Mobile QR Code QR CODE

Journal of the Korea Concrete Institute

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

  1. μ •νšŒμ›,μ„œμšΈμ‹œλ¦½λŒ€ν•™κ΅ 건좕곡학과 μŠ€λ§ˆνŠΈμ‹œν‹°μœ΅ν•©μ „κ³΅ 박사과정
  2. μ •νšŒμ›,μ„œμšΈμ‹œλ¦½λŒ€ν•™κ΅ λ„μ‹œλ°©μž¬μ•ˆμ „μ—°κ΅¬μ†Œ 박사후연ꡬ원
  3. μ •νšŒμ›,μ„œμšΈμ‹œλ¦½λŒ€ν•™κ΅ λ°©μž¬κ³΅ν•™κ³Ό 쑰ꡐ수
  4. μ •νšŒμ›,μ„œμšΈμ‹œλ¦½λŒ€ν•™κ΅ 건좕곡학과 μŠ€λ§ˆνŠΈμ‹œν‹°μœ΅ν•©μ „κ³΅ ꡐ수, κ΅μ‹ μ €μž



콘크리트, λ™κ²°μœ΅ν•΄, μ €ν•­μ„±λŠ₯, 인곡신경망, μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜
Concrete, Freeze-thaw, Resistance, Artificial neural network (ANN), Relative dynamic modulus of elasticity(RDME)

1. μ„œ λ‘ 

κ΅­λ‚΄ 겨울철 μ˜¨λ„μ˜ λ³€ν™” 폭이 크기 λ•Œλ¬Έμ— 콘크리트 ꡬ쑰물은 주기적인 λ™κ²°μœ΅ν•΄ ν”Όν•΄λ₯Ό κ²½ν—˜ν•˜κ³  μžˆλ‹€. Fig. 1에 λ‚˜νƒ€λ‚Έ 바와 같이 λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 내ꡬ성 μ €ν•˜ν˜„μƒμ€ 주둜 λ‚΄λΆ€ λ―Έμ„Έκ· μ—΄ 및 ν‘œλ©΄λ°•λ¦¬μ™€ 같은 문제λ₯Ό λ°œμƒμ‹œν‚¨λ‹€. 이둜 μΈν•˜μ—¬ ꡬ쑰물의 미관이 μ†μƒλ˜λ©°, λΆ€μž¬μ˜ λ‚΄λ ₯이 κ°μ†Œν•¨μœΌλ‘œμ¨ ꡬ쑰물의 수λͺ…이 단좕될 수 μžˆλ‹€(Neville, 1995). λ™κ²°μœ΅ν•΄ ν”Όν•΄λ₯Ό μž…μ€ 콘크리트 ꡬ쑰물의 내ꡬ성λŠ₯ μ €ν•˜λ‘œ μΈν•œ μ‚¬νšŒμ  λΉ„μš©μ΄ μ§€μ†μ μœΌλ‘œ λ°œμƒλ˜κ³  μžˆλŠ” 것이닀. 콘크리트 배합에 λ”°λ₯Έ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯을 νŒŒμ•…ν•˜κΈ° μœ„ν•΄μ„œλŠ” μ‹€ν—˜μ΄ ν•„μˆ˜μ μ΄μ§€λ§Œ 이λ₯Ό μœ„ν•΄μ„œλŠ” μƒλ‹Ήν•œ μ‹œκ°„κ³Ό λΉ„μš©μ΄ μ†Œμš”λœλ‹€. λ”°λΌμ„œ, μ‹ μ†ν•˜λ©΄μ„œλ„ 신뒰성이 높은 콘크리트 배합에 λ”°λ₯Έ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ˜ κ°œλ°œμ„ ν†΅ν•˜μ—¬ μ΄λŸ¬ν•œ λ¬Έμ œμ μ„ 극볡할 수 μžˆλ‹€.

Lee(2018)λŠ” 콘크리트의 탄산화와 λ™κ²°μœ΅ν•΄μ— λŒ€ν•œ 볡합열화λ₯Ό ν‰κ°€ν•˜μ˜€λ‹€. Choi(2021)λŠ” λ™κ²°μœ΅ν•΄ 싸이클 μˆ˜μ— λ”°λ₯Έ 압좕강도, λ°˜λ°œκ²½λ„, μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜ 및 초음파 λΉ„μ„ ν˜•μ„±μ„ μΈ‘μ •ν•˜μ˜€μœΌλ©°, SEM을 ν†΅ν•˜μ—¬ 미세균열을 λΆ„μ„ν•˜μ˜€λ‹€. Yoon(2017)은 넓은 콘크리트 보λ₯Ό λŒ€μƒμœΌλ‘œ P파, S파, R파의 νŽ„μŠ€ 속도λ₯Ό μΈ‘μ •ν•˜κ³  각 파의 μ’…λ₯˜λ³„ 속도에 λŒ€ν•œ 톡계적 뢄포λ₯Ό μ‘°μ‚¬ν•˜μ˜€λ‹€. You(2010)λŠ” μ‹€ν—˜μ„ ν™œμš©ν•˜μ—¬ FRP λ³΅ν•©μ²΄μ˜ λ‚΄κ΅¬νŠΉμ„±μ„ νŒŒμ•…ν•˜μ˜€μœΌλ©°, 섀계에 λ°˜μ˜ν•˜κΈ° μœ„ν•΄μ„œ FRP λ³΅ν•©μ²΄μ˜ 내ꡬ성 μ €ν•˜μ— λŒ€ν•œ λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό κ΅¬μΆ•ν•˜μ˜€λ‹€. μ΅œκ·Όμ—λ„ 콘크리트 내ꡬ성 평가 μ‹€ν—˜λΏλ§Œ μ•„λ‹ˆλΌ, 내ꡬ성 평가λͺ¨λΈμ— κ΄€ν•œ 연ꡬ도 ν™œλ°œνžˆ μˆ˜ν–‰λ˜κ³  μžˆλ‹€(Yan, 2020).

