฉํ•œ ์†์ƒํ‰๊ฐ€๊ธฐ๋ฒ•์„ ์ œ์‹œ ํ•˜๊ณ  ์ด๋ฅผ ์ฒ ๊ณจ ๊ตฌ์กฐ๋ฌผ ์˜ˆ์ œ์— ์ ์šฉํ•˜์—ฌ ํ•ด์„์ ์œผ๋กœ ๊ฒ€์ฆํ•˜ ์˜€๋‹ค. ์ œ์‹œ๋œ ๊ธฐ๋ฒ•์€ ํ•œ์ •๋œ ๋ฐ์ดํ„ฐ๋งŒ์œผ๋กœ๋„ ๋‹ค์ˆ˜์˜ ๊ฒฝ๋ฏธํ•œ ์†์ƒ ์œ„์น˜ ๋ฐ ์ •๋„๋ฅผ ํƒ์ง€ํ•  ์ˆ˜ ์žˆ์—ˆ๋‹ค. ์†์ƒ์„ ํ‰๊ฐ€ํ•˜๊ธฐ ์œ„ํ•ด MSE๊ฐ€ ์ฃผ์š” ์ง€ํ‘œ๋กœ ์ด์šฉ๋˜์—ˆ๋‹ค. Kang et al. (2012)๋Š” Particle Swarm Optimization (PSO)์™€ Artificial Immune System์„ ์ ‘ ๋ชฉํ•˜์—ฌ ํ–ฅ์ƒ๋œ PSO๋ฅผ ํ™œ์šฉํ•œ ์†์ƒํ‰๊ฐ€๊ธฐ๋ฒ•์„ ์ œ์‹œํ•˜์˜€๋‹ค. ๊ณ ์œ ์ง„๋™์ˆ˜์™€ ๋ชจ๋“œํ˜•์ƒ ๋ฐ์ดํ„ฐ๊ฐ€ ์†์ƒํ‰๊ฐ€๋ฅผ ์œ„ํ•œ ๋ชฉ์ ํ•จ์ˆ˜ ํ‰๊ฐ€์— ํ™œ์šฉ๋˜์—ˆ๋‹ค. ๋ณด, ํŠธ๋Ÿฌ์Šค ๊ตฌ์กฐ๋ฌผ ์˜ˆ์ œ์— ์ด๋ฅผ ์ ์šฉํ•˜์—ฌ ๊ธฐ์กด ์ตœ์ ํ™”๊ธฐ๋ฒ•(PSO)์— ์˜ํ•œ ๊ฒฐ๊ณผ์™€ ๋น„๊ต ๊ฒ€์ฆํ•˜์˜€๋‹ค. ์†์ƒ ์œ„์น˜ ๋ฐ ํฌ๊ธฐ๋ฅผ ํšจ๊ณผ์ ์œผ๋กœ ํƒ์ง€ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค. Seyedpoor (2012)๋Š” ๋‘ ๋‹จ๊ณ„๋กœ ์ด๋ฃจ์–ด์ง„ ์†์ƒํ‰๊ฐ€๊ธฐ๋ฒ•์„ ์ œ์‹œํ•˜๊ณ , ๋ณด ๋ฐ ํŠธ๋Ÿฌ์Šค ๊ตฌ์กฐ๋ฌผ ์˜ˆ์ œ๋ฅผ ํ†ตํ•ด ํ•ด์„์ ์œผ๋กœ ๊ฒ€์ฆํ•˜์˜€๋‹ค. ์ด๋Š” ๋จผ์ € MSE์˜ ๋ณ€ํ™”๋ฅผ ํ†ตํ•ด ๊ตฌ์กฐ๋ฌผ์˜ ์†์ƒ ์œ„์น˜๋ฅผ ํƒ์ง€ํ•˜๊ณ , ๊ทธ ๊ฒฐ๊ณผ์— PSO๊ธฐ๋ฒ•์— ์ ์šฉํ•˜์—ฌ ์†์ƒ ์ˆ˜์ค€์„ ์˜ˆ์ธกํ•˜๋„๋ก ํ•˜์˜€๋‹ค. Perera et al. (2007)์€ Modal Flexibility๊ณผ Modal Parameter๋ฅผ ๋™์‹œ์— ๊ณ ๋ คํ•œ ๋‹ค๋ชฉ์  ์ตœ์ ํ™”๊ธฐ๋ฒ•์„ ์ œ์‹œํ•˜์˜€์œผ๋ฉฐ, ๋ณด ๊ตฌ์กฐ ๋ฌผ ์˜ˆ์ œ์— ์ ์šฉํ•˜์—ฌ ํ•ด์„์ ์œผ๋กœ ๊ฒ€์ฆํ•˜์˜€๋‹ค. ๋‹ค์ˆ˜์˜ ์ง€ํ‘œ๋ฅผ ๊ณ ๋ คํ•จ์œผ๋กœ์จ ๊ฐ ์ง€ํ‘œ๊ฐ€ ๊ฐ€์ง€๋Š” ๋‹จ์ ์„ ๋ณด์™„ํ•˜๋„๋ก ํ•˜์˜€๋‹ค.

๋ณธ ์—ฐ๊ตฌ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง(Artificial Neural Network, ANN)์„ ์ด์šฉํ•ด ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ๊ณจ์กฐ์˜ ์ ‘ํ•ฉ๋ถ€ ์†์ƒ์„ ์˜ˆ์ธกํ•˜๋Š” ๊ธฐ๋ฒ•์„ ์ œ ์•ˆํ•œ๋‹ค. ์ €์ž๋Š” ์„ ํ–‰ ์—ฐ๊ตฌ(Kim and Choi, 2016)๋ฅผ ํ†ตํ•ด ๊ธฐ๋‘ฅ ์—ด์— ์„ค์น˜๋œ ๋ณ€ํ˜•๋ฅ ์„ผ์„œ๋กœ๋ถ€ํ„ฐ์˜ ์‘๋‹ต๊ฐ’์„ ์ด์šฉํ•ด ๊ตฌ์กฐ๋ฌผ์˜ ํšก์‘๋‹ต(ํšก๋ณ€์œ„, ํšก๊ฐ€์†๋„)์„ ์˜ˆ์ธกํ•˜๋Š” ๊ธฐ๋ฒ•์„ ์ œ์‹œํ•˜์˜€๋‹ค. ์„  ํ–‰ ์—ฐ๊ตฌ ๊ฒฐ๊ณผ๋ฅผ ์ด์šฉํ•ด ๊ตฌ์กฐ๋ฌผ๋ ˆ๋ฒจ์˜ ๋ฐ์ดํ„ฐ(๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ ๋“œํ˜•์ƒ ๋“ฑ)์™€ ๋ถ€์žฌ๋ ˆ๋ฒจ์˜ ๋ฐ์ดํ„ฐ(ํœจ๋ชจ๋ฉ˜ํŠธ)๋ฅผ ์กฐํ•ฉํ•œ ์†์ƒํ‰ ๊ธฐ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค. ๊ตฌ์กฐ๋ฌผ์˜ ๋™ํŠน์„ฑ(๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ)๊ณผ ๋ชจ๋ฉ˜ํŠธ ์‘๋‹ต์ด ์ž…๋ ฅ๊ฐ’์œผ๋กœ ์‚ฌ์šฉ๋˜๋ฉฐ, ์ง€์  ๋ฐ ๋ณด-๊ธฐ๋‘ฅ ์ ‘ํ•ฉ๋ถ€ ์˜ ํšŒ์ „๊ฐ•์„ฑ์˜ ์†์ƒ์ง€ํ‘œ๊ฐ€ ์ถœ๋ ฅ๊ฐ’์œผ๋กœ ์„ค์ •๋œ๋‹ค. ์†์ƒ์ง€ํ‘œ๋Š” ์†์ƒ์ˆ˜์ค€์„ ๋‚˜ํƒ€๋‚ธ๋‹ค. 5์ธต ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ๊ณจ์กฐ ์˜ˆ์ œ์˜ ์ˆ˜์น˜ํ•ด์„ ์„ ํ†ตํ•ด ํ›ˆ๋ จ ๋ฐ ๊ฒ€์ฆ์šฉ ๋ฐ์ดํ„ฐ๋ฅผ ์ƒ์„ฑํ•œ๋‹ค. ์˜ˆ์ œ ๊ฒ€์ฆ์„ ํ†ตํ•ด ์ œ์‹œํ•˜๋Š” ๊ธฐ๋ฒ•์˜ ์†์ƒ ์œ„์น˜ ๋ฐ ์ˆ˜์ค€ ์˜ˆ์ธก ์ •ํ™•์„ฑ์„ ๋ถ„์„ํ•œ๋‹ค.

