Characterization of the fatigue behaviour of fibre-reinforced plastics through accelerated fatigue tests of composite beams and plates with electromagnetic shakers

Increasing oil prices have intensified the trend to use composite materials more and more in the so-called primary components of cars and in aerospace applications. As a consequence the fatigue behaviour of composites becomes an issue again. During the last few decades, the problem of fatigue of fibre-reinforced plastics was partially avoided because the application of composites was mainly limited to so-called secondary components and the maximum strain in composite components was arbitrarily limited to 4000 microstrain (fatigue limit for BVID: Bearly Visible Impact Damage). This threshold was so low that fatigue damage was rarely a problem and design could be limited to static loading conditions [1-3]. For primary components with substantial weight savings, this is no longer possible. Moreover a number of unforeseen failures with large composite wind turbine blades have proved that fatigue can be a problem if not properly accounted for [4-6]. In order to determine the fatigue life time of composite structures, designers and engineers are still relying heavily on three types of experimental tests which all have their distinct drawbacks: (1) laboratory tests are often limited to uniaxial fatigue tests on rectangular specimens with a simple lay-up,
often with the sole purpose of establishing the S-N curve. The practical use of these S-N curves is fairly limited because almost all composite structures are subject to a multiaxial stress state under in-service fatigue loading [7-12],
(2) next there are the laboratory tests that generate a multiaxial stress state into the specimen. The most commonly used methods are: (i) biaxial loading of a plane cruciform specimen [13-23], en (ii) a combination of tension/compression and torsion on tubular specimens [24-26]. Both methods require an expensive test set-up and a complex specimen geometry. Stress concentrations at the grips complicate considerably the experiments [27,28],
(3) full-scale fatigue tests generate a lot of usefull information, but the construction of the test facilities and running the fatigue tests is very expensive [29,30].
On the last 'International Conference on Fatigue of Composites' (September 2007) - the most important international conference on fatigue of composites - many speakers stressed the absolute necessity for much faster and simpler fatigue tests capable of providing a lot more information in a much shorter time [31-33].