졜근 곡학 λΆ„μ•Όμ—μ„œλŠ” λ³΅μž‘ν•œ λ©”μ»€λ‹ˆμ¦˜κ³Ό 큰 λΆˆν™•μ‹€μ„±μ„ 가진 μž…λ ₯ 및 좜λ ₯ λ³€μˆ˜λ‘œ 이루어진 문제λ₯Ό ν•΄κ²°ν•˜κΈ° μœ„ν•˜μ—¬ 인곡신경망(Artificial Neural Network, ANN) λͺ¨λΈμ„ ν™œμš©ν•œ 연ꡬ가 적극적으둜 μ§„ν–‰λ˜κ³  μžˆλ‹€(Nimlyat, 2017; Baek, 2010; Stankovi, 2018; Popescu, 2006). 인곡신경망은 λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό ν™œμš©ν•˜μ—¬ μž…λ ₯μΈ΅κ³Ό 좜λ ₯μΈ΅ μ‚¬μ΄μ˜ λ³΅μž‘ν•œ λΉ„μ„ ν˜•μ  관계λ₯Ό ν•™μŠ΅ν•˜κ³  이λ₯Ό ν†΅ν•˜μ—¬ κ²°κ³Όκ°’μ˜ 였차λ₯Ό μ΅œμ†Œν™”ν•  수 μžˆλ‹€. λ˜ν•œ, 이미 ν•™μŠ΅λœ 인곡신경망 λͺ¨λΈμ€ μž…λ ₯값이 주어지면 μ‹ μ†ν•˜κ²Œ 좜λ ₯값을 μ˜ˆμΈ‘ν•  수 μžˆλ‹€λŠ” μž₯점이 μžˆλ‹€. μ΄λŸ¬ν•œ 인곡신경망 μ•Œκ³ λ¦¬μ¦˜μ„ 콘크리트 λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가에 μ μš©ν•˜λ©΄ 콘크리트 λ°°ν•© 데이터λ₯Ό ν™œμš©ν•˜μ—¬ μ‹€μ‹œκ°„μœΌλ‘œ μ €ν•­μ„±λŠ₯을 μ‚°μ •ν•  수 μžˆμ„ 것이닀.

λ”°λΌμ„œ, 이 μ—°κ΅¬μ—μ„œλŠ” λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ„ κ°œλ°œν•˜κΈ° μœ„ν•˜μ—¬ κΈ°μ‘΄ 연ꡬ(Yu, 2017; Shang, 2009; Shang, 2013; Wu, 2016; Duan, 2013; Janssen, 1994)λ₯Ό 기반으둜 μ½˜ν¬λ¦¬νŠΈμ— λŒ€ν•œ λ™κ²°μœ΅ν•΄ μ‹€ν—˜ λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό κ΅¬μΆ•ν•˜μ˜€μœΌλ©°, 인곡신경망 λͺ¨λΈμ„ ν™œμš©ν•˜μ—¬ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ„ κ°œλ°œν•˜μ˜€λ‹€. λ˜ν•œ, νšŒκ·€λΆ„μ„μ„ μˆ˜ν–‰ν•˜μ—¬ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 산정식을 μ œμ•ˆν•˜μ˜€λ‹€.

Fig. 1 Types of freeze-thaw damage
../../Resources/ksm/jksmi.2023.27.6.144/fig1.png

2. λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈ

2.1 λ°μ΄ν„°λ² μ΄μŠ€ ꡬ좕

λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ˜ κ°œλ°œμ„ μœ„ν•˜μ—¬ μ½˜ν¬λ¦¬νŠΈμ— λŒ€ν•œ λ™κ²°μœ΅ν•΄ μ‹€ν—˜ κ΄€λ ¨ λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό κ΅¬μΆ•ν•˜μ˜€λ‹€. κΈ°μ‘΄ 연ꡬ(Yu, 2017; Shang, 2009; Shang, 2013; Wu, 2016; Duan, 2013; Cho, 2007; Janssen, 1994)λ₯Ό ν† λŒ€λ‘œ 총 196개의 λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό κ΅¬μΆ•ν•˜μ˜€λ‹€. Table 1μ—λŠ” λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 λ°μ΄ν„°λ² μ΄μŠ€ ꡬ좕에 ν™œμš©λœ μ°Έκ³ λ¬Έν—Œμ— λŒ€ν•œ 정보λ₯Ό λ‚˜νƒ€λ‚΄μ—ˆμœΌλ©°, Table 2μ—λŠ” λ°μ΄ν„°λ² μ΄μŠ€ ꡬ좕에 ν™œμš©λœ μ‹œν—˜μ²΄μ˜ 일뢀λ₯Ό λŒ€ν‘œμ μœΌλ‘œ λ‚˜νƒ€λ‚΄μ—ˆλ‹€. λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ€ 일반적으둜 μ‚¬μš©λ˜λŠ” 콘크리트 배합을 μ—°κ΅¬λ²”μœ„λ‘œ ν•œμ •ν•˜μ˜€μœΌλ©°, 특수 ν˜Όν™”μ œ 등을 μ‚¬μš©ν•œ 데이터λ₯Ό μ œμ™Έν•˜μ˜€λ‹€. λ”°λΌμ„œ, 평가λͺ¨λΈμ˜ μž…λ ₯λ³€μˆ˜λŠ” 콘크리트 μ €ν•­μ„±λŠ₯에 큰 영ν–₯을 λ―ΈμΉœλ‹€κ³  μ•Œλ €μ§„ λ¬Ό/μ‹œλ©˜νŠΈ λΉ„, μž”κ³¨μž¬λΉ„, κ³΅κΈ°λŸ‰, 압좕강도 및 동결-μœ΅ν•΄ 사이클 수둜 μ„€μ •ν•˜μ˜€λ‹€. λ˜ν•œ, 좜λ ₯λ³€μˆ˜λŠ” 콘크리트 λ™κ²°μœ΅ν•΄ 저항성을 νŒλ‹¨ν•  수 μžˆλŠ” μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜(Relative Dynamic Modulus of Elasticity, RDME)둜 μ„€μ •ν•˜μ˜€λ‹€. μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜λŠ” μ½˜ν¬λ¦¬νŠΈκ°€ μ˜¨λ„λ³€ν™”μ— λ”°λ₯Έ λ―Έμ„Έν•œ 균열을 μ΅œμ†Œν™”ν•˜κ³  νƒ„μ„±μ μœΌλ‘œ μœ μ§€ν•  수 μžˆλŠ” λŠ₯λ ₯을 μ˜λ―Έν•˜λŠ” κ²ƒμœΌλ‘œ μ΄ν•΄λ˜κ³  있으며, 콘크리트의 내ꡬ성 μ €ν•˜λ₯Ό ν‰κ°€ν•˜λŠ” κ°€μž₯ μ€‘μš”ν•œ μ§€ν‘œλ‘œ 널리 ν™œμš©λ˜κ³  μžˆλ‹€(Yu, 2017; Cho, 2007; Zhao, 2013).