2. ์ธ๊ณต์‹ ๊ฒฝ๋ง

์ธ๊ณต์‹ ๊ฒฝ๋ง์˜ ๊ตฌ์กฐ๋Š” Fig. 1๊ณผ ๊ฐ™์ด ์ž…๋ ฅ์ธต(Input layer), ์€ ๋‹‰์ธต(Hidden layer), ์ถœ๋ ฅ์ธต(Output layer)๋กœ ๊ตฌ์„ฑ๋œ๋‹ค. ์ž…๋ ฅ์ธต ์—์„œ๋Š” ์ž…๋ ฅ์ธต์„ ์ด๋ฃจ๋Š” ๋…ธ๋“œ์˜ ๊ฐœ์ˆ˜๋งŒํผ์˜ ๊ฐ’์ด ์ž…๋ ฅ๋œ๋‹ค. ์ฐธ๊ณ ๋กœ Fig. 1์— ํ‘œ์‹œ๋œ ์ž…๋ ฅ์ธต์„ ์ด๋ฃจ๋Š” ๋…ธ๋“œ๋Š” 3๊ฐœ์ด๋‹ค. ๊ฐ ์ž…๋ ฅ๊ฐ’์€ ๊ฐ€์ค‘์น˜๋ฅผ ํ†ตํ•œ ์„ ํ˜•์กฐํ•ฉ์„ ํ†ตํ•ด ๊ฐ’์ด ๋ณ€ํ™˜๋˜์–ด ์€ ๋‹‰์ธต์„ ์ด๋ฃจ๋Š” ๊ฐ ๋…ธ๋“œ์— ์ „๋‹ฌ๋œ๋‹ค. ์ด๋Ÿฌํ•œ ๊ฐ’์€ ์€๋‹‰์ธต์— ์œ„ ์น˜ํ•œ ํ™œ์„ฑํ•จ์ˆ˜์˜ ์ž…๋ ฅ๊ฐ’์ด ๋œ๋‹ค. ํ™œ์„ฑํ•จ์ˆ˜์—๋Š” step ํ•จ์ˆ˜, sigmoidํ•จ์ˆ˜, ReLU ํ•จ์ˆ˜ ๋“ฑ ๋‹ค์–‘ํ•œ ์ข…๋ฅ˜๊ฐ€ ์žˆ๋‹ค. ํ™œ์„ฑํ•จ์ˆ˜์˜ ์ถœ๋ ฅ๊ฐ’์€ ๊ฐ๊ฐ ๋‹ค์‹œ ๊ฐ€์ค‘์น˜๊ฐ€ ๊ณฑํ•ด์ง€๊ณ  ์€๋ฆญ์ธต ๋‚ด์˜ ๋‹ค๋ฅธ ๋…ธ ๋“œ์˜ ์ถœ๋ ฅ๊ฐ’๊ณผ ํ•ฉํ•ด์ ธ ์ถœ๋ ฅ์ธต์˜ ๋…ธ๋“œ์— ์ „๋‹ฌ๋œ๋‹ค. ์ถœ๋ ฅ์ธต ๋‚ด ์˜ ๊ฐ ๋…ธ๋“œ์˜ ๊ฐ’์€ ์ถœ๋ ฅ๊ฐ’์„ ์˜๋ฏธํ•˜๊ฒŒ ๋œ๋‹ค.

Fig. 1

Organization of Artificial Neural Network(ANN)

JKSMI-22-107_F1.jpg

์‚ฌ์ „์— ์ˆ˜์ง‘๋œ(์ž…๋ ฅ, ์ถœ๋ ฅ) ๋ฐ์ดํ„ฐ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง ๋‚ด์˜ ๊ฐ€์ค‘ ์น˜๊ฐ’์„ ๊ฒฐ์ •ํ•˜๋Š”๋ฐ ํ•™์Šต๋ฐ์ดํ„ฐ๋กœ ํ™œ์šฉ๋œ๋‹ค. ์‚ฌ์ „์— ์ˆ˜์ง‘๋œ (์ž…๋ ฅ, ์ถœ๋ ฅ) ๋ฐ์ดํ„ฐ๋Š” ์ผ์ข…์˜ (์ž…๋ ฅ, ์ •๋‹ต) ๊ด€๊ณ„๋ฅผ ์˜๋ฏธํ•˜๋ฉฐ, ์ธ๊ณต์‹ ๊ฒฝ๋ง์— ๋™์ผํ•œ ์ž…๋ ฅ ๋ฐ์ดํ„ฐ๋ฅผ ์ž…๋ ฅํ•˜์˜€์„ ๋•Œ ์ •๋‹ต๊ณผ ์˜ค์ฐจ๊ฐ€ ์ ์€ ์ถœ๋ ฅ๊ฐ’์„ ์ƒ์„ฑํ•˜๋„๋ก ๋ฐ˜๋ณต๊ณ„์‚ฐ์„ ํ†ตํ•ด ๊ฐ€์ค‘์น˜ ๊ฐ’์„ ๊ฒฐ์ •ํ•œ๋‹ค. ์ด๋Ÿฌํ•œ ํ•™์Šต๊ณผ์ •์„ ํ†ตํ•ด ์ƒ์„ฑ๋œ ์‹ ๊ฒฝ๋ง์€ ์ž… ๋ ฅ๊ฐ’์„ ์•Œ๊ณ  ์žˆ์œผ๋‚˜ ์ •๋‹ต๊ฐ’์€ ๋ชจ๋ฅด๋Š” ์ƒํ™ฉ์—์„œ ์ž…๋ ฅ๊ฐ’์„ ์‹  ๊ฒฝ๋ง์— ์ „๋‹ฌํ•˜์—ฌ ์ •๋‹ต๊ฐ’์„ ์˜ˆ์ธกํ•˜๋Š”๋ฐ ํ™œ์šฉํ•  ์ˆ˜ ์žˆ๋‹ค. ์ธ๊ณต ์‹ ๊ฒฝ๋ง์€ ๊ฑด์ถ•, ํ† ๋ชฉ ๋“ฑ ๋‹ค์–‘ํ•œ ๊ณตํ•™๋ถ„์•ผ์—์„œ ์„ค๊ณ„, ์‹œ์Šคํ…œ ์‹ ๋ณ„, ์†์ƒํ‰๊ฐ€, ๋น„์šฉ ์˜ˆ์ธก, ๊ตฌ์กฐ์ตœ์ ํ™”, ์ง€์ง„ ์˜ˆ์ธก ๋“ฑ์˜ ๋ชฉ์ ์œผ ๋กœ ๋„๋ฆฌ ์‚ฌ์šฉ๋˜๊ณ  ์žˆ๋‹ค(Rafiq et al., 2001; Flood and Kartam, 1994; Wu et al., 1992; Adeli, 2001).

