ABSTRACT
Any engineering structure will be subjected to external forces arising from service conditions or the environment in which it works. If it is in equilibrium, the resultant of the external forces will be zero, but the structure will nevertheless experience a load that will tend to deform it. Internal forces set up within the material will react upon this load. Fatigue is the failure of a material under fluctuating stresses each of which is believed to produce minute amounts of plastic, irreversible, strain [2]. The behaviour of materials under fatigue is usually described by a fatigue life or S-N curve in which the number of cycles, N to produce a failure with a stress peak of S is plotted against N. This curve has the profile similar to an exponential decay that becomes asymptotic to a minimum stress level Sn, which is the stress value at which the component under test is assumed to have an infinite life. This means that stresses below this value applied to the member will never cause it to fatigue, Structures are usually designed to keep the maximum stress to a level below Sn. This usually implies over-design in terms of physical size and material usage in consideration of cases where the member occasionally experiences loading that exceeds Sn. However, in some cases like in the case of aircraft component design, weight is a critical issue constraining the design to the point that only allows it to have a finite life. It is therefore clear that the number of times Sn is exceeded and by how much will have an important bearing on the prescribed life of the component. Such designs inevitably require frequent testing to establish whether the design-prescribed life of the component has been shortened, Almost all the preliminary work in designing the fatigue sensor has been carried out by the Fatigue Monitoring Bureau (FMB) at UMIST, Here, a brief description of the sensor is given, The rate of growth of the crack in the gauge can be determined from fracture mechanics principles using the Paris law as follows [3]:
where C and m are material constants, and M is the stress intensity factor range in the gauge. A typical value for m for many metallic materials is m = 3 and values of C can be chosen to allow for different materials, mean stress and environmental conditions. Where the gauge design is such that the stress intensity factor is independent of the crack length, the crack growth equation can be integrated directly. This gives the relationship between change in crack length in the gauge, /o,.a, gauge factor, q, stress range, S, and number of cycles, N, in the underlying structure as
da = C( S)3 N dN q (2)
showing that the crack growth in the gauge is directly proportional to stress in the structure and the number of cycles, and can be expressed as ff N. By choice of the appropriate gauge factor q from Finite Element analysis, the gauge geometry can be chosen to give an appropriate change in crack length to match weld design performance. In the case of variable amplitude fatigue loading, the fatigue damage can be represented as 'LS/ni where ni is the number of cycles at stress range Si. Since the principles rely on change in crack length in the gauge to indicate fatigue damage, it is necessary to have an initial crack present of sufficient length for constant gauge factor performance (a/W>0.3). The sensor operates by simply following the flexions of the component under test. Since it already has a predefined crack in it, cyclic loading will cause the crack to propagate along the width of the coupon. The rate at which this occurs will be dependant on the magnitude of the inflexions and the material properties of the sensor and hence the component under investigation.