Table 1 Information of training database from literature

Reference

Total number of data

Freeze-thaw test Method

Yu et al.(2017)

60

ASTM C666;

GB/T 50082-2009*

Shang et al.(2009)

7

GB/T 50082-2009*

Shang et al.(2013)

33

GB/T 50082-2009*

Wu et al.(2016)

28

GB/T 50082-2009*

Duan et al.(2013)

29

ASTM C666

Cho et al.(2007)

25

ASTM C666

Janssen et al.(1994)

14

ASTM C666

*GB/T 50082-2009 is a Chinese standard, whose details are very similar to ASTM Standard C666.
Table 2 Database of freeze-thaw resistance

No.

Water-cement ratio, w/c

Sand-aggregate ratio, s/a

Air content (%)

Compressive strength, fc (MPa)

Number of freeze-thaw cycles

Relative dynamic modulus of elasticity, RDME (%)

1

0.320

0.613

3.1

81.1

50

98

2

0.320

0.613

3.1

81.1

100

98

3

0.320

0.613

3.1

81.1

150

98

4

0.320

0.613

3.1

81.1

200

93

5

0.250

0.613

2.9

91.4

50

101

6

0.250

0.613

2.9

91.4

100

98

7

0.250

0.613

2.9

91.4

150

97

8

0.250

0.613

2.9

91.4

200

96

9

0.320

0.613

3.2

74.2

50

100

10

0.250

0.613

3.2

74.2

100

98

2.2 인곡신경망 ν•™μŠ΅

인곡신경망은 λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό 기반으둜 ν•˜μ—¬ μž…λ ₯λ³€μˆ˜μ™€ 좜λ ₯λ³€μˆ˜ μ‚¬μ΄μ˜ λ³΅μž‘ν•œ λΉ„μ„ ν˜•μ  관계λ₯Ό ν’€μ–΄κ°€λŠ” 데에 ν™œμš©ν•  수 μžˆλ‹€. 인곡신경망 λͺ¨λΈμ€ μž…λ ₯μΈ΅, 은닉측 및 좜λ ₯측으둜 κ΅¬μ„±λ˜μ–΄ μžˆλ‹€. 인곡신경망 λͺ¨λΈμ—λŠ” λ‰΄λŸ°μ΄λΌλŠ” μ •λ³΄μ²˜λ¦¬ μΈμžκ°€ μ‘΄μž¬ν•˜λ©°, λ‰΄λŸ°μ€ μž…λ ₯μΈ΅κ³Ό 좜λ ₯μΈ΅ 사이 κ΄€κ³„μ˜ 강도λ₯Ό ν‘œν˜„ν•˜λŠ” μ„œλ‘œ λ‹€λ₯Έ κ°€μ€‘μΉ˜(weight) 및 μ—­μΉ˜(bias)둜 μ—°κ²°λ˜μ–΄ μžˆλ‹€. ν•™μŠ΅ 과정은 κ°€μ€‘μΉ˜ 값을 μ§€μ†μ μœΌλ‘œ μˆ˜μ •ν•˜μ—¬ 였차λ₯Ό μ΅œμ†Œν™”ν•˜λŠ” λ°©ν–₯으둜 μ§„ν–‰λœλ‹€(Russel, 2010; Cho, 2015; Cho, 2017; Kang, 2019; Darkhanbat 2021).

Fig. 2에 λ‚˜νƒ€λ‚Έ 바와 같이 인곡신경망 ν•™μŠ΅μ—μ„œ μ‚¬μš©λœ μž…λ ₯측은 λ¬Ό/μ‹œλ©˜νŠΈ λΉ„, μž”κ³¨μž¬λΉ„, κ³΅κΈ°λŸ‰, 압좕강도 및 λ™κ²°μœ΅ν•΄ 사이클 수둜 5개, 은닉측은 9개의 λ‰΄λŸ°, 좜λ ₯측은 μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜λ‘œ κ΅¬μ„±ν•˜μ˜€λ‹€. λ°μ΄ν„°λ² μ΄μŠ€ 쀑 인곡신경망 λͺ¨λΈμ˜ ν•™μŠ΅, 검증 및 확인에 ν™œμš©λœ λ°μ΄ν„°μ˜ λΆ„λ°°λŠ” 각

각 70%, 15%와 15%둜 κ³„νšν•˜μ˜€λ‹€. Fig. 3에 λ‚˜νƒ€λ‚Έ 바와 같이 μž…λ ₯μΈ΅μ—μ„œ μ€λ‹‰μΈ΅μœΌλ‘œμ˜ 전달을 μœ„ν•œ ν™œμ„±ν™” ν•¨μˆ˜λ‘œ μŒκ³‘νƒ„μ  νŠΈ μ‹œκ·Έλͺ¨μ΄λ“œ(Hyperbolic tangent sigmoid) ν•¨μˆ˜λ₯Ό ν™œμš©ν•˜μ˜€μœΌλ©°, 좜λ ₯μΈ΅μ—μ„œλŠ” μˆœμˆ˜μ„ ν˜•ν•¨μˆ˜(Pure linear function)λ₯Ό μ μš©ν•˜μ—¬ 인곡신경망 ν•™μŠ΅μ„ μˆ˜ν–‰ν•˜μ˜€λ‹€(Hagan, 1994; Lee, 2016; Nielsen, 1989).