3. ๋ณ€ํ˜•๋ฅ  ๊ธฐ๋ฐ˜ ํšก์‘๋‹ต ์˜ˆ์ธก

์†์ƒํ‰๊ฐ€์—์„œ ์ฃผ๋กœ ์‚ฌ์šฉ๋˜๋Š” ๊ตฌ์กฐ๋ฌผ๋ ˆ๋ฒจ์˜ ๊ฐ€์†๋„์‘๋‹ต์€ ๊ตญ๋ถ€์ ์ธ ์†์ƒ์„ ํŒŒ์•…ํ•˜๋Š”๋ฐ ํ•œ๊ณ„๊ฐ€ ์žˆ๋Š” ๊ฒƒ์œผ๋กœ ์•Œ๋ ค์ ธ ์žˆ ๊ธฐ ๋•Œ๋ฌธ์—(Wang and Deng, 1999; Kim and Melhem, 2004), ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ๋ถ€์žฌ๋ ˆ๋ฒจ์˜ ๋ชจ๋ฉ˜ํŠธ ์‘๋‹ต์„ ํ•จ๊ป˜ ์‚ฌ์šฉํ•˜๋Š” ์†์ƒํ‰ ๊ฐ€๊ธฐ๋ฒ•์„ ์ œ์‹œํ•˜๊ณ ์ž ํ•œ๋‹ค. ์ด๋ฅผ ์œ„ํ•ด ๋ณธ ์ €์ž์˜ ์„ ํ–‰ ์—ฐ๊ตฌ ๊ฒฐ ๊ณผ(Kim and Choi, 2016)์—์„œ ์ œ์‹œ๋œ ๋ณ€ํ˜•๋ฅ  ์‘๋‹ต์„ ์ด์šฉํ•œ ๊ฑด ์ถ•๋ฌผ์˜ ํšก์‘๋‹ต ๊ธฐ๋ฒ•์„ ์‚ฌ์šฉํ•œ๋‹ค. ๋ณธ ์ ˆ์—์„œ๋Š” ์ด์— ๋Œ€ํ•œ ๋‚ด์šฉ ์„ ์š”์•ฝ ์ •๋ฆฌํ•˜์˜€์œผ๋ฉฐ, ์ž์„ธํ•œ ์‚ฌํ•ญ์€ Kim and Choi(2016)์„ ์ฐธ๊ณ ํ•˜๊ธฐ ๋ฐ”๋ž€๋‹ค

3.1. ๊ธฐ๋‘ฅ ๋ถ€์žฌ์˜ ํœจ๋ชจ๋ฉ˜ํŠธ ๋ถ„ํฌ ์˜ˆ์ธก

๊ธฐ๋‘ฅ๋ถ€์žฌ์˜ ํœจ๋ชจ๋ฉ˜ํŠธ ๋ถ„ํฌ๋ฅผ ์ถ”์ •ํ•˜๊ธฐ ์œ„ํ•ด ๋ณ€ํ˜•๋ฅ  ์„ผ์„œ๋Š” Fig. 2์™€ ๊ฐ™์ด ๊ธฐ๋‘ฅ์˜ ์–‘ ๋‹จ๋ถ€์— ์ด 4๊ฐœ์”ฉ ์„ค์น˜๋œ๋‹ค๊ณ  ๊ฐ€์ •ํ•œ ๋‹ค. ๊ธฐ๋‘ฅ์˜ ์–‘๋‹จ๋ถ€์— 2๊ฐœ์”ฉ ๋ณ€ํ˜•๋ฅ  ์„ผ์„œ๋ฅผ ์„ค์น˜ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ๋‹จ๋ถ€๋ฉด์—์„œ์˜ ๋ณ€ํ˜•๋ฅ  ๋ถ„ํฌ๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ๋‹ค. ์‹ (1)๊ณผ ๊ฐ™์ด ๋‹จ๋ฉด ์˜ ๋ณ€ํ˜•๋ฅ  ฮต์— ํƒ„์„ฑ๊ณ„์ˆ˜ E์™€ ๋ถ€์žฌ ๋‹จ๋ฉด๊ณ„์ˆ˜ Z๋ฅผ ๊ณฑํ•˜๋ฉด ํ•ด๋‹น ์œ„์น˜์—์„œ์˜ ๋ชจ๋ฉ˜ํŠธ M๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ๋‹ค.

Fig. 2

Installment of strain sensors

JKSMI-22-107_F2.jpg

(1)
M = ฮต E Z

ํ•œํŽธ, ์ผ๋ฐ˜์ ์œผ๋กœ ๊ธฐ๋‘ฅ ๋ถ€์žฌ ๋‚ด์—์„œ๋Š” ๋ถ„ํฌํ•˜์ค‘ ๋˜๋Š” ํšกํ•˜ ์ค‘์ด ์ž‘์šฉํ•˜์ง€ ์•Š๋Š”๋‹ค๊ณ  ๊ฐ€์ •ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ๊ธฐ๋‘ฅ ๋ถ€์žฌ์˜ ํœจ๋ชจ ๋ฉ˜ํŠธ ๋ถ„ํฌ๋Š” ์–‘ ๋‹จ๋ถ€์˜ ๋ชจ๋ฉ˜ํŠธ๋ฅผ ์„ ํ˜•์œผ๋กœ ์ด์–ด์ฃผ๋ฉด ์–ป์„ ์ˆ˜ ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ ๋ถ€์žฌ ์น˜์ˆ˜ ๋ฐ ํƒ„์„ฑ๊ณ„์ˆ˜ ๊ฐ’์„ ์•Œ๊ณ  ์žˆ๊ณ , ๊ธฐ๋‘ฅ ์–‘ ๋‹จ๋ถ€์— 2๊ฐœ์”ฉ ๋ณ€ํ˜•๋ฅ  ์„ผ์„œ๋ฅผ ์„ค์น˜ํ•˜๋ฉด ํ•ด๋‹น ๋ถ€์žฌ์˜ ํœจ๋ชจ๋ฉ˜ํŠธ ๋ถ„ํฌ๋ฅผ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋‹ค.

3.2. ๊ธฐ๋‘ฅ์—ด์˜ ํšก๋ณ€์œ„ ๋ฐ ํšก๊ฐ€์†๋„ ์˜ˆ์ธก

๊ฑด๋ฌผ ๋‚ด ํŠน์ • ๊ธฐ๋‘ฅ์—ด์„ ์ด๋ฃจ๋Š” ๋ชจ๋“  ๊ธฐ๋‘ฅ์— ๋™์ผํ•œ ๋ฐฉ๋ฒ•์œผ ๋กœ ๋ณ€ํ˜•๋ฅ  ์„ผ์„œ๊ฐ€ ์„ค์น˜๋œ๋‹ค๊ณ  ๊ฐ€์ •ํ•˜๋ฉด, ํ•ด๋‹น ๊ธฐ๋‘ฅ์—ด์˜ ํœจ๋ชจ ๋ฉ˜ํŠธ ๋ถ„ํฌ๋ฅผ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋‹ค. ์ผ๋ฐ˜์ ์ธ ๊ณจ์กฐ์˜ ํšก๋ณ€์œ„๋Š” ํœจ๋ณ€ ํ˜•์— ์˜ํ•ด ์ง€๋ฐฐ๋˜๊ธฐ ๋•Œ๋ฌธ์— ์ถ•๋ ฅ ๋ฐ ์ „๋‹จ๋ ฅ์— ์˜ํ•œ ํšก๋ณ€์œ„ ๊ธฐ ์—ฌ๋„๋Š” ์ž‘์œผ๋ฉฐ, ํœจ๋ชจ๋ฉ˜ํŠธ์— ์˜ํ•œ ํšก๋ณ€์œ„ ๊ธฐ์—ฌ๋„๋งŒ ๊ณ ๋ คํ•ด๋„ ์‹ ๋ขฐํ•  ๋งŒํ•œ ๊ฒฐ๊ณผ๊ฐ’์„ ์–ป์„ ์ˆ˜ ์žˆ๋‹ค(Hibbeler, 2011). ๋”ฐ๋ผ์„œ ๊ธฐ๋‘ฅ์—ด์˜ ํœจ๋ชจ๋ฉ˜ํŠธ ๋ถ„ํฌ๋ฅผ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฒƒ์€ ํ•ด๋‹น ๊ตฌ์กฐ ๋ฌผ์˜ ํšก๋ณ€์œ„ ์‘๋‹ต์„ ์–ป์„ ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•˜๋ฉฐ, ์ด๋Š” ์„ ํ–‰ ์—ฐ๊ตฌ ๋ฅผ ํ†ตํ•ด ๊ฒ€์ฆ๋˜์—ˆ๋‹ค(Kim and Choi, 2016). ์„ ํ–‰ ์—ฐ๊ตฌ(Kim and Choi, 2016)์—์„œ๋Š” ์ฒ˜์ง๊ฐ๋ฒ•(Slope Deflection Method)์„ ์ด ์šฉํ•˜์—ฌ Fig. 3๊ณผ ๊ฐ™์ด ๊ธฐ๋‘ฅ์˜ ํšก์ฒ˜์ง ฮ”๊ณผ ์ฒ˜์ง๊ฐ ฮธ์„ ๊ตฌํ•˜๊ณ , ์ด๋ฅผ ๋ˆ„์ ํ•˜์—ฌ ๊ธฐ๋‘ฅ์—ด์˜ ํšก๋ณ€์œ„๋ฅผ ๊ตฌํ•˜์˜€๋‹ค.

