인공신경망을 이용한 굴패각 혼합토의 비배수 전단거동 예측
(The) prediction of undrained shear behavior of mixed soil with oyster shells by using Artificial Neural Network
인공신경망 굴패각 혼합토;
- 원문 URL
The existing structural models can be easily understood by the obvious mathematical formulas and assumptions, which is a good point, many material constants should be decided in advance. To decide these parameters, additional experiments and the best suited technology are needed. Because of that, the reliance of the prediction of shear behavior relatively falls and the empirical approach to the experimental results is becoming important. Especially, in overconsolidation, the shear behavior of the mixed soil with oyster shells is very hard to predict, because it has a very complicated form by the moisture content, structure, stress state, stress history of the mixed soil. In this study, the shear behavior of the mixed soil with oyster shells based on a proper property of the mixed soil with oyster shells and overconsolidation test is predicted to apply the dynamic neural network, which has a feedback process from hidden layer or output layer to input layer and the constitutive equation. According to the result of predicting the undrained shear behavior of overconsolidated sample by using Artificial Neural Network, stress-strain behavior and pore water pressure are predicted analogously to the real behavior. A side face in a portrait, the constitutive equation is not predicted analogously to the real behavior. Also the comparison of the actual measurement and estimated figures shows that E-J model comes into more close contact to the actual measurement.