이 μ—°κ΅¬μ—μ„œ μ œμ•ˆν•œ μ•Œκ³ λ¦¬μ¦˜μ„ 기반으둜 예츑된 κ²°κ³Όκ°’($t_{n}$)은 μ•„λž˜μ™€ 같은 ν•¨μˆ˜λ‘œ ν‘œν˜„ν•  수 μžˆλ‹€.

(1)
$t_{n}=\left[1+f_{2}\left(W_{2}f_{1}\left(W_{1}\left(1-\left | 2\left(p_{n}-1\right)\right |\right)+b_{1}\right)+b_{2}\right)\right]/2$

μ—¬κΈ°μ„œ, $p_{n}$은 μ •κ·œν™”λœ μž…λ ₯κ°’, $W_{1}$, $W_{2}$ 및 $b_{1}$, $b_{2}$λŠ” 각각 κ°€μ€‘μΉ˜μ™€ μ—­μΉ˜, $f_{1}$은 μž…λ ₯μΈ΅κ³Ό 은닉측 κ°„μ˜ ν™œμ„±ν•¨μˆ˜λ₯Ό, $f_{2}$λŠ” 은닉측과 좜λ ₯μΈ΅ κ°„μ˜ ν™œμ„±ν•¨μˆ˜λ₯Ό λ‚˜νƒ€λ‚Έλ‹€. Table 3μ—λŠ” λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯의 인곡신경망 ν•™μŠ΅κ³Όμ •μ— ν™œμš©λœ λ°μ΄ν„°λ“€μ˜ μ΅œλŒ“κ°’κ³Ό μ΅œμ†Ÿκ°’μ„ λ‚˜νƒ€λ‚΄μ—ˆλ‹€.

Fig. 2 Freeze-thaw resistance prediction model
../../Resources/ksm/jksmi.2023.27.6.144/fig2.png
Fig. 3 ANN structure
../../Resources/ksm/jksmi.2023.27.6.144/fig3.png
Table 3 Range of input parameters and RDME

Parameters

Min

Max

Water-cement ratio, w/c

0.1

0.6

Sand-aggregate ratio, s/a

0.06

0.82

Air content (%)

1.1

7.0

Compressive strength, fc (MPa)

14.0

92.7

Number of freeze-thaw cycles

10.0

1143.0

Relative dynamic modulus of elasticity, RDME (%)

25.1

101.0

2.3 인곡신경망 ν•™μŠ΅κ²°κ³Ό

이 μ—°κ΅¬μ—μ„œλŠ” 인곡신경망 λͺ¨λΈ 및 νšŒκ·€μ‹μ˜ 정확도 ν‰κ°€μ§€ν‘œλ‘œ κ²°μ •κ³„μˆ˜μ™€ 평균 μ˜€μ°¨μœ¨μ„ ν™œμš©ν•˜μ˜€λ‹€. κ²°μ •κ³„μˆ˜(R2)λŠ” 주어진 데이터에 λŒ€ν•œ νšŒκ·€ λͺ¨λΈμ˜ μ„€λͺ…λ ₯을 λ‚˜νƒ€λ‚΄λŠ” μ§€ν‘œμ΄λ©°, 일반적으둜 κ²°μ •κ³„μˆ˜κ°€ λ†’μ„μˆ˜λ‘ λͺ¨λΈμ΄ 주어진 데이터λ₯Ό 더 잘 μ„€λͺ…ν•œλ‹€κ³  ν•  수 μžˆλ‹€. λ‹€λ§Œ, λΆ„μ•Ό 및 상황에 따라 인곡신경망 λͺ¨λΈμ— μš”κ΅¬ν•˜λŠ” 정확도가 크게 달라지며, 인곡신경망 및 콘크리트의 λ™κ²°μœ΅ν•΄ κ΄€λ ¨ μ—°κ΅¬μžλ“€(In, 2013)은 κ²°μ •κ³„μˆ˜κ°€ 0.70 이상이면 μƒλ‹Ήνžˆ 높은 정확도λ₯Ό κ°–λŠ”λ‹€κ³  λ³΄κ³ ν•˜κ³  μžˆλ‹€. λ˜ν•œ, 이 μ—°κ΅¬μ—μ„œλŠ” μ œμ•ˆ 인곡신경망 λͺ¨λΈμ˜ μ„±λŠ₯을 ν‰κ°€ν•˜κΈ° μœ„ν•΄ μΆ”κ°€μ μœΌλ‘œ 평균 μ˜€μ°¨μœ¨μ„ ν™œμš©ν•˜μ˜€μœΌλ©°, 평균 였차율의 κ²½μš°μ—λ„ κ²°μ •κ³„μˆ˜μ™€ λ§ˆμ°¬κ°€μ§€λ‘œ 정확도λ₯Ό ν™•λ³΄ν•˜μ˜€λ‹€κ³  νŒλ‹¨ν•  수 μžˆλŠ” μ˜€μ°¨μœ¨μ€ λΆ„μ•Ό, λ°μ΄ν„°μ˜ 뢄포 및 ν™˜κ²½μ— 따라 크게 달라진닀. 콘크리트의 λ™κ²°μœ΅ν•΄ κ΄€λ ¨ μ—°κ΅¬μžλ“€(Yang, 2002)은 μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜ 평가λͺ¨λΈμ˜ 였차율이 μ•½ 15% μ΄ν•˜μΈ κ²½μš°μ— μΆ©λΆ„ν•œ μ„±λŠ₯을 ν™•λ³΄ν•˜μ˜€λ‹€κ³  λ³΄κ³ ν•˜κ³  있으며, μ €μžλ“€μ€ 이λ₯Ό μ°Έκ³ ν•˜μ—¬ 인곡신경망 λͺ¨λΈ 및 νšŒκ·€μ‹μ˜ ν‰κ°€μ§€ν‘œλ‘œ ν™œμš©ν•˜μ˜€λ‹€.

