Discussion. Model Analysis and Testing as a Design Tool.

Abstract

Consultant) said that concrete specimens cut out of an actual reinforced concrete structure showed variations in the initial value of the elastic modulus of 14 -5 million Ib/sq. in. How would the Authors reproduce that in a small model? 47. How many of the models described in the Paper were mounted on compressible supports in order to simulate the effect of relative settlement of the foundations? If some of them were not, then why not? 48. Next, two classical questions which contractors and engineers had been asking for many years. How thick was a 6 in. wall, and which dircction, if any, was vertically upwards? 49. In the comfort of a research laboratory, the answers to these and similar questions appeared to be simple. But when one was building a battery of silos with sliding shutters and the temperature was falling rapidly towards freezing, the answers tended to become somewhat stochastic. For instance, a 8f-in. slab had been found to vary from 73 in. to 94 in. How would the Authors propose to introduce this random element into a model?50. The first report on the collapse of the cooling towers at Ferrybridge said that the towers developed an instability failure when the wind velocity was about 100 mile/h. On the other hand, the models had been able to stand up to an equivalent wind velocity of 200 mile/h. If it was accepted that the wind effect varied as the square of the velocity, the actual towers went down under a loading of unity, whereas the models stood up under a loading of 4. The discrepancy was 300%. How could the Authors reconcile such a discrepancy with the optimistic claims put forward in the Paper? Manning) were the first people in the world to design and build long span aluminium domes. He suggested that at the outset they should make a model; but the members of the full size dome were so small that it proved impossible to reproduce them to a small enough scale in structural alloy, so Mr Hamilton built a full scale dome and test loaded it. The experience that was gained in fabricating and erecting the dome confirmed what they had known instinctively from the start, that the extrusion tolerances, the fabrication tolerances, and the sequence of erection of the members could easily double or treble the stresses in any one member. It was impossible to reproduce that effect quantitatively in a model. Also. in an all-welded structure, the method of clamping the various members to-. gether, the type of welding used, and particularly the sequence in which they were welded, could make a big difference to the locked up welded stresses. This also could not be reproduced quantitatively in a model. 52. There were six major sources of information on which structural design was Proc. Irzstn cio. Engrs, 33 (Feb.) 1966, 183-199. 677 Downloaded by [] on [12/09/16]. Mr Hamilton and he (Mr