Fig. 3

Slope deflection method

JKSMI-22-107_F3.jpg

ํ•œํŽธ, ์ฒ˜์ง๊ฐ๋ฒ•์— ์˜ํ•ด ์˜ˆ์ธก๋œ ๊ฐ ์ธต์˜ ํšก๋ณ€์œ„ ์‘๋‹ต y(t)์€ ์‹ (2)์™€ ๊ฐ™์ด ์ด์ค‘๋ฏธ๋ถ„์„ ํ•˜๋ฉด ๊ฐ ์ธต์˜ ํšก๋ฐฉํ–ฅ ๊ฐ€์†๋„ ์‘๋‹ต a(t)์„ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋‹ค.

(2)
a ( t ) = โˆ’ y ( t โˆ’ 3 ) + 4 y ( t โˆ’ 2 ) โˆ’ 5 y ( t โˆ’ 1 ) + 2 y ( t ) ( ฮ” t ) 2

์˜ˆ์ธก๋œ ํšก๋ณ€์œ„ ํ˜น์€ ํšก๊ฐ€์†๋„ ์‘๋‹ต ์ด๋ ฅ๊ฐ’์€ ๊ตฌ์กฐ๋ฌผ์˜ ๊ณ  ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ, ๊ฐ์‡ ๋น„ ๋“ฑ๊ณผ ๊ฐ™์€ ๋™ํŠน์„ฑ์„ ํŒŒ์•…ํ•˜๋Š”๋ฐ ํ™œ์šฉ๋  ์ˆ˜ ์žˆ๋‹ค. Brincker et al.(2001)์€ ์‘๋‹ต๋ฐ์ดํ„ฐ๋งŒ์œผ๋กœ ๋ชจ๋‹ฌ ํŒŒ๋ผ๋ฏธํ„ฐ๋ฅผ ์‹๋ณ„ํ•  ์ˆ˜ ์žˆ๋Š” Frequency Domain Decomposition (FDD)์„ ์ œ์‹œํ•˜์˜€์œผ๋ฉฐ, ์ด๋Š” ๊ต๋Ÿ‰ ๋ฐ ๊ฑด๋ฌผ์˜ ์ƒ์‹œ ์ง„๋™ ๋ฐ์ด ํ„ฐ๋กœ๋ถ€ํ„ฐ ๋™ํŠน์„ฑ์„ ์ถ”์ถœํ•˜๋Š”๋ฐ ํ™œ์šฉ๋˜๊ณ  ์žˆ๊ณ  ์žˆ๋‹ค(Weng et al., 2008; Michel et al., 2010). ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ์‹ (2)๋กœ๋ถ€ํ„ฐ ์–ป ์€ ํšก๊ฐ€์†๋„ ์ด๋ ฅ ๋ฐ์ดํ„ฐ๋ฅผ FDD๊ธฐ๋ฒ•์— ์ ์šฉํ•˜์—ฌ ๊ตฌ์กฐ๋ฌผ์˜ ๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ ๊ฐ’์„ ์ถ”์ถœํ•œ๋‹ค.

4. ์˜ˆ์ œ ์ ์šฉ

4.1. ์˜ˆ์ œ ๊ฐœ์š”

๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” Fig. 4์™€ ๊ฐ™์€ 5์ธต 1๊ฒฝ๊ฐ„ ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ๊ณจ์กฐ ์˜ˆ ์ œ๋ฅผ ์‚ฌ์šฉํ•œ๋‹ค. ์ธต๋†’์ด์™€ ๊ฒฝ๊ฐ„ ๊ธธ์ด๋Š” ๊ฐ๊ฐ 3.96 m, 9.14 m์ด ๋‹ค. 3์žฅ์—์„œ ์„ค๋ช…ํ•œ ๋ณ€ํ˜•๋ฅ  ๋ฐ์ดํ„ฐ๋ฅผ ์ด์šฉํ•ด ๊ตฌ์กฐ๋ฌผ์˜ ํšก์‘๋‹ต ์„ ๊ณ„์‚ฐํ•˜๊ธฐ ์œ„ํ•ด ์™ผ์ชฝ ๊ธฐ๋‘ฅ์—ด์— ๊ฐ ์ธต์˜ ๊ธฐ๋‘ฅ๋‹น 4๊ฐœ์”ฉ ๋ณ€ํ˜• ๋ฅ ๊ณ„๊ฐ€ ์„ค์น˜๋˜์–ด ์ด 20๊ฐœ์˜ ์„ผ์„œ๊ฐ€ ์‚ฌ์šฉ๋œ๋‹ค๊ณ  ๊ฐ€์ •ํ•œ๋‹ค.

Fig. 4

Example structure

JKSMI-22-107_F4.jpg

๊ทธ๋ฆฌ๊ณ  ๋ณธ ์˜ˆ์ œ์˜ ๋ถ•๊ดด๋ชจ๋“œ๋Š” ์ง€์ ๊ณผ ๋ณด์—์„œ ํžŒ์ง€๊ฐ€ ๋ฐœ์ƒ ํ•˜๋Š” ๋ณด-ํžŒ์ง€ ๋ถ•๊ดด๋ชจ๋“œ(Beam-hinge collapse mechanism)๋ฅผ ๊ฐ€์ง„๋‹ค๊ณ  ๊ฐ€์ •ํ•˜๊ณ , ์ด๋ฅผ ํ†ตํ•ด ์ง€์ ๊ณผ ๋ณด ์ ‘ํ•ฉ๋ถ€๋ฅผ ํšŒ์ „ ์Šคํ”„ ๋ง์„ ์ด์šฉํ•ด ๋ชจ๋ธ๋งํ•˜๋„๋ก ํ•œ๋‹ค.

์†์ƒ ์ „ํ›„์˜ ํšŒ์ „์Šคํ”„๋ง ๊ฑฐ๋™์€ ์„ ํ˜• ๊ฑฐ๋™์„ ํ•œ๋‹ค๊ณ  ๊ฐ€์ • ํ•œ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ๋งŒ์•ฝ ์ง€์ง„์— ์˜ํ•œ ์†์ƒ์€ ํšŒ์ „ ์Šคํ”„๋ง์ด ์œ„์น˜ํ•œ ๊ณณ์—์„œ๋งŒ ๋ฐœ์ƒํ•˜๋ฉฐ, ์ด๋Š” ํ•ด๋‹น ์Šคํ”„๋ง์˜ ๊ฐ•์„ฑ ์ €ํ•˜๋กœ ๊ณ ๋ ค๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ๊ฐ€์ •ํ•œ๋‹ค.

๊ตฌ์กฐ๋ฌผ์˜ ์ง„๋™์„ ๋ฐœ์ƒ์‹œํ‚ค๊ธฐ ์œ„ํ•ด ๊ฐ€์ง„๊ธฐ๊ฐ€ ์ตœ์ƒ๋ถ€์ธต์— ์œ„ ์น˜ํ•œ๋‹ค๊ณ  ๊ฐ€์ •ํ•œ๋‹ค. ๊ฐ€์ง„๊ธฐ๋ฅผ ํ†ตํ•ด ๋ฐฑ์ƒ‰์žก์Œ(White noise) ๋ฐ ์กฐํ™”ํ•˜์ค‘์„ ๋ฐœ์ƒ์‹œํ‚ค๊ณ , ๊ธฐ๋‘ฅ ๋‹จ๋ถ€(๋ณ€ํ˜•๋ฅ ์ด ์„ค์น˜๋˜์—ˆ๋‹ค๊ณ  ๊ฐ€์ •๋œ ๊ณณ)์—์„œ ๋ณ€ํ˜•๋ฅ  ๊ฐ’์„ ์–ป๊ณ  ์ด๋ฅผ ํ†ตํ•ด ํœจ๋ชจ๋ฉ˜ํŠธ, ํšก๊ฐ€์† ๋„ ์‘๋‹ต, ๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ ๋“ฑ์„ ์–ป๋Š”๋‹ค.