인곡신경망 λͺ¨λΈμ˜ ν•™μŠ΅κ³Όμ •μ„ ν†΅ν•˜μ—¬ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 예츑 λͺ¨λΈμ˜ κ°€μ€‘μΉ˜μ™€ μ—­μΉ˜λ₯Ό λ„μΆœν•˜μ˜€μœΌλ©°, Table 4에 λ‚˜νƒ€λ‚΄μ—ˆλ‹€. Fig. 4 및 Table 5μ—λŠ” μ œμ•ˆ 인곡신경망 λͺ¨λΈμ˜ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ ν•™μŠ΅κ²°κ³Όλ₯Ό λ‚˜νƒ€λ‚Έ 것이닀. ν•™μŠ΅κ²°κ³Όμ˜ κ²°μ •κ³„μˆ˜(Coefficient of determination, R2)λŠ” 0.93, 평균 μ˜€μ°¨μœ¨μ€ μ•½ 5%λ₯Ό λ³΄μ˜€μœΌλ©°, μ΄λŠ” 인곡신경망 λͺ¨λΈμ΄ 맀우 μš°μˆ˜ν•œ ν•™μŠ΅κ²°κ³Όλ₯Ό λ„μΆœν•˜μ˜€λ‹€λŠ” 것을 μ˜λ―Έν•œλ‹€.

Fig. 4 ANN training results
../../Resources/ksm/jksmi.2023.27.6.144/fig4.png
Table 4 Weight and bias

Hidden neurons

Bias 1 (b1)

Weight 1 (W1)

Weight 2 (W2)

Water-cement ratio, w/c

Sand-aggregate ratio, s/a

Air content (%)

Compressive strength, fc (MPa)

Number of freeze-thaw cycles

Relative dynamic modulus of elasticity, RDME (%)

1

-1.409

-0.428

-3.987

-2.584

1.339

-0.784

2.600

2

4.206

2.227

-1.698

-0.059

0.090

-4.733

-6.563

3

-2.395

-2.257

-3.958

1.121

-2.783

-2.419

4.783

4

3.133

0.138

0.570

0.910

-2.289

1.299

-0.204

5

-1.988

0.689

3.581

3.100

0.219

1.519

-0.727

6

0.850

-4.592

-1.342

2.144

1.214

-4.618

-1.740

7

1.570

-1.840

-0.081

-1.733

4.216

-1.062

0.921

8

3.692

-0.314

1.702

2.181

2.961

-1.665

3.654

9

-2.460

-1.799

1.044

0.917

-1.357

1.799

-1.561

Bias 2 (b2) = 1.704

Table 5 Training results of ANN model

Case No.

Relative dynamic modulus of elasticity,

RDME (%)

[Experiment]

Relative dynamic modulus of elasticity,

RDME (%)

[Prediction]

Error, %

1

98

98.94

0.96

2

98

96.20

1.84

3

98

92.70

5.41

4

93

87.73

5.66

5

93

84.38

9.27

6

81

80.23

0.95

7

72

68.86

4.36

8

56

61.47

9.77

9

48

53.10

10.63

10

101

102.28

1.27

AVG.

4.91

2.4 인곡신경망 검증

λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ˜ 검증은 μ œμ•ˆ 인곡신경망 λͺ¨λΈμ˜ ν•™μŠ΅μ— ν™œμš©λ˜μ§€ μ•Šμ€ μ°Έκ³ λ¬Έν—Œμ—μ„œ μˆ˜μ§‘ν•œ 20개의 μƒˆλ‘œμš΄ λ°μ΄ν„°λ‘œ κ΅¬μ„±ν•˜μ—¬ μˆ˜ν–‰ν•˜μ˜€λ‹€(Trottier, 2022; Choi, 1997; Shang, 2012). Fig. 5 및 Table 6에 λ‚˜νƒ€λ‚Έ 바와 같이 검증 결과의 평균 였차율이 10.4%둜 λ‚˜νƒ€λ‚¬μœΌλ©°, κ²°μ •κ³„μˆ˜(R2)λŠ” 0.70으둜 λ‚˜νƒ€λ‚˜μ„œ μ œμ•ˆλœ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ΄ 높은 정확도λ₯Ό ν™•λ³΄ν•œ κ²ƒμœΌλ‘œ νŒλ‹¨λœλ‹€.

Fig. 5 ANN validation results
../../Resources/ksm/jksmi.2023.27.6.144/fig5.png
Table 6 Validation results of ANN model

No.

