Global-local progressive failure analysis of composite panels including skin-stringer debonding and intralaminar damage

Abstract

An increasing application of fibre-reinforced composites in aircraft and aerospace engineering contributed to the rise of development of highly effective numerical tools. On the one hand, computational simulations help to overcome the problems related to expensive experimental testings by partially substituting them. This in turn could result in decrease of design and certification costs. On the other hand, the need of effective numerical methods emerged from the desire to increase the limit loads and thus, to further exploit possible reserves of composite components. However, the complexity of problems commonly associated with the modelling of composites requires thorough understanding of the material and structural behaviour. For this reason, an efficient and reliable progressive failure analysis capability is required. Moreover, it is indispensable to be able to combine different levels of precision in one model to effectively examine large structures. Extensive usage of composite stiffened panels is justified by their slenderness, which results in desired light weight jointly with high stiffness in designated direction due to accordingly aligned reinforcing parts called stringers or stiffeners. These stiffeners not only prevent the skin of the panel from premature buckling under compressive loading, but they also increase the overall structural strength leading to final failure detected far beyond initial buckling. Successfully employed in modelling fuselages and wing boxes as primary components, composite stiffened panels have gained recognition. However, substantial enhancements are to be envisaged in terms of computational methods, as various effects, such as damage initiation and propagation, plasticity or impact damage, for example, require investigation at different levels of accuracy. That leads to development of various multiscale and global-local methods. A literature review has been conducted with a special attention to damage mechanisms and their importance as well as a following discussion dedicated to the existing multiscale algorithms with their strong and weak points. To address a need in computationally efficient strategies, during this work a novel global-local coupling approach has been developed that is able to model progressive separation of the skin and the stringer together with intralaminar damage in stiffened CFRP panels under compression. The main goal of this methodology is to examine the damage at two levels of accuracy, taking advantage of the fast calculations at the global level and assessing in detail the damage propagation at the local level. An appropriate information exchange between the global and local levels in both directions is particularly challenging and it has been achieved in the demonstrated global-local approach. According to the proposed method, at the global level a linear elastic coarse model with shell elements is employed to detect the probable areas of damage. Afterwards, local models are generated using a fine mesh with solid elements. Kinematic constraints are used as boundary conditions to prescribe corresponding displacements from the global to the local model. The damage evolution is simulated by means of a material degradation model and cohesive elements are applied at the local level. To ensure a full information exchange between the two levels, a transfer of the reduced material properties from the local to the global level is carried out. As the relative element sizes of the global and local models are different, a special homogenization procedure is implemented that preserves energies dissipated between the local and global levels. The global-local steps are executed until the final failure takes place. The new developed approach is illustrated on the basis of one-stringer and multi-stringer laminate panels considering intact and predamaged cases with initial debonding with an aim to demonstrate advantages of the proposed method for modelling progressive failure in the stiffened composite panels with localized damage.