๊ตฌ์กฐํ•ด์„์€ OpenSees๋ฅผ ์‚ฌ์šฉํ•œ๋‹ค. ๊ธฐ๋‘ฅ, ๋ณด ๋“ฑ์˜ ๋ถ€์žฌ๋Š” ํƒ„ ์„ฑ๋ถ€์žฌ๋ฅผ ์‚ฌ์šฉํ•˜๋ฉฐ, ํŒจ๋„์กด์˜ ๊ฐ•์„ฑ ๋ฐ ๊ฐ•๋„ ํšจ๊ณผ๋Š” ๊ณ ๋ คํ•˜์ง€ ์•Š๋Š”๋‹ค. ํšŒ์ „์Šคํ”„๋ง์˜ ๊ฐ•์„ฑ์€ 6EI/L(E ํƒ„์„ฑ๊ณ„์ˆ˜, I ๋‹จ๋ฉด์ด์ฐจ๋ชจ ๋ฉ˜ํŠธ, L ๋ถ€์žฌ๊ธธ์ด)์„ ํ†ตํ•ด ์„ค์ •ํ•œ๋‹ค. ๋‹ค์ด์–ดํ”„๋žจ ๋ฐ P-delta ํšจ ๊ณผ๋Š” ๊ณ ๋ ค๋˜๋ฉฐ, gravity-frame์— ์˜ํ•œ ๊ธฐ์—ฌ๋Š” ๊ณ ๋ คํ•˜์ง€ ์•Š๋Š”๋‹ค.

๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” Backpropagation Algorhitm(BP)์„ ์‚ฌ์šฉํ•˜ ์—ฌ ์‹ ๊ฒฝ๋ง ๋‚ด์˜ ๊ฐ€์ค‘์น˜๋ฅผ ๊ฒฐ์ •ํ•˜๋„๋ก ํ•œ๋‹ค. Fig. 1์—์„œ ์‹ ๊ฒฝ๋ง ์˜ ์ž…๋ ฅ์ธต์—๋Š” ๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ, ๊ธฐ๋‘ฅ ๋‹จ๋ถ€์˜ ๋ชจ๋ฉ˜ํŠธ ๊ฐ’ ์„ ์ž…๋ ฅํ•œ๋‹ค. ์˜ˆ์ œ ๊ตฌ์กฐ๋ฌผ์˜ ๊ฒฝ์šฐ 1์ฐจ์™€ 2์ฐจ ๋ชจ๋“œ์˜ ์งˆ๋Ÿ‰์ฐธ์—ฌ ์œจ์„ ํ•ฉํ•˜๋ฉด 90% ์ด์ƒ ๊ฐ’์„ ๋‚˜ํƒ€๋‚ด๊ธฐ ๋•Œ๋ฌธ์—, ๊ณ ์œ ์ง„๋™์ˆ˜์™€ ๋ชจ๋“œํ˜•์ƒ์€ 1์ฐจ์™€ 2์ฐจ๋งŒ ๊ณ ๋ คํ•œ๋‹ค. ๋”ฐ๋ผ์„œ ๊ณ ์œ ์ง„๋™์ˆ˜๋ฅผ ์ž… ๋ ฅํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” 2๊ฐœ์˜ ๋…ธ๋“œ๊ฐ€ ํ•„์š”ํ•˜๋ฉฐ, ๋ชจ๋“œํ˜•์ƒ์„ ์ž…๋ ฅํ•˜ ๊ธฐ ์œ„ํ•ด์„œ๋Š” 10๊ฐœ(5์ธต*2๊ฐœ ๋ชจ๋“œ)์˜ ๋…ธ๋“œ๊ฐ€ ํ•„์š”ํ•˜๋‹ค. ํ•œ๋ฉด, ๊ธฐ ๋‘ฅ ๋‹จ๋ถ€์˜ ํœจ๋ชจ๋ฉ˜ํŠธ๊ฐ’์„ ์ž…๋ ฅํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” 10๊ฐœ(2๊ฐœ/๊ธฐ๋‘ฅ*5 ๊ธฐ๋‘ฅ)์˜ ๋…ธ๋“œ๊ฐ€ ํ•„์š”ํ•˜๋‹ค. ์ด๋ฅผ ์ข…ํ•ฉํ•˜๋ฉด, ์ด 22๊ฐœ์˜ ์ž…๋ ฅ๋…ธ๋“œ ๊ฐ€ ์‚ฌ์šฉ๋œ๋‹ค. ์€๋‹‰์ธต์˜ ๋…ธ๋“œ๋Š” 20๊ฐœ๋กœ ์„ค์ •ํ•œ๋‹ค. ์€๋‹‰์ธต์˜ ํ™œ ์„ฑํ•จ์ˆ˜๋Š” sigmoidํ•จ์ˆ˜๊ฐ€ ์‚ฌ์šฉ๋œ๋‹ค.

์ถœ๋ ฅ์ธต์˜ ๋…ธ๋“œ์—๋Š” ํšŒ์ „์Šคํ”„๋ง์˜ ์†์ƒ์ง€ํ‘œ(DF)์ด ์—ฐ๊ฒฐ๋œ ๋‹ค. DF๋Š” ์‹ (3)๊ณผ ๊ฐ™์ด ํ™œ์šฉ๋œ๋‹ค. DF๊ฐ€ 1.0์ด๋ฉด ์†์ƒ์ด ์—†๋‹ค ๋Š” ๊ฒƒ์„ ์˜๋ฏธํ•˜๋ฉฐ, DF๊ฐ€ 0.8์ด๋ฉด ํ•ด๋‹น ํšŒ์ „์Šคํ”„๋ง์ด ์œ„์น˜ํ•œ ๊ณณ์— ์†์ƒ์ด ๋ฐœ์ƒํ•˜์—ฌ ํšŒ์ „์Šคํ”„๋ง์˜ ๊ฐ•์„ฑ๊ฐ’์ด 20% ์ €๊ฐํ•˜์˜€ ์Œ์„ ์˜๋ฏธํ•œ๋‹ค. ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ์ธต๋ณ„๋กœ ๋…๋ฆฝ์ ์ธ ์†์ƒ์ด ๋ฐœ์ƒ ํ•  ์ˆ˜ ์žˆ๋Š” ๋ฐ˜๋ฉด์—, ๋™์ผ ์ธต ๋‚ด์—์„œ๋Š” ๋™์ผํ•œ ์†์ƒ์œจ์ด ๋ฐœ์ƒํ•œ ๋‹ค๊ณ  ๊ฐ€์ •ํ•œ๋‹ค. ๋”ฐ๋ผ์„œ Fig. 4์—์„œ ๋™์ผ ์ธต์— ์œ„์น˜ํ•œ ํšŒ์ „์Šคํ”„ ๋ง์˜ DF๋Š” ๋™์ผํ•œ ๋ณ€์ˆ˜๊ฐ€ ์—ฐ๊ฒฐ๋˜์–ด ์ด 6๊ฐœ์˜ DF ๋ณ€์ˆ˜๊ฐ€ ์‚ฌ์šฉ ๋œ๋‹ค. ์ถœ๋ ฅ์ธต์—๋Š” ์ด 6๊ฐœ์˜ ๋…ธ๋“œ๊ฐ€ ์„ค์ •๋˜์–ด ๊ฐ DF ๊ฐ’์„ ์ถœ๋ ฅ ํ•˜๊ฒŒ ๋œ๋‹ค.

