Fractures in the forms of joints and micro-cracks are commonly found in natural rocks, and their failure mechanism strongly depends on the crack coalescence pattern between pre-existing flaws. The determination of the failure behavior of non-persistent joints is an engineering problem that involves several parameters as mechanical properties of rock, normal stress and the ratio of joint surface to total shear surface. The impact of these parameters on the Crack coalescence is investigated through the use of computational tools called Neural Networks. A number of networks of threshold logic units were tested, with adjustable weights. The computational method for the training process was a back-propagation learning algorithm. In this paper, the input data for crack coalescence consists of values of geotechnical and geometrical parameters. As an output, the network estimates the crack type coalescence (i.e. mode I, mode II or mode I-II) that can be used for stability analysis of rock mechanic structures. The performa
Since their introduction, research into the area of artificial neural networks and their applications continue to captivate scientists and engineers from a variety of disciplines. This growing interest among researchers is stemming from the fact that these learning machines have an excellent performance in the issues of non-linear function approximation, data classification, clustering, and non-parametric regression or as simulations of the behavior biological neurons and pattern recognition. This paper investigates the validity of utilizing artificial neural networks in the prediction of the crack coalescence mode (i.e. mode I, mode II or mode I-II) that can be used for stability analysis of rock mechanic structures. According to Jing and Hudson [1], Jing [2] all numerical modeling methods (analytical methods, basic numerical methods, Finite Element Method, Boundary Element Method, Distinct Element Method, hybrid methods, extended numerical methods and fully coupled models) attempt to achieve one-to-one mechanism mapping in the model. In other words, a one-to-one mechanism occurring in reality is modeled directly such as a clear stress–strain relationship. The term ‘one-to-one mapping’ refers to the direct modeling of geometry and physical mechanisms, either specifically or through equivalent properties. The neural network approach is a ‘non one-to-one mapping’ method. In such a model, mechanism mapping is not totally direct. This model provides predicting capabilities; this is why it has been used for rock and soils parameter identification and prediction. The neural network modeling approach has already been applied to a variety of subjects in rock and soil mechanics (Millar and Hudson, [3]; Goh, [4]; Sklavounos and Sakellariou, [5]; Goh, [6]; Najjar and Basheer, [7]; Gangopadhyay et al., [8]; Deng and Lee, [9]; Chang et al., [10]; Yang and Rosenbaum, [11]) The approach to the problem of crack coalescence from the 268 perspective of artificial neural networks is not an easy task and requires sophisticated modeling techniques, experience, deep knowledge of engineering and a vast amount of experimental data.
