|
ANALYSIS
" Mathematics, rightly viewed, possesses not only truth, but supreme beauty, a beauty, cold and austere, like that of sculpture."
Bertrand Russell
The quantitative analysis of the behaviour of fabric structures under severe loadings, has developed to the point where they can now, be engineered in every way to the same performance criteria as a permanent structure. With the assistance of computers and very high level of engineering, the consulting engineer is now able to provide analysis on even the most complex of fabric shapes.
Because fabric membranes undertake relatively large shape changes under wind load, ( in comparison with other structures ) the analysis must prove the surface form under load in it's pretension state, and in its loaded state. This means the membrane must be analysed in a series of unequally different geometry's to prove non-linear performance. On a conventional structure, if the loads increase, then the stresses increase proportionally. On a non-linear structure, the resultant load is not constant. Ultimately the maximum load that can be imposed upon a membrane is that which matches the maximum tensile strength of the skin. As the skin is 2 dimensional, (or effectively so), and the thickness of the material is minimal, (to the extent that it effectively has no thickness), there can be no bending or compressive resistance.
The skin must remain in tension at all times. Therefore the maximum stress that can be accepted by the membrane is the ultimate tensile resistance in the skin. The fabric roof, (once the equilibrium shape has been determined), is, under analysis, given a loading equal to the highest pressure expected, which deforms the initial equilibrium state. This deflection is taken up by the fabric membrane in response to the load. Because the fabric membrane has a certain give to it, which acts as a shock absorber, the membrane redistributes the load across the whole surface, and higher wind loads can usually be taken by membranes than by traditional tented structures. In this way, with loading on one corner of the membrane and lift on the other corner of the membrane, the deflective shape transmits load from one high down force area, to the other high lift area, without too many problems.
It is nessesary to allow for a given pre-tension in the erected structure as a value during the initial form finding. This pretension, sometimes as high as 1-1.5 tons, per lineal metre of membrane, usually measured at the perimeter is required to provide the out of plane stiffness to prevent buckling or wrinkling and is an intrinsic part of the actual membrane design.
The resulting figures are taken by the engineer performing the analysis and are compared to 'rule of thumb' estimates and results from previous similar successful structures as a rough check to determine that these precise numbers are accurate. Once the maximum displacement has been checked of the membrane; it's maximum shape change, it is often a good idea to re-run the shape finding software, with the nodes in that position to determine whether the pre-determined form matches the maximum shape deformation or whether local high stress points or even by generating high stress points in one area, relaxation of the membrane occurs in another area. All of these must be allowed for and the best and most complex of software achieves all of these in an interactive package that combines all the functions described and in a form of dynamic loop of shape finding, applying pressure, measuring the dynamic response, checking the maximum displacement and checking the patterned shape against that maximum displacement and in the pretension state, to give a software package that is truly rounded. Various offices have forms of this, and usually the engineers prices reflect the sophistication of the software, rather than the actual skill of the engineer although it is usually only the most skilled and experienced engineers that use the top end software. This process of the design of fabri membranes is further complicated by the addition of stiffening members or webbing's or rigid members within the membrane, particularly when the stretch characteristics of the stiffening members may be different from the membrane itself, for example; the membrane will have one stretch characteristic of co-efficient, and a webbing sewn to a fabric strip and then welded to the membrane will have considerably different stretch characteristics, therefore both must be tested and the stretch correction factors allowed for in the dimensions and when the webbing is welded to the fabric membrane each must be stretched to the correct tensions prior to welding tone to the other. This will result in the webbing puckering the fabric membrane that it is attached to and when loose in the workshop and will only pull out flat when have been tensioned to the correct amount in their service conditions.
|