Water-cement ratio, w/c

Sand-aggregate ratio, s/a

Air content (%)

Compressive strength, fc (MPa)

Number of freeze-thaw cycles

Relative dynamic modulus of elasticity, %

(Experiment)

Relative dynamic modulus of elasticity, %

(Predicted)

Error

(%)

1

0.35

0.51

6.5

19

100

65.3

60.4

7.5

2

0.35

0.51

6.5

19

150

52.8

66.2

25.4

3

0.35

0.53

6.5

20

50

78.5

70.9

9.6

4

0.35

0.53

6.5

20

100

67.2

75.3

11.9

5

0.35

0.52

6.5

19

50

82.4

61.4

25.4

6

0.35

0.53

6.5

21

50

81.9

74.4

9.2

7

0.35

0.53

6.5

21

100

69.4

78.1

12.4

8

0.35

0.71

6.5

26

50

97.1

112.9

16.3

9

0.35

0.73

6.5

27

50

114.1

114.2

0.1

10

0.35

1.09

6.5

41

50

97.3

110.2

13.3

11

0.35

1.14

6.5

42

150

67.0

65.9

1.6

12

0.35

1.11

6.5

62

150

68.5

63.0

8.0

13

0.45

0.54

2

26.5

300

76.8

87.4

13.8

14

0.45

0.54

2

26.5

258

83.2

79.0

5.0

15

0.45

0.54

2

26.5

268

83.1

81.1

2.5

16

0.45

0.54

2

26.5

276

80.5

82.6

2.6

17

0.45

0.54

2

26.5

280

79.9

83.5

4.5

18

0.45

0.54

2

26.5

292

79.8

85.9

7.6

19

0.550

0.49

1.9

42.5

25

74.1

70.5

4.9

20

0.550

0.49

1.9

42.5

50

46.8

59.1

26.3

AVG.

10.40

3. νšŒκ·€λΆ„μ„μ„ ν†΅ν•œ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈ

λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ μ‚°μ •μ‹μ˜ λ³€μˆ˜λ₯Ό μ„€μ •ν•˜κΈ° μœ„ν•˜μ—¬ 각 λ³€μˆ˜μ— λŒ€ν•œ μ€‘μš”λ„ 뢄석을 μˆ˜ν–‰ν•˜μ˜€λ‹€. Fig. 6에 λ‚˜νƒ€λ‚Έ 바와 같이 κ³΅κΈ°λŸ‰, λ¬Ό/μ‹œλ©˜νŠΈλΉ„, 압좕강도 및 μž”κ³¨μž¬λΉ„ μˆœμ„œλ‘œ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯에 큰 영ν–₯을 λ―ΈμΉ˜λŠ” κ²ƒμœΌλ‘œ λ‚˜νƒ€λ‚¬λ‹€.

μ€‘μš”λ„ 뢄석과 ν•¨κ»˜ Fig. 7에 λ‚˜νƒ€λ‚Έ 바와 같이 λ¬Ό/μ‹œλ©˜νŠΈλΉ„, μž”κ³¨μž¬μœ¨, κ³΅κΈ°λŸ‰, 콘크리트의 압좕강도 및 λ™κ²°μœ΅ν•΄ 사이클 μˆ˜μ™€ μƒλŒ€λ™νƒ„μ„±κ³„μˆ˜μ— λŒ€ν•œ 상관관계λ₯Ό λΆ„μ„ν•˜μ˜€μœΌλ©°, 상관관계 뢄석 κ²°κ³Όλ₯Ό 기반으둜 νšŒκ·€λΆ„μ„μ„ 톡해 식 (2)와 같은 λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ ($RDME$) 평가식을 λ‹€μŒκ³Ό 같이 λ„μΆœν•˜μ˜€λ‹€.

(2)
$RDME =\dfrac{A\times\beta^{0.2}\times f_{c}^{0.3}}{0.15\times\alpha\times N^{0.2}}$

μ—¬κΈ°μ„œ, AλŠ” κ³΅κΈ°λŸ‰, π›½λŠ” μž”κ³¨μž¬μœ¨, fcλŠ” 압좕강도, π›ΌλŠ” λ¬Ό/μ‹œλ©˜νŠΈλΉ„, N은 λ™κ²°μœ΅ν•΄ 사이클 수λ₯Ό μ˜λ―Έν•œλ‹€. 식 (2)에 μ˜ν•œ μ‚°μ •κ°’ μ‹€ν—˜κ²°κ³Όκ°’μ„ λΉ„κ΅ν•œ κ²°κ³Ό, Table 7κ³Ό Fig. 8에 λ‚˜νƒ€λ‚Έ 바와 같이 λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ ν‰κ°€μ‹μ˜ 평균 였차율과 κ²°μ •κ³„μˆ˜(R2)λŠ” 각각 11.8%와 0.72둜 λ‚˜νƒ€λ‚¬λ‹€.

Fig. 6 Importance evaluation results
../../Resources/ksm/jksmi.2023.27.6.144/fig6.png
Fig. 7 Correlation for each variables with RDME
../../Resources/ksm/jksmi.2023.27.6.144/fig7.png
Fig. 8 Results of Resistance evaluation equation
../../Resources/ksm/jksmi.2023.27.6.144/fig8.png
Table 7 Comparison between equation with experiment

No.

Water-cement ratio, w/c

Sand-aggregate ratio, s/a

Air content (%)

Compressive strength, fc (MPa)

Number of freeze-thaw cycles

Relative dynamic modulus of elasticity, %

(Experiment)

Relative dynamic modulus of elasticity, %

(Equation)

Error

(%)

𝛼

𝛽

A

fck

N

RDME

RDME

1

0.320

0.613

3.1

81.1

50

98.0

100.1

2.2

2

0.320

0.613

3.1

81.1

300

72.0

70.0

2.8

3

0.250

0.613

2.9

91.4

150

97.0

99.7

2.8

4

0.250

0.613

2.9

91.4

200

96.0

94.2

1.9

5

0.250

0.613

2.9

91.4

225

95.0

92.0

3.2

6

0.250

0.613

2.9

91.4

250

92.0

90.1

2.1

7

0.400

0.600

4

40

25

95.1

95.6

0.5

8

0.400

0.600

4

40

50

86.6

83.2

3.9

9

0.400

0.600

4

40

150

68.3

66.8

2.2

10

0.550

0.539

4

55

10

96.2

89.9

6.5

AVG.

