LSTM 순환 신경망을 이용한 재료의 단축하중 하에서의 응력-변형률 곡선예측 연구
Prediction of the Stress-Strain Curve of Materials under Uniaxial Compression by Using LSTM Recurrent Neural Network
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
이 논문에서는 재료의 단축하중 하에서의 응력-변형률 곡선을 예측하기 위하여 순환 신경망의 일종인LSTM(Long Short-Term Memory) 알고리즘을 사용하였다. 석고와 규사를 혼합해 만든 재료에 일축압축시험을 수행하여 얻은 응력-변형률 데이터를 이용하였으며, 낮은 응력 구간의 초반 데이터를 활용해서 파괴 전까지의 거동을 예측하였다. 앞부분의 데이터를 활용하여 단계적으로 뒤쪽 구간의 값을 예측하는 LSTM 순환 신경망의 구조상 큰 응력에 대응하는 변형률을 예측할 경우에는 앞쪽 구간의 오차가 누적되어 실측값과 차이가 늘어났으나 전반적으로 높은 정확도로 응력-변형률 곡선을 예측하였다. 예측에사용한 초기 데이터의 길이가 늘어나는 경우 정확도는 조금 증가했다. 그러나 접선을 이용한 단순 예측과의 성능 차이는 초기 데이터의 길이가 작은 경우에 두드러졌으며, 적은양의 데이터로도 응력-변형률곡선 전체 구간의 예측을 가능하게 한다는 점으로부터 신경망 모델의 필요성을 확인하였다.
Abstract
LSTM (Long Short-Term Memory) algorithm which is a kind of recurrent neural network was used to establish a model to predict the stress-strain curve of an material under uniaxial compression. The model was established from the stress-strain data from uniaxial compression tests of silica-gypsum specimens. After training the model, it can predict the behavior of the material up to the failure state by using an early stage of stress-strain curve whose stress is very low. Because the LSTM neural network predict a value by using the previous state of data and proceed forward step by step, a higher error was found at the prediction of higher stress state due to the accumulation of error. However, this model generally predict the stress-strain curve with high accuracy. The accuracy of both LSTM and tangential prediction models increased with increased length of input data, while a difference in performance between them decreased as the amount of input data increased. LSTM model showed relatively superior performance to the tangential prediction when only few input data was given, which enhanced the necessity for application of the model.
참고문헌 (12)
- Arindam, D., "Estimation of Burger model parameters using inverse formulation", International Journal of Geotechnical Engineering, 6(3), pp.261~274 (2012)[DOI]
- 변훈, "인공 퇴적암의 모래입자 크기와 함량이 역학적 성질에 미치는 영향에 관한연구", 터널과 지하공간, 28(2), pp.156~169 (2018), 한국암반공학회[DOI]
- Gers, F., "Learning precise timing with LSTM recurrent networks", Journal of Machine Learning Research, 3, pp.115~143 (2002)
- Guan, Z., "A new rheological model and its application in mountain tunnelling", Tunnelling and Underground Space Technology, 23(3), pp.292~299 (2008)[DOI]
- Hardy, H.R., "Analysis of the inelastic deformation of geologic materials in terms of mechanical models", Spring Meeting, Society for Experimental Stress Analysis (1967)
- Hudson, J.A., "Engineering Rock Mechanics: an introduction to the principle", pp.215~ (1997), Elsevier Science
- Karami, M., "A new time-dependent constitutive model and its application in underground construction", ISRM International Symposium on Rock Mechanic for Resources, Energy and Environment, pp.21~26 (2013)
- Li, Y., "damage rheology model applied to analysis of splitting failure in underground caverns of Jinping I hydropower station", International Journal of Rock Mechanics and Mining Sciences, 71, pp.224~234 (2014)[DOI]
- Perez-Ortiz, J., "Kalman filters improve LSTM network performance in problems unsolvable by traditional recurrent nets", Neural Networks, 16(2), pp.241~250 (2003)[DOI]
- Skrzypek, J.J., "Constitutive Equations for Isotropic and Anisotropic Linear Viscoelastic Materials, Mechanics of Anisotropic Materials, Engineering Materials", pp.57~85 (2015), Springer
- Zaremba, W., "Recurrent neural network regularization", ~ (2014)
- Zhao, D., "Displacement prediction of tunnels based on a generalised Kelvin constitutive model and its application in a subsea tunnel", Tunnelling and Underground Space Technology, 54, pp.29~36 (2016)[DOI]