نوع مقاله : یادداشت فنی
نویسندگان
دانشکدهی مهندسی عمران، دانشگاه تبریز
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
In spite numerous investigations about bearing capacity of a driven pile in noncohesive soil, calculation of that is a complicated trend. Any prediction from numerical analysis was highly dependent on the model adopted for modeling the soil behavior. However, setting up a realistic model that would be able to calculate the Load carrying capacity of pile is rather difficult. Most of the researches show that the capability (i.e. pattern recognition and memorization) of an ANN is suitable for inherent uncertainties and imperfections found in geotechnical engineering problems considering its successful application without any restriction. The combination of the wavelet transforms theory with the basic concept of neural networks leads to a new mapping network called
neural network adaptive wavelets or wavenets which is proposed as an alternative to feed-forward neural networks for approximating arbitrary nonlinear functions. A wavelet network is a feed-forward neural network using wavelets as activation functions of its hidden layers neurons. In this network, both the position and dilation of the wavelets are optimized beside the weights. In one special approach of this network construction so called
wavenet, the position and dilation of the wavelets are fixed and the weights are optimized. In this research, considering late mentioned procedure for available experimental data, the potential for applying neural network and its adaptive wavelets (wavenets) has been shown for predicting the load carrying capacity of a pile driven in noncohesive soil. The validation tests showed the artificial intelligence solutions clearly outperformed in predictive accuracy under varying training and testing conditions and these methods can be employed for predicting the load carrying capacity of a pile driven in noncohesive soil in comparison with other computational and time consuming methods considering complexity of the soil characteristic. Numerical results indicate that substituting wavelet function as feed-forward neural network transfer functions can enhance the network performance and efficiency. Therefore proposed wavenet with feedforward neural network structure (wavenet) that uses SLOG1 wavelet function as its hidden layer activation functions is much better in comparison to the standard feed-forward in terms of performance generality .
)