11.8

4. κ²° λ‘ 

이 μ—°κ΅¬μ—μ„œλŠ” κΈ°μ‘΄ 연ꡬλ₯Ό ν† λŒ€λ‘œ 콘크리트 λ™κ²°μœ΅ν•΄ μ‹€ν—˜ λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό κ΅¬μΆ•ν•˜μ˜€μœΌλ©°, 이λ₯Ό ν™œμš©ν•˜μ—¬ 인곡신경망 평가λͺ¨λΈ 및 νšŒκ·€λΆ„μ„ 산정식을 μ œμ•ˆν•˜μ˜€λ‹€. 이 연ꡬλ₯Ό ν†΅ν•˜μ—¬ λ‹€μŒκ³Ό 같은 결둠을 얻을 수 μžˆμ—ˆλ‹€.

1. κ΅¬μΆ•λœ λ°μ΄ν„°λ² μ΄μŠ€λ₯Ό 기반으둜 인곡신경망 λͺ¨λΈμ„ ν™œμš©ν•˜μ—¬ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ„ λ„μΆœν•˜μ˜€λ‹€. λ„μΆœλœ 인곡신경망 기반 콘크리트 λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯ 평가λͺ¨λΈμ˜ 평균 였차율 및 κ²°μ •κ³„μˆ˜λŠ” 각각 10.4% 및 0.70둜 λ‚˜νƒ€λ‚˜ 평가λͺ¨λΈμ΄ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯을 맀우 μš°μˆ˜ν•œ μ •ν™•λ„λ‘œ ν‰κ°€ν•˜λŠ” κ²ƒμœΌλ‘œ νŒλ‹¨λœλ‹€.

2. λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ μ‚°μ •μ‹μ˜ λ³€μˆ˜λ₯Ό μ„€μ •ν•˜κΈ° μœ„ν•˜μ—¬ 각 λ³€μˆ˜μ— λŒ€ν•œ μ€‘μš”λ„ 뢄석을 μˆ˜ν–‰ν•˜μ˜€λ‹€. κ³΅κΈ°λŸ‰(μ€‘μš”λ„: 0.30), λ¬Ό/μ‹œλ©˜νŠΈλΉ„(μ€‘μš”λ„: 0.23), 압좕강도(μ€‘μš”λ„: 0.15) 및 μž”κ³¨μž¬λΉ„(μ€‘μš”λ„: 0.13) μˆœμ„œλ‘œ λ™κ²°μœ΅ν•΄ μ €ν•­μ„±λŠ₯에 λ―ΈμΉ˜λŠ” 영ν–₯이 큰 κ²ƒμœΌλ‘œ λ‚˜νƒ€λ‚¬λ‹€.

3. νšŒκ·€λΆ„μ„μ„ 톡해 λ„μΆœλœ λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯ μ‚°μ •μ‹μ˜ 평균 였차율과 κ²°μ •κ³„μˆ˜(R2)λŠ” 각각 11.8%와 0.72둜 λ‚˜νƒ€λ‚˜μ„œ, μ‹€μš©μ μ΄λ©΄μ„œλ„ 비ꡐ적 μš°μˆ˜ν•œ 정확도λ₯Ό ν™•λ³΄ν•œ κ²ƒμœΌλ‘œ νŒλ‹¨λœλ‹€.

4. 이 μ—°κ΅¬μ—μ„œλŠ” λ™κ²°μœ΅ν•΄ μž‘μš©μ„ λ°›λŠ” 콘크리트의 μ €ν•­μ„±λŠ₯에 큰 영ν–₯을 λ―ΈμΉœλ‹€κ³  μ•Œλ €μ§„ ν•„μˆ˜μ μΈ λ³€μˆ˜λ§Œμ„ ν† λŒ€λ‘œ 평가λͺ¨λΈμ„ μ œμ•ˆν•˜μ˜€μ§€λ§Œ, μΆ”ν›„ λ‹€μ–‘ν•œ ν˜Όν™”μ œμ— λŒ€ν•œ λ³€μˆ˜λ₯Ό μΆ”κ°€ν•¨μœΌλ‘œμ¨ μ œμ•ˆλͺ¨λΈμ˜ μ μš©λ²”μœ„κ°€ λ”μš± ν™•λŒ€λ  κ²ƒμœΌλ‘œ κΈ°λŒ€λœλ‹€.

κ°μ‚¬μ˜ κΈ€

이 μ„±κ³ΌλŠ” μ •λΆ€(κ³Όν•™κΈ°μˆ μ •λ³΄ν†΅μ‹ λΆ€)의 μž¬μ›μœΌλ‘œ ν•œκ΅­μ—°κ΅¬μž¬λ‹¨μ˜ 지원을 λ°›μ•„ μˆ˜ν–‰λœ μ—°κ΅¬μž„(No. RS-2023-00220019).

