Sunday, October 28, 2012

2012 Week 44: State of Play

Past few weeks been drawing up a few cold-formed steel sheds, my dad doing the engineering. I will be doing workshop details later, probably last minute before Christmas to get the steel ordered and available for erection early next year in January, whilst the steel suppliers otherwise shut down.

Produced calculations for aluminium glazing channel to be used for frame less or cantilevered glass balustrade. Then tested the channel whilst otherwise testing the proposed laminated glazing. Everyone was hoping to break the glass but didn't happen. Proposing to retest the glazing as the grout didn't properly fill the channel.

Otherwise done calculations for several framed balustrades. Seems that balustrades are an ignored component in multi-storey building design. All very nice reducing the thickness of the slab, and getting an extra storey in the height: but the edge of the slab requires some substance into which the balustrade framing can be anchored. Most especially important, if the architects are opposed to base plates. Compound the hassles of getting adequate anchorage in the slab, with the problems of connecting the aluminium in the first place whilst trying to avoid welding. If weld aluminium then reduce its strength, so somehow the aluminium has to be connected to a base plate with out welding, or connected to other structure with out base plate. Then there is the issue that the aluminium tubes themselves don't have adequate strength, but this can be compensated for by the inserts used to attach to the support structure. So balustrades went from relatively simple to some what complex composite structures.

Also been working on design of golf safety net. These are large nets which are placed at the perimeter of a golf course to protect roads and housing from flying golf balls. I reduced the wind load to that typically used for light poles, antenna's and similar structures, rather than directly use the BCA criteria. Such structure is a class 10 structure and covered by BCA: volume 2, which some how seems inappropriate. I was only requested to check the poles. So there is no real  design of the sports safety net. I checked the posts for:

1) Drag force on the posts
2) Cable reactions normal to plane of the net.
3) Cable reactions in the plane of the net.

Not having any specification for the requirements of the net or the posts. That is I would expect the post supplier to have been provided with a specification for the post: either forces on the post, or the required size of the post. Not having any of this, using MS Excel I simply did a goal seek on the catenary  formula, to get the minimum cable tension which would achieve equilibrium for the forces normal to the net. I checked pier sizes using the Rutledge formula.

Got request for further information. Not having section property tables for the circular hollow section (CHS) proposed, I lazily used Multiframe steel designer to check the members. Seems I didn't make it clear that the post was checked for biaxial stresses: the forces normal to the net and those in the plane of the net. The forces in the plane of the net are higher than those normal to the net. These forces are an important factor to design of structures supporting flexible cables. If had a solid plate wall supported between posts then would only really have the forces normal to the plane of the wall to consider. For internal posts the lateral forces in the plane of the net cancel each other, on condition that the net either side of a post is equally loaded. For the end post this is not so, only got a net on one side, however the cables supporting the net also act as guy ropes for the post, so post size can be kept down.

The request wants the wind load to be increased or the posts designed for the breaking load of the cables on the basis that the cable should break before the post. I disagree with this philosophy: when tension cables snap they can cut people in two, take their heads off. The structure should be robust and ductile. The nets should break free of the cables releasing loads from cable and posts. If the cables get overloaded then the posts should fail by forming a plastic hinge at some point above the base, the posts folding up, and reducing their profile to the wind.

The basis of the wind loading code is not altogether to prevent collapse of a structure, since the design load always has potential to be exceeded, rather the main task is to keep the structure and its components anchored to the site.

My design philosophy therefore would be that first the net detaches from the cables, not fully but along at least one edge. By doing so the net releases load from the cables and the posts, it would move from being in a vertical plane to being horizontal: if the wind can keep it horizontal. My first thoughts were that the lower edge should detach. But second thoughts suggested that the top edge should detach, the net would then drop  under gravity: either to the floor or a lower level. The load on the post would then be reduced. This would therefore require that the attachments of the net differ along the two support edges.

If the nets fail to detach and release load then, the heads of the posts would be pulled closer together under the lateral loads. It would be preferable that the posts are tapered: either constant external diameter with varying internal wall thickness, or constant wall thickness with varying external diameter. The post then only has such strength as it needs at each height: if the nets are overloaded then the posts should bend and yield reducing profile to wind and avoid further loading which would lead to fracture. All would then be retained on site, but in need of repair. That's the philosophy, demonstrating suitability is another issue.

So been refreshing memory how to use lights by Dr. Vinicius F. Arcaro, University of Campinas, Brazil. Experimented with about 10 years ago, for sail-shade assessment, at which time I wrote MS Excel/vba subroutines to graphically display the model and results in IntelliCAD 2000, because the AutoCAD script (.scr) which it produces was incompatible with IntelliCAD. Thus far managed to build a model of one segment of the golf-net. To do so also made use of formfinding routines operating in GID. Idea is to consider the net as a coarsely woven fabric and so treat as tension membrane, supported on cables, rather than model the whole net. Though at this point not certain how going to do that.

Not sure of the benefit though if I have no control over the nets and support cables.

The other issue is to provide more information on the design of the piers, and a request that the piers be designed as laterally loaded piles.

Any case designing the posts for the breaking load of the cable would make the steel CHS posts massive, much larger than the timber matchsticks supporting a much larger golf-net standing near by. There is also another golf-net of the larger size to be provided at a future date.

So it seems to me that some design philosophy is required for the design of the nets. Yet another situation where buy something as if off-the-shelf and get engineering done for some bit, on an as needs basis. Get even more complex if the buyer and the certifier both turn out to be the same government authority. Now that definitely would be a lack of independence: and not in the best interests of the community.

Very little is as simple as it first appears.