NUMERICAL SIMULATION OF BOX FOLDS IN MULTILAYERS

Box folds are characterized by two separate axial surfaces separated by a broad flat zone (i.e. they are double-hinged). These unique geometric features indicate specific deformation mechanisms during folding. It is generally understood that box folds result from the buckling of conjugate kink bands in anisotropic multilayers. In addition, buckling of modeling clay layers in laboratory experiments (with and without competence contrasts) has shown that flexural slip may play a role in the development of box folds. In this study, 2D plane strain finite element analysis is used to simulate visco-elastic multilayer buckle folding for a variety of conditions. In order to quantitatively evaluate fold shape (and to confirm the development of box folds), the relationship between aspect ratio, P (P= 2*amplitude/wavelength), and dip angle at the inflection point is used. In effective single layer setups (where the folding layers have the same competence) buckle folds develop for sinusoidal initial perturbations and conjugate kink band perturbations, and the occurrence of flexural slip significantly affects the resulting box fold shape. Increasing amounts of slip result in flat-topped box folds including M-shaped subsidiary folds. Lower amounts of slip result in more rounded box folds. For true multilayers setups (i.e. the folding layers have alternating competence), box folds form for a variety of initial perturbation geometries and material properties in the center of the multilayer fold stack. The results show that for sinusoidal, chevron and random white noise initial perturbations, box folds gradually develop through a progression from sinusoidal to parabolic to box folds. Key control parameters for the development of box folds are the ratio of incompetent layer thickness to competent layer thickness, s/h, the viscosity contrast of competent to incompetent layers, the friction coefficient, and the number of folding layers. The resulting box fold shapes from the numerical models are in good agreement with observations from the field and modeling clay laboratory experiments and thus enable a more quantitative assessment of dynamic fold development.