References

1 
Neville, A. M. (1995), Properties of Concrete, Fourth and Final Edition, Logman.URL
2 
Lee, S. T., and Park, K. P. (2018), Resistance to Freezing and Thawing of Concrete Subjected to Carbonation, Journal of Korea Academia-Industrial cooperation Society, 19(2), 623-631.DOI
3 
Choi, H. J., Kim, R. R., Lee, J. S., and Min, J. Y. (2021), Evaluation of Freeze-Thaw Damage on Concrete Using Nonlinear Ultrasound, ournal of the Korea Institute for Structural Maintenance and Inspection, 25(4), 56-64.DOI
4 
Yoon, Y. G., Lee, I. B., Sa, M. H., and Oh, T. K. (2017), A Study on the Statistical Distribution of Ultrasonic Velocities for the Condition Evaluation of Concrete Wide Beam, Journal of Korean Society of Safety, 32(2), 98-104.DOI
5 
You, Y. C., Choi, K. S., and Kim, K. H. (2010), Freezing-Thawing Resistance of Fiber Reinforced Polymers in Strengthening RC Members, Journal of the Korea Institute for Structural Maintenance and Inspection, 14(1), 182-189.DOI
6 
Yan, W., Wu, Z., Niu, F., Wan, T., and Zheng, H. (2020), Study on the service life prediction of freeze- thaw damaged concrete with high permeability and inorganic crystal waterproof agent additions based on ultrasonic velocity, Constuction and Building Materials, 259, 1-11.DOI
7 
Nimlyat, P. S., Audu, A. U., Ola-Adisa, E. O., and Gwatau, D. (2017), An evaluation of the fire safety measures in high-rise buildings in Nigeria, Sustainable Cities and Society, 35, 774-785.DOI
8 
Baek, E. S., Baek, G. J., Shin, H., Song, M. J., Kook, C. H., Kim, S. W. (2010), A study on the awareness of fire safety and evacuation guide system, Jornal of Korean Institute of Fire Science and Engineering, 24(6), 45-53.URL
9 
Stankovi, G, Petelin, S., Vidmar, P., and Perkovi. M., (2018), Impact of LNG Vapor Dispersion on Evacuation Routes inside LNG Terminals, Journal of Mechanical Engineering, 64(3), 176-184.URL
10 
Popescu, I., Nikitopoulos, D.,Constantinou, P., and Nafornita, I. (2006), Comparison of ANN Based Models for Path Loss Prediction in Indoor Environment, IEEE Vehicular Technology Conference.DOI
11 
Yu, H., Ma, H., and Yan, K., (2017), An equation for determining freeze-thaw fatigue damage in concrete and a model for predicting the service life, Construction and Building Materials, 137, 104-116.DOI
12 
Shang, H., Song, Y., and Ou, J., (2009), Behavior of Air-Entrained Concrete After Freeze-Thaw Cycles, Acta Mechanica Solida Sinica, 22(3), 261-266.DOI
13 
Shang, H. S., and Yi, T. H., (2013), Freeze-Thaw Durability of Air-Entrained concrete, The Scientific World Journal, 2013, 1-6.DOI
14 
Wu, H., Liu, Z., Sun, B., and Yin, J., (2016), Experimental investigation on freeze-thaw durability of Portland cement pervious concrete (PCPC), Construction and Building Materials, 117, 63-71.DOI
15 
Duan, A., Tian, Y., Dai, J. G., and Jin, W. L., (2013), A stochastic damage model for evaluating the internal deterioration of concrete due to freeze-thaw action, Materials and Structures, 47, 1025-1039.DOI
16 
Cho, T., (2007), Prediction of cyclic freeze-thaw damage in concrete structures based on response surface method, Construction and Building Materials, 21(12), 2031-2040.DOI
17 
Janssen, D. J., and Snyder, M. B. (1994), Resistance of Concrete to Freezing and Thawing, Highway Research Program National Research Council, Washington DC.URL
18 
Russell, S., Norvig, P. (2010), Artificial Intelligence A Modern Approach.URL
19 
Cho, H. C., Lee, D. H., Ju, H. J., Kim, K. S., Kim, K. H., and Paulo, J. M. (2015), Monteiro. Remaining Service Life Estimation of Reinforced Concrete Buildings based on Fuzzy Approach, Computers and Concrete, 15(6), 879-902.URL
20 
Cho, H. C., Lee, D. H., Ju, H. J., Park, H. C., Kim, H. Y., and Kim, K. S. (2017), Fire Damage Assessment of Reinforced Concrete Structures Using Fuzzy Theory, Applied Sciences, 7(5), 1-16.DOI
21 
Kang, H., Cho, H. C, Choi, S. H., Heo, I. W., Kim, H. Y., and Kim, K. S. (2019), Estimation of Heating Temperature for Fire- Damaged Concrete Structures Using Adaptive Neuro-Fuzzy Inference System, Materials, 12, 1-17.DOI
22 
Darkhanbat, K, Heo, I. W., Han, S. J., Cho, H. C., and Kim, K. S. (2021), Real-Time Egress Model for Multiplex Buildings under Fire Based on Artificial Neural Network, Applied Sciences, 11 1-22.DOI
23 
Hagan, M. T., and Menhaj, M. (1994), Training feedforward networks with the Marquardt algorithm, IEEE Transactions on Neural Networks, 5(6), 989-993.DOI
24 
Lee, J. S., and Suh, K. D. (2016), Calculation of Stability Number of Tetrapods Using Weight and Biases of ANN Model, Journal of Korean Society of Coastal and Ocean Engineers, 28(5), 277-283.DOI
25 
Nielsen, R. H. (1989), Theory of the Backpropagation Neural Network, IEEE IJCNN, New York , 1593-1605.DOI
26 
Trottier, C., Grazia, M. T., Macedo, H. F., Sanchez, L. F. M., Andrade, G. P. de., Souza. D. J. de., Naboka, O., Fathifazl, G., Nkinamubanzi, P. C., and Demers, A. (2022), Freezing and Thawing Resistance of Fine Recycled Concrete Aggregate (FRCA) Mixtures Designed with Distinct Techniques, Materials, 15(4), 1-23.DOI
27 
Choi, S. G., and Kim, S. B. (1997), A Study on Freezing and Trawing Resistance of Concrete with the Ratio of Ground Granulated Blast-Furnace Slag Replacement, Journal of Korea Concrete Institute, 9, 149-155.DOI
28 
Shang, H. S., Yi, T. H., and Song,Y. P. (2012), Behavior of Plain Concrete of a High Water-Cement Ratio after Freeze-Thaw Cycles, Materials, 5(9), 1698-1707.DOI
29 
Zhao. N., and Zhang, A. (2013), Prediction of Concrete Freezing Resistance under Site Environment based on ANN, Applied Mechanics and Materials, 438-439, 202-206.DOI
30 
Im, S. Y., and Chun. B. S. (2013), The Prediction of Flow and Strength of Controlled Low-Strength Material Using Artificial Neural Networks, Research Institute of Korean Traditional Dance, 3, 164-183.URL
31 
Yang, S. I., Yoon, Y. S., Lee, S. H., and Kim, G. D., (2002), High Performance Concrete Mixture Design using Artificial Neural Networks, Korea Concrete Institute, (2002.05a), 545-550.URL