It is very sound policy to collect and examine information on joints exposed to natural weathering conditions, rather than to depend solely on laboratory experiments. It has, however, been emphasised that because so many factors can affect joint strengths, extreme care must be taken when interpreting published performance data — and particularly data from durability trials. Nevertheless, the comparison of the results of outdoor exposure to tropical and temperate climates with those from laboratory testing would be expected to give a valuable indication of actual service behaviour(5). Such data have been few but the most complete account of trials organised by the Royal Aircraft Establishment was given by Cotter(93); some of their results are collected in Fig. 4.20. In Table 4.6 are summarised a number of durability studies, including outdoor exposure trials, each selected for their key findings and data interpretation, from a range of relevant experiences. Krieger(116) further advised that thirty years of airframe experience had not shown a failure in the cohesive mode which could be attributed to the environment; all problems had been at the interface.
In the early 1980s, the Wolfson Bridge Research Unit at Dundee University in Scotland embarked upon a programme for the environmental exposure of double lap-shear joints constructed with steel adherends united by cold-curing epoxies. The adhesive type and duration of exposure to each of seven environments were varied over a period of about 15 years. The joints were exposed in both an unstressed and in a stressed condition, and tested to destruction periodically both ‘statically’ and in fatigue. Rooftop exposure to the climate in Dundee (cool temperate) and Mauritius (hot humid) were included. All joints in the natural exposure environments fared worse than their counterparts exposed to laboratory-controlled conditions. Predictably, exposure to the hot/wet Mauritian climate resulted in the fastest and greatest degradation. The condition of the bonded interface of gritblasted joints after 2 years’ exposure can be seen in Fig. 4.22, and classic joint perimeter corrosion may be observed. In fact, all residual joint ‘strengths’ were generally higher than the initial control values, with the notable exception of joints constructed with an epoxy-polysulphide adhesive. This fact emphasises the probability that the plasticising effect of imbibed moisture has made a positive contribution, through improved bondline stress redistribution for this particular joint configuration (of bonded area 80 mm x 25 mm wide). The data from natural exposure of course reflect the actual weather during the period of exposure, which may vary considerably from season to season.