Plastic analysis and design of steel frames

[civeng80]

Hello all.

I attended a 1 hour seminar on plastic design of steel frames a few days ago and have been wondering if any of you use it or are thinking of using it for design of frames particularly portal frames.

The lecturer suggested not may engineers use it because they think its too difficult particularly when the frame becomes a little different to the solved frames in the literature.

He suggested a general method of solution using conventional structural analysis programs (elastic analysis) and together with a spreadsheet evaluating a plastic analysis (not quite sure how he did it).

Another comment that was made by someone was how weird it was that we use Limit state loads then use elastic analysis methods for the structure and then a plastic approach to design the section (AS 4100 Australian code).

Wondering if there is anyone with any comments at all on plastic design and analysis?

 

[dik]

I've been using plastic design for about 40 years and for multispan warehouses and that sort of thing it can't be beat. It's generally less costly from a material, fabrication and erection vantage and is easier to undertake.

Dik

 

[rowingengineer]

Let’s do a breakdown of the situation,
There are a few analysis methods but lets focus on the two you nominate
1. Elastic analysis
2. Plastic analysis

Then there are section properties
1. Elastic section
2. Plastic section

Any combination of these is acceptable to the laws of statics, as long as in plastic design the hinge points are correct. The plastic/yield line method is reserved for portal frames and continuous beams analysis generally for steel structures, as when you get above two hinges it can become complex. This method is also used heavily in concrete slabs in the UK, I don’t like it myself due to the fact that deflections are not part of the analysis and need to be done separately, extremely important for concrete. For the portal frame analysis The AIS (or when they were published AISC) has a two publications: “An Approximate Simple Plastic Analysis of Portal Frame Structures” and “Design of Unbraced Cantilevers. N S Trahair Deflections in Pinned-Base Haunched Gable Frame. S Parsanejad”. These will give you the tools to design a portal frame based on the plastic method. Worth reading “Full scale experiments of a steel portal frame building” as so you can get a feel for the situation.

I personally use the elastic method with semi-rigid base plate connections; will get you just as light a section as the plastic, but also give you the ability to access deflections.

As for the elastic and plastic section, the reason AS4100 lets you use plastic section for compact sections, and then elastic analysis is because this is a lower bond solution. The plastic analysis is a higher bound solution.


When in doubt, just take the next small step.

 

[BAretired]

Plastic design is a valid theory and is recognized by many codes including the Canadian CSA S16 code. I have used it occasionally for the design of simple frames. For more complex structures, it is not easy to determine the critical mechanism involved in collapse, particularly with a combination of lateral and gravity loads.

You said

Another comment that was made by someone was how weird it was that we use Limit state loads then use elastic analysis methods for the structure and then a plastic approach to design the section (AS 4100 Australian code).

Perhaps it seems a little weird, but not altogether weird. Under service loads, the structure is usually in the elastic range, so the elastic analysis method is the appropriate one to use. Under limit states loads, there is more opportunity to distribute moments freely which means the elastic analysis is conservative even if not the most efficient.

BA

 

[abusementpark]

Low-Rise Portal Frames are definitely the easiest to apply it to. However, I find sway often dictates that I have heavier sections than strength would require.

 

[civeng80]

With Limit state design in particular (not Plastic Design) how does one know when calculating deflections that some parts of the structure have not yielded? Because if some part has yielded then elastic theory for deflection is not valid.

Its guaranteed in working stree design but not in Limit State design.

Also does anyone have any good reference on Plastic design for large span portal frames such as 50m spans?

Incidentally for such large spans would deflection conditions prevail so that elastic or plastic analysis would not make a difference ?

 

[csd72]

civeng80,

I can remember one of my bosses in Melbourne was a very experienced designer of industrial buildings.

He explained to me that they used to design portal frames by the plastic method before computers.

He said, 'when we analysed them one the computer we found out how much these things actually deflect.'

They do all their portal designs elastically.

It is also important that we do not mistake least weight frames for least cost. The economies are usually in the buildability of the details and not in shaving off the weight.

Not directly related but I would recommend anyone view the videocast Field fixes on the AISC website as this gives a lot of pitfalls to avoid. I learnt quite a bit from it.

http://www.aisc.org/content.aspx?id=2164

 

[civeng80]

Yes, Ive seen some older portal frames (designed by WSD) where deflection were not usually checked at all. This is about 30 years ago.

I think deflection criteria would make the sections heavier.

 

[dik]

It really shines for mult-span continuous beams as found for warehouses, etc. I've only used it for a few portals and generally only when it was not possible to use cross-bracing.

It permits design without concern to alternating loads (there are some exceptions).

There are a couple of 'hiccups'... if deflection is an issue, then calculating a deflected shape/envelope can be a little problematic. Normally with continuous construction, deflection isn't as great an issue. As a guide, the deflections will be approx 1/3 of simple span construction. In addition there is a restriction in the classes of sections that can be used.

 

[youngstructural]

A great deal of the older portal frame structures do not actually have deflection problems because when we analyse them by our "advanced" methods, we're neglecting many other factors. I'll give you two very good examples: 1) The cladding on most structures adds to the "true" stiffness of the frame and limits the actual deflections. and 2) The loads we apply in the code are modified gusts distributed evenly accross the whole structure. They are conservative in most cases.

Didn't deflect severely before, won't start now that the computer has become involved.

I know I'm over simplifying. I know the structures haven't been around for all that long, but there are many of them in lots of locations, and I'd hazard to guess that if this hasn't been identified as a problem, there isn't one.

Cheers,

YS

B.Eng (Carleton), P.Eng (Ontario), MIPENZ (Structural-New Zealand)
Working in Canada, and missing my adoptive New Zealand family... at least I brought the little Kiwi with me!

 

[csd72]

YS makes some good points.

I have seen small industrial sheds built with no bracing and with enough fixings to make the cladding act as a stressed skin diaphragm. There can be erection issues with this as well as the difficulty in quality control.

If you are in Queensland then the ultimate wind often governs rather than deflection and it may be considerably more efficient to use plastic design.

As for the high distribution of wind, the Australian code covers it with a multiple surface reduction factor and an area factor.

But often these building have just been lucky enough not to be subject to a high level of wind.

 

[rowingengineer]

While YS has a point about the cladding, this is only true for small width building. The main contributor to deflection reductions that isn't taken into account in most analysis of portals is the base plate connection, most people assume a pinned end connection. This connection as a rule of thumb under service loads is about 20% of the column stiffness. under ultimate is will be more like 5-10% as a rule of thumb.

When in doubt, just take the next small step.