THE DIURNAL HEATING HYPOTHESIS
The observations made during the first failure investigation are repeated here in summary. Figures 3 and 4 show typical damage observed in the upper skin of the PC sheeting. There appear to be oval holes generally about 5 mm wide (i.e. the web spacing), holes elongated along a web, cracks along web positions, and places where the damage is of the form of a flap in the surface. Figure 5 shows a schematic diagram illustrating the manner of attachment of the PC roofing sheets to the steel structure via 'J' bolts. Clearly, such an attachment system will restrain the expansion and contraction of the PC sheeting in two orthogonal directions, along the web and perpendicular to it. The thermal gradient experienced during a diurnal cycle could be large, ranging in the summer from perhaps 40o-50oC during the day down to 5o-10oC just before sunrise. Daily differences in winter would be smaller, but could still be as much as 20o-30oC. The coefficient of thermal expansion of PC is 6.5 x 10-5 /oC, i.e. an expansion of around 0.195 mm over the 300 mm distance per 10oC change in temperature, and some 0.585 mm over the 900 mm fixing distance per 10oC temperature change. These expansions would tend to be concentrated at the web positions and are large enough to be likely to initiate cracks in the PC sheets.
This information was interpreted by the investigators to provide a failure hypothesis in which the failure was the result of normal 'wear and tear', and hence was not covered by the insurance policy. Think about the information presented and decide what this set of investigators would have concluded regarding the origins of the damage. Check your deductions with the applet given below.