JEC COMPOSITES MAGAZINE - Issue #74 - July 2012 - 47

knock-down factors are applied to design structures following a safe-life concept. However, to establish the durability of a structure, quite a few cases must be covered in the strength evaluation. One can think of the applicability of the S-N data obtained with specific lay-ups to a structure where other lay-up configurations are used. Subsequently, variable-amplitude load spectra must be related to constant-amplitude load data, including the effect of mean stress variations. In addition, the ageing of the polymer composite, the exposure to environmental temperature and moisture has to be accounted for in high cycle fatigue, not to mention the potential deviation of primary load paths, or load deviations. To substantiate the applied knock-down factors, fatigue tests are often duplicated to test the same specimen configuration under the same applied loads, but in different environmental conditions. This implies performing many long-term fatigue tests to obtain the necessary data for validating the applied knock-down factors; consider, for example, the design lives of wind turbine blades that may be in the range of 108-109 cycles. In other words, a large database of S-N data is often created, which must then be correlated by empirical relations to substantiate certain design configurations [3].

tion of changes in structural usage. Potential heavier usage leads to increased load spectra for which damage tolerance cannot simply be validated with the component or full-scale test performed. Similarly, the increase in design life, which is commonplace these days with aeronautical structures considering the work being performed on life extension programs, cannot be substantiated with this component or full-scale test either. The fact that for a given design life no growth has been observed for certain structural damage does not necessarily imply that extending that life will not show damage either. There is a validation limit to the no-growth approach, which is determined by the life for which the component test has been performed.

Mechanistic description approach
Both cases clearly indicate the limits of the current approach to the strength evaluation of composite structures. Although incorrect, it is often assumed that a more constructive approach requires complex modelling and prediction methods with the necessity to use micro-mechanical modelling and multi-scale modelling. In the following sections, previous achievements and results on polymer composites and fibre metal laminates (FML) are used to illustrate that the way towards analysing composite structures is more straightforward than it seems. Initially, this may lead to a reduction in the amount of tests, with the tests being more specific in nature, but therefore more generic in application. Let us look in more detail at general fatigue tests on composites to observe what happens between the first load cycle and final failure. Some researchers indicate S-N test regimes for which specific failure modes predominate. In Fig. 2, for example, transverse-ply cracking, delamination and fibre fracture are illustrated with respect to the initial stiffness of the material. Although this representation in principle is correct, it apparently does not completely describe the observed and sometimes reported phenomena. ‘Dominant’ was therefore added in Fig. 2, because a distinction between the failure mechanisms is Fig. 2: The dominant failure modes in fatigueoften not possible loaded composites [2] due to interactions between mechanisms. This can be illustrated with the upper curve in Fig. 2; damage forms initially by transverse ply cracking which at some moment in time deflects at ply interfaces into delamination that No74 July 2012 /

Damage tolerance approach
Because initial or impact damage cannot always be prevented during operation, designers have to consider the application of damage tolerant design principles. To consider the damage tolerance of a structure, the aforementioned S-N data cannot be used straightforwardly, because the way the damage progresses is different from the way damage forms in an un-notched fatigue life specimen until its failure. Yet, the aeronautical industry applies no-growth design concepts, because growth cannot be predicted with theoretical analysis and requires too many component tests to be validated based on tests. With a no-growth concept, it is considered sufficient to perform component or full-scale fatigue tests with the maximum damage that could potentially be expected at the maximum applied loads. In other words, if the upper damage limit together with the upper load limit in a component or full-scale test does not lead to damage growth, the design is considered covered with the no-growth concept. This approach, however, has several limitations, if not complications. The component or full-scale fatigue test cannot be used to evaluate what consequences greater or different damage could have. In addition, because of their size, these tests cannot include environmental influences. These influences must be considered by the above-mentioned – but for the current case less representative – fatigue tests that include environmental aspects. Another complication is the evalua-

jec composites magazine 47

JEC COMPOSITES MAGAZINE - Issue #74 - July 2012

Table of Contents for the Digital Edition of JEC COMPOSITES MAGAZINE - Issue #74 - July 2012