(3)
์†์ƒํ›„ํšŒ์ „๊ฐ•์„ฑ = โ€‰ D F ร— ์†์ƒ์ „ํšŒ์ „๊ฐ•์„ฑ

๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จํ•˜๊ณ  ์†์ƒ ์˜ˆ์ธก์˜ ์ •ํ™•์„ฑ ์„ ํ‰๊ฐ€ํ•˜๊ธฐ ์œ„ํ•ด ์ด 729๊ฐ€์ง€์˜ ์†์ƒ์‹œ๋‚˜๋ฆฌ์˜ค๋ฅผ ๊ณ ๋ คํ•˜์˜€๋‹ค. ๊ฐ ํšŒ์ „์Šคํ”„๋ง์˜ DF๋Š” 1.00, 0.75, 0.50์™€ ๊ฐ™์ด ์ด 3๊ฐ€์ง€์˜ ๊ฐ’ ์„ ๊ฐ€์งˆ ์ˆ˜ ์žˆ๋„๋ก ํ•˜์—ฌ 729(=36)๊ฐ€์ง€์˜ ์†์ƒ์‹œ๋‚˜๋ฆฌ์˜ค์— ๋Œ€ํ•œ ๊ตฌ์กฐํ•ด์„์„ ์ˆ˜ํ–‰ํ•˜๊ณ  ์ด๋“ค์˜ ๊ฒฐ๊ณผ๋กœ๋ถ€ํ„ฐ ์ž…๋ ฅ์ธต๊ณผ ์ถœ๋ ฅ์ธต์— ํ•„์š”ํ•œ ๋ณ€์ˆ˜๋“ค์˜ ๊ฐ’์„ ์ •๋ฆฌํ•˜์˜€๋‹ค.

์ด 729๊ฐ€์ง€ ์†์ƒ์‹œ๋‚˜๋ฆฌ์˜ค ์ค‘ 656๊ฐ€์ง€ ์‹œ๋‚˜๋ฆฌ์˜ค(์•ฝ 90%)์˜ ๊ฒฐ๊ณผ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จ ๋ฐ ๊ฒ€์ฆํ•˜๋Š”๋ฐ ์‚ฌ์šฉํ•˜๋ฉฐ, ๋‚˜๋จธ์ง€ 73๊ฐ€์ง€ ์‹œ๋‚˜๋ฆฌ์˜ค(์•ฝ 10%)์˜ ์†์ƒ ์˜ˆ์ธก์˜ ์ •ํ™•์„ฑ์„ ํ‰๊ฐ€ํ•˜๋Š” ๋ฐ ํ™œ์šฉํ•œ๋‹ค. 73๊ฐ€์ง€์˜ ์‹œ๋‚˜๋ฆฌ์˜ค ๋ฐ์ดํ„ฐ์—๋Š” ๊ฐ ์‹œ๋‚˜๋ฆฌ์˜ค์˜ ์ž…๋ ฅ๊ฐ’์„ ์‹ ๊ฒฝ๋ง์— ์ž…๋ ฅํ•  ๊ฒฝ์šฐ ์ถœ๋ ฅ๋˜์–ด์•ผ ํ•  ์ •ํ™•ํ•œ ์ถœ๋ ฅ ๊ฐ’์„ ๊ฐ€์ง€๊ณ  ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ 656๊ฐ€์ง€ ์‹œ๋‚˜๋ฆฌ์˜ค ๋ฐ์ดํ„ฐ๋ฅผ ํ†ตํ•ด ์ƒ์„ฑํ•œ ์‹ ๊ฒฝ๋ง์— ์ƒˆ๋กœ์šด 73๊ฐ€์ง€ ์‹œ๋‚˜๋ฆฌ์˜ค์˜ ์ž…๋ ฅ๊ฐ’์„ ๋„ฃ๊ฒŒ ๋˜๋ฉด ์‹ ๊ฒฝ๋ง์— ์˜ํ•œ ์˜ˆ์ธก๊ฐ’์„ ์–ป์„ ์ˆ˜ ์žˆ์œผ๋ฉฐ, ์ด๋ฅผ ์ฐธ๊ฐ’๊ณผ ๋น„ ๊ตํ•  ์ˆ˜ ์žˆ๋‹ค. ์ด๋Ÿฌํ•œ ์‹ ๊ฒฝ๋ง์˜ ์†์ƒ ์˜ˆ์ธก๊ฐ’์˜ ์˜ค์ฐจ๋ฅผ ํ‰๊ฐ€ํ•˜ ๋Š” ์„ค๋ช…๋„๋ฅผ Fig. 5์— ๋‚˜ํƒ€๋‚ด์—ˆ๋‹ค.

Fig. 5

Illustration of damage prediction by using ANN

JKSMI-22-107_F5.jpg

4.2. ์ ์šฉ ๊ฒฐ๊ณผ

Fig. 4์™€ ๊ฐ™์€ ์˜ˆ์ œ ๊ตฌ์กฐ๋ฌผ์˜ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๋ฐ์ดํ„ฐ๋ฅผ ์ด์šฉํ•˜ ์—ฌ ์‹ ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จํ•˜๊ณ  ์ด๋ฅผ ํ†ตํ•ด ์ ‘ํ•ฉ๋ถ€ ์†์ƒ์„ ์˜ˆ์ธกํ•œ ๊ฒฐ๊ณผ๋Š” Fig. 6๊ณผ ๊ฐ™๋‹ค. Fig. 6(a)๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จ์‹œํ‚ค๋Š” ๊ณผ์ •์—์„œ ๋ฐœ์ƒํ•œ ์˜ค์ฐจ์ด๋ฉฐ, Fig. 6(b)๋Š” ํ›ˆ๋ จ์„ ํ†ตํ•ด ์ƒ์„ฑ๋œ ์ธ ๊ณต์‹ ๊ฒฝ๋ง์— ์ƒˆ๋กœ์šด ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๋ฐ์ดํ„ฐ๋ฅผ ์ ์šฉํ•˜์—ฌ ํšŒ์ „๊ฐ•์„ฑ ๊ฐ’์„ ์˜ˆ์ธกํ•œ ๊ฒฐ๊ณผ์ด๋‹ค. Fig. 5์—์„œ ๋‚˜ํƒ€๋‚œ ๊ฒƒ์ฒ˜๋Ÿผ ๋ณธ ์—ฐ๊ตฌ์—์„œ ์‚ฌ์šฉ๋œ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๋ฐ์ดํ„ฐ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ์ƒ์„ฑํ•˜๊ธฐ ์œ„ํ•ด ์‚ฌ ์šฉ๋˜๋Š” ํ›ˆ๋ จ์šฉ ๋ฐ์ดํ„ฐ์™€ ์ด๋ฅผ ๊ฒ€์ฆํ•˜๊ธฐ ์œ„ํ•ด ์‚ฌ์šฉ๋˜๋Š” ๋ฐ์ดํ„ฐ ๋ฅผ ๊ตฌ๋ถ„๋œ๋‹ค. ๊ทธ๋ฆผ์— ๋‚˜ํƒ€๋‚œ ๊ฒƒ์ฒ˜๋Ÿผ ๋ชจ๋“  ์†์ƒ์ง€ํ‘œ(DF1~DF6) ์— ๋Œ€ํ•œ ์˜ค์ฐจ๊ฐ€ 0.002 ์ดํ•˜์˜ ์ˆ˜์ค€์ธ ๊ฒƒ์„ ํ™•์ธํ•  ์ˆ˜ ์žˆ๋‹ค. ์ด ๋Š” ์†์ƒ์œ„์น˜ ๋ฐ ์ •๋„๋ฅผ ๋†’์€ ์ •ํ™•๋„๋ฅผ ๊ฐ€์ง€๊ณ  ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ์Œ ์„ ์˜๋ฏธํ•œ๋‹ค. Fig. 6์—์„œ ์‚ฌ์šฉ๋œ ๋ฐ์ดํ„ฐ๋Š” ๊ฐ ์†์ƒ์ง€ํ‘œ(DF)์˜ ๊ฐ’์„ 0.50, 0.75, 1.00์œผ๋กœ ํ•œ์ •์‹œํ‚ค๊ณ  ๊ฒ€์ฆํ•œ ๊ฒฐ๊ณผ์ด๋‹ค(Method 1). ์ฆ‰, ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จ์‹œํ‚ค๋Š”๋ฐ ๋™์ผํ•œ ์กฐ๊ฑด์—์„œ ๋ฐœ์ƒํ•˜ ๋Š” ์†์ƒ ๋ฐ ํšŒ์ „๊ฐ•์„ฑ๊ฐ’์— ๋Œ€ํ•ด์„œ ์ •ํ™•ํžˆ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ์Œ์„ ํ™• ์ธํ•  ์ˆ˜ ์žˆ๋‹ค.

