Лако-красочные материалы — производство

Fatigue and creep are intimately related, being facets of a common property of an adhesive. Thus, the adhesive’s visco-elastic response is a major determinant of the fatigue life of the joint with which it is constructed. For most applications fatigue resistance in shear is of overriding importance, and slow-cycle loading is likely to be the most detrimental in allowing cumulative creep. Fatigue performance is again related to the joint configuration, and the resulting nature of the stress and strain distribution within the adhesive layer.

Romanko and Knauss(74) stated that an accurate representation of the adhesive stress-strain relation over the range of loads and environments expected in service is a prerequisite for understanding fatigue behaviour. They recommend monitoring of the load-deflection history of the adhesive layer in model joints, in the form of Fig. 4.16, of which the first few cycles of response are atypical. Romanko et al. (33, 74-76) monitored the creep strain by clip gauge, and described fully their data interpretation; Althof(42) monitored the strain with sophisticated extensometry. Krieger(77, 78) demonstrated that irreversible damage was caused by running a slow-cycle fatigue test in which the peak amplitude exceeded the adhesive’s elastic limit. Marceau et al.(19) observed creep rupture at low frequency cycling of thick-adherend lap-shear joints.

The results of fatigue or Wohler tests are customarily presented as S-N curves, in which the stress as a percentage (5) of the ultimate strength determined in a ‘static’ test is plotted against the number of cycles (N) at that stress to failure, on a logarithmic scale (Fig. 4.17). Matting and Draugelates(80) were the first to collect a large amount of data on adhesive joints of the single overlap configuration, and they showed that the survival time of joints from a large batch was distributed in a log normal distribution provided that the stresses involved were less than the so-called endurance limit. Above the endurance limit there will be a scatter about the Wohler line, and below the survival time is indefinite (which does not mean that the joints are unbreakable!). Marceau et al.(79) confirmed that there was a frequency dependence, observing that the fracture modes of thick adherend lap shear joints changed, indicating creep-rupture at low frequencies. Allen et al.(81), in the culmination of many years’ work, attempted to establish the reality of an endurance limit by combining stress analysis and fracture mechanics concepts to describe crack growth in the joints employed. Whilst enabling the calculations of critical crack lengths of flaw sizes, it was not established that cycling above the endurance limit induced crack propagation any more than cycling below it. Additionally, their approach must be tempered by two important facts, namely that purely elastic response to load was assumed and, secondly, that the cleavage stress calculated at the end of the joint was in error. The results of fatigue testing of double lap-shear joints constructed with cold-curing epoxies and steel adherends are reported by Mays et al. (82, 83).

Fracture

Brittle fracture arising from crack formation and propagation applies more generally to joint failure rather than to joint behaviour. However, as indicated earlier, fracture energy (Gc) or fracture toughness (Kc) may be related to a critical crack length in a structure. Structures may therefore be designed to accommodate cracks of a pre-determined length and remain serviceable. It will be recalled that, like other material properties, Gc and Кc are not unique values. For instance, the plasticising action of elevated temperature and/or moisture absorbed by the adhesive is beneficial in subduing brittle fracture of the adhesive material.

Polymers tend to have rather lower fracture strengths than materials such as metals or ceramics, but not concrete! The theory of brittle fracture applies for polymers as for metals, but with greater emphasis on the development of a plastic zone around the tip of the growing crack (Fig. 4.5). The brittle mechanism is favoured in unmodified epoxies, and as a result of reducing the temperature, increasing the strain-rate or specimen thickness, and having sharp notches. Traditionally, susceptibility to brittle fracture has been assessed by some form of impact testing.

A high Gc determined from short-term tests does not guarantee resistance to long-term crack growth under transient or sustained loads, or in the presence of moisture. For example, crack growth occurring from fatigue loading may be plotted against G or К for different loading frequencies, in the characteristic form of Fig. 4.18. Comprehensive review material related to the fracture of bonded joints may be found in many texts (25-29), whilst Kinloch and Young(84) offer a monograph on the fracture behaviour of polymers.