Fig. 6

Errors from ANN(Method 1)

JKSMI-22-107_F6.jpg

์˜ˆ์ธก๋œ ํšŒ์ „๊ฐ•์„ฑ๊ฐ’์„ ์ด์šฉํ•ด ๊ตฌ์กฐ ๋ชจ๋ธ๋ง ๋ฐ ํ•ด์„์„ ์ˆ˜ํ–‰ ํ•˜์—ฌ ๊ณ ์œ ํŠน์„ฑ๊ฐ’์„ ์–ป๊ณ , ์ด๋ฅผ ์ฐธ๋œ ํšŒ์ „๊ฐ•์„ฑ๊ฐ’์œผ๋กœ ์ •์˜๋œ ๊ตฌ์กฐ๋ฌผ์˜ ๊ณ ์œ ํŠน์„ฑ๊ฐ’๊ณผ ๋น„๊ตํ•˜์˜€๋‹ค. ๊ทธ ๊ฒฐ๊ณผ๋Š” Fig. 7๊ณผ ๊ฐ™๋‹ค. Fig. 7(a)๋Š” 1์ฐจ, 2์ฐจ ๊ณ ์œ ์ง„๋™์ˆ˜์˜ ์˜ค์ฐจ๋ฅผ ๋‚˜ํƒ€๋‚ด๋ฉฐ, Fig. 7(b) ๋Š” 1์ฐจ, 2์ฐจ ๋ชจ๋“œํ˜•์ƒ์˜ ์œ ์‚ฌ๋„๋ฅผ ์˜๋ฏธํ•œ๋‹ค. ๋ชจ๋“œํ˜•์ƒ์˜ ์œ ์‚ฌ ๋„๋Š” ์‹ (4)์™€ ๊ฐ™์ด Modal Assurance Criteria(MAC)์„ ์‚ฌ์šฉํ•œ ๋‹ค(Chang and Kim, 2008). MAC๊ฐ’์€ 0๊ณผ 1์‚ฌ์ด์˜ ๊ฐ’์„ ๊ฐ€์ง€ ๋ฉฐ, 1์— ๊ทผ์ ‘ํ• ์ˆ˜๋ก ๋‘ ๋ชจ๋“œํ˜•์ƒ์˜ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ๋†’๋‹ค.

Fig. 7

Comparison of dynamic properties(Method 1)

JKSMI-22-107_F7.jpg

(4)
M A C i j = [ { ฯ• i } T { ฯ• j } ] 2 [ { ฯ• i } T { ฯ• i } ] [ { ฯ• j } T { ฯ• j } T ]

์—ฌ๊ธฐ์„œ, ฯ•i์™€ ฯ•j๋Š” ๋น„๊ตํ•˜๋Š” ๋‘ ๋ชจ๋“œํ˜•์ƒ์„ ์˜๋ฏธํ•œ๋‹ค. ๊ณ ์œ  ์ง„๋™์ˆ˜์™€ ๋ชจ๋“œํ˜•์ƒ์ด ๋ชจ๋‘ ์œ ์‚ฌํ•˜๊ฒŒ ๋‚˜ํƒ€๋‚˜๋Š” ๊ฒƒ์„ Fig. 7(a) ๊ณผ 7(b)๋ฅผ ํ†ตํ•ด ํ™•์ธํ•  ์ˆ˜ ์žˆ๋‹ค.

Fig. 6๊ณผ 7์€ ํ›ˆ๋ จ์šฉ ๋ฐ์ดํ„ฐ์™€ ๋™์ผํ•œ ์กฐ๊ฑด(DF๊ฐ’์ด 0.50, 0.75, 1.00 ์ค‘ ํ•œ ๊ฐ’๋งŒ์„ ๊ฐ€์ง)์—์„œ ์˜ˆ์ธก๋œ ๊ฒฐ๊ณผ์ด๋‹ค. ํ›ˆ๋ จ ๋ฐ ์ดํ„ฐ์™€ ๋‹ค๋ฅธ ์ƒํ™ฉ์— ๋Œ€ํ•œ ์˜ˆ์ธก ์ •ํ™•๋„๋ฅผ ํŒŒ์•…ํ•˜๊ธฐ ์œ„ํ•ด์„œ ๊ฐ DF๋Š” 0.5์—์„œ 1.0์‚ฌ์ด ์ค‘ ์ž„์˜ ๊ฐ’์„ ๋žœ๋คํ•˜๊ฒŒ ๊ฐ€์ง€๋„๋ก ์„ค์ • ํ•œ ํ›„, ์ด 100๊ฐœ์˜ ์ƒˆ๋กœ์šด ๋ฐ์ดํ„ฐ๋ฅผ ์ƒ์„ฑํ•˜์—ฌ ์˜ค์ฐจ๋ฅผ ๋ถ„์„ํ•˜ ์˜€๋‹ค(Method 2). ์˜ˆ๋ฅผ ๋“ค๋ฉด, ํ•œ ๊ฒฝ์šฐ์˜ ํšŒ์ „๊ฐ•์„ฑ๊ฐ’(DF1~DF6) ์€ (0.6673, 0.9531, 0.6470, 0.7349, 0.8268, 0.6951)์ด๋‹ค. ํšŒ์ „ ๊ฐ•์„ฑ์˜ ์˜ˆ์ธก ์ •ํ™•๋„๋Š” Fig. 6์˜ ๊ฒฐ๊ณผ๋ณด๋‹ค๋Š” ์˜ค์ฐจ๊ฐ€ ํฌ์ง€๋งŒ, Fig. 8์— ๋‚˜ํƒ€๋‚œ ๊ฒƒ์ฒ˜๋Ÿผ ์˜ค์ฐจ ์ˆ˜์ค€์€ 0.06 ์ดํ•˜๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค. ์ด ๋Ÿฌํ•œ ์˜ค์ฐจ๋Š” Fig. 9(a)์™€ 9(b)์— ํ‘œ์‹œ๋œ ๊ฒƒ์ฒ˜๋Ÿผ ์ฐธ๊ฐ’์— ๊ทผ์ ‘ํ•œ ๊ณ ์œ ์ง„๋™์ˆ˜์™€ ๋ชจ๋“œํ˜•์ƒ์„ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋Š” ์ˆ˜์ค€์ด๋‹ค.

Fig. 8

Comparison of damage factor values(Method 2)

JKSMI-22-107_F8.jpg
Fig. 9

Comparison of dynamic properties(Method 2)

JKSMI-22-107_F9.jpg

5. ๊ฒฐ ๋ก 

๋ณธ ์—ฐ๊ตฌ๋Š” ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ์ด์šฉํ•ด ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ๊ณจ์กฐ์˜ ์ ‘ํ•ฉ๋ถ€ ์†์ƒ์„ ์˜ˆ์ธกํ•˜๋Š” ๊ธฐ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค. ์ธ๊ณต์‹ ๊ฒฝ๋ง์„ ํ›ˆ๋ จํ•˜๊ณ  ์†์ƒ ์˜ˆ์ธก์˜ ์ •ํ™•์„ฑ์„ ํ‰๊ฐ€ํ•˜๊ธฐ ์œ„ํ•ด ์ด 829๊ฐ€์ง€์˜ ์†์ƒ์‹œ๋‚˜ ๋ฆฌ์˜ค(DF๊ฐ’์ด 0.50, 0.75, 1.00 ์ค‘ ํ•œ ๊ฐ’๋งŒ ๊ฐ€์ง€๋Š” ๊ฒฝ์šฐ 729๊ฐ€ ์ง€, DF๊ฐ’์ด 0.5์—์„œ 1.0 ์‚ฌ์ด ์ค‘ ์ž„์˜์˜ ๊ฐ’์„ ๊ฐ€์ง€๋Š” 100๊ฐ€์ง€) ๋ฅผ ๊ณ ๋ คํ•˜์˜€๋‹ค. ์ธ๊ณต์‹ ๊ฒฝ๋ง์˜ ์ž…๋ ฅ์ธต์—๋Š” ๊ธฐ๋‘ฅ ๋ถ€์žฌ์˜ ํœจ๋ชจ ๋ฉ˜ํŠธ, ๊ณ ์œ ์ง„๋™์ˆ˜, ๋ชจ๋“œํ˜•์ƒ ์ •๋ณด๊ฐ€ ์‚ฌ์šฉ๋˜๋ฉฐ, ์ถœ๋ ฅ์ธต์—๋Š” ๊ตฌ ์กฐ๋ฌผ ์ ‘ํ•ฉ๋ถ€์˜ ํšŒ์ „๊ฐ•์„ฑ ์†์ƒ์ง€ํ‘œ๊ฐ€ ์‚ฌ์šฉํ•œ๋‹ค.

5์ธต 1๊ฒฝ๊ฐ„ ์ฒ ๊ณจ๋ชจ๋ฉ˜ํŠธ๊ณจ์กฐ ์˜ˆ์ œ ๊ฒ€์ฆ ๊ฒฐ๊ณผ, ๋‹ค์ˆ˜์˜ ์œ„์น˜์— ์„œ ๋ฐœ์ƒํ•˜๋Š” ์†์ƒ์— ๋Œ€ํ•ด ์œ„์น˜๋ณ„ ์†์ƒ์ •๋„๋ฅผ ์‹ ๋ขฐํ•  ์ˆ˜ ์žˆ๋Š” ์ˆ˜์ค€์œผ๋กœ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์œผ๋กœ ํ™•์ธ๋œ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ธ๊ณต์‹ ๊ฒฝ ๋ง ๊ธฐ๋ฐ˜ ์˜ˆ์ธก์„ฑ์„ ํŒŒ์•…ํ•˜๊ธฐ ์œ„ํ•ด ํ›ˆ๋ จ์šฉ ๋ฐ์ดํ„ฐ์™€ ๋™์ผํ•œ ๋ฐฉ ๋ฒ•(Method 1)์œผ๋กœ ์ƒ์„ฑ๋œ ๋ฐ์ดํ„ฐ ๊ฒฐ๊ณผ(Figs. 6๊ณผ 7)๋ณด๋‹ค ๋ฐ์ด ํ„ฐ ๋ฒ”์œ„๋Š” ๋™์ผํ•˜์ง€๋งŒ ์ƒ์„ฑ๋ฐฉ๋ฒ•์„ ๋‹ค๋ฅด๊ฒŒ ํ•  ๊ฒฝ์šฐ(Method 2) ์˜ ๊ฒฐ๊ณผ(Figs. 8๊ณผ 9)๋Š” ๋‹ค์†Œ ์˜ค์ฐจ๊ฐ€ ํฌ๊ฒŒ ๋ฐœ์ƒํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ ํƒ€๋‚ฌ๋‹ค.

 ๊ฐ์‚ฌ์˜ ๊ธ€

์ด ๋…ผ๋ฌธ์€ 2017๋…„๋„ ์ •๋ถ€(๊ต์œก๋ถ€)์˜ ์žฌ์›์œผ๋กœ ํ•œ๊ตญ์—ฐ๊ตฌ์žฌ ๋‹จ์˜ ์ง€์›์„ ๋ฐ›์•„ ์ˆ˜ํ–‰๋œ ๊ธฐ์ดˆ์—ฐ๊ตฌ์‚ฌ์—…์ž„(No. NRF-2017R1 D1A1B03034978)

 

1 
(2013), Damage Location Detection of Shear Building Structures Using Mode Shape., Journal of the Korea Institute for Structural Maintenance and Inspection, 17(1), 124-132.
2 
(2014), Damage Detection of Shear Building Structures Using Dynamic Response., Journal of the Korea Institute for Structural Maintenance and Inspection, 18(3), 101-107.
3 
(2003), Damage Identification in Beam-Type Structures: Frequency-Based Method vs Mode-Shape-Based Method., Eng. Struct., 25(1), 57-67.
4 
(2001), Vibration-Based Damage Detection in Civil Engineering: Excitation Sources and Temperature Effects., Smart Mater. Struct., 10(3)DOI
5 
(1998), A Summary Review of Vibration-Based Damage Identification Methods., Shock and Vibration Digest, 30(2), 91-105.
6 
(2011), Vibration-Based Damage Identification Methods: A Review and Comparative Study., Struct. Health Monit., 10(1), 83-111.
7 
(1999), Damage Detection with Spatial Wavelets., Int. J. Solids Struct., 36(23), 3433-3468.
8 
(2004), Damage Detection of Structures by Wavelet Analysis., Eng. Struct., 26(3), 347-362.
9 
(1991), Damage Detection from Changes in Curvature Mode Shapes., J. Sound Vibrat., 145(2), 321-332.
10 
(2000), Structural Damage Detection from Modal Strain Energy Change., J. Eng. Mech., 126(12), 1216-1223.
11 
(1994), Damage Detection in Structures using Changes in Flexibility., J. Sound Vibrat., 169(1), 3-17.
12 
(2010), Damage Detection Method based on Element Modal Strain Energy Sensitivity., Adv. Struct. Eng., 13(6), 1075-1088.
13 
(2008), Using Incomplete Modal Data for Damage Detection in Offshore Structures., Ocean Eng., 35(17), 1793-1799.
14 
(1998), Structural Damage Detection From Incomplete and Noisy Modal Test Data., J. Eng. Mech., 124(11), 1280-1288.
15 
(2000), Damage Localization by Directly Using Incomplete Mode Shapes., J. Eng. Mech., 126(6), 656-660.
16 
(2003), Damage Detection Using Spatially Incomplete Frequency Response Funcitons., Mech. Syst. Signal Process., 17(3), 519-532.
17 
(2015), Structural Damage Detection Using Modal Strain Energy and Hybrid Multiobjective Optimization., Comput. Aided Civ. Infrastruct. Eng., 30, 347-358.
18 
(2012), Damage Detection based on Improved Particle Swarm Optimization using Vibration Data., Appl. Soft Comput., 12, 2329-2335.
19 
(2012), A Two Stage Method for Structural Damage Detection Using A Modal Strain Eenrgy Based Index and Particle Swarm Optimization., Int. J. Non-linear Mech., 47, 1-8.
20 
(2007), An Evolutionary Multiobjetive Framework for Structural Damage Localization and Quantification., Eng. Struct., 29, 2540-2550.
21 
(2016), A Numerical Study to Estimate the Lateral Responses of Steel Moment Frames Using Strain Data., Journal of the Korea Institute for Structural Maintenance and Inspection, 20(6), 113-119.
22 
(2001), Neural Network Design for Engineering Applications., Comput. Struc., 79(17), 1541-1552.
23 
(1994), Neural Networks in Civil Engineering I: Principles and Understanding., J. Comput. Civ. Eng., 8(2), 131-148.
24 
(1992), Use of Neural Networks in Detection of Structural Damage., Comput. Struc., 42(4), 649-659.
25 
(2001), Neural Networks in Civil Engineering: 1989-2000., Comput. Aided Civ. Infrastruct. Eng., 16(2), 126-142.
26 
(2011), Structural Analysis, 451-486.
27 
(2001), Modal Identification of Output-only Systems Using Frequency Domain Decomposition., Smart Mater. Struct., 10, 441-445.
28 
(2008), Output-only Modal Identification of a Cable-Stayed Bridge Using Wireless Monitoring Systems., Eng. Struct., 30, 1820-1830.
29 
(2010), Full-scale Dynamic Response of an RC Building Under Weak Seismic Motions Using Earthquake Recordings, Ambient Vibrations and Modelling., Earthquake Eng. Struct. Dynam., 39, 419-441.
30 
(2008), Estimation of Displacement Response from the Measured Dynamic Strain Signals Using Mode Decomposition Technique., KSCE J. Civ. Eng., 28, 507-515.