Ice Load on Conveyor Box Truss & Tension-only Members

[ARS97]

A few questions.

1) When designing box trusses for an elevated conveyor system, it is my opinion that consideration of ice load (ASCE 7-10 Chapter 10) would be appropriate. I consider these trusses to be "ice-sensitive" structures due to the exposure and relatively small members used. In some cases, this ice load can be appreciable, depending on the location and member profiles used. Does anyone agree, or am I over-designing?

2) What's everyone's thoughts on using tension-only members within a box truss, particularly if the members are supporting transverse loads, such as selfweight & ice?

3) Wind load pressures typically get fairly high (normally 50-65 psf for most applications that I deal with) for box trusses due to the force coefficient used. (I typically use figure 29.5-2 in ASCE 7-10.) I will simply apply this pressure to the first-exposed face of the truss as uniform loads on each member of that face. The load is obviously dependant on the members profile dimension. I've seen some engineers simply apply the wind load as nodal loads on the panel points. I guess I'm just looking for everybody's opinion on what's the most appropriate.

Thanks for taking the time..........please feel free to provide your thoughts!

 

[aggman]

In regards to the wind loading I do not have ASCE 7-10 yet. We maintain the applicable building code of the area we are working in so we are still working with ASCE 7-05 mostly. That being said I do not know what Figure 29.5-2 is so I can't comment on that. If you want to post the figure than maybe I can have some thoughts on that. As such with ASCE 7-05 we typically use Figure 6-23 for trussed towers to determine our shape factor for the structure. This gives us a range of Cf around 2.6 - 3.2 based on the structure. If shielding were not accounted for you would have a Cf of 4.0, assuming Cf = 2.0 for a structural shape (angle, wide-flange, etc) and two frames. If you review the wind loads on petrochemical structures and any other wind testing data you can find charts that give the nominal total structure shape factor based on the length, frame spacing, and number of frames. The results will be similar to what you find ASCE 7-05 using Figure 6-23. Obviously one can reason what the wind load would be with a Cf of 2.0 on the windward frame and then the net balance on the opposite frame. As far as we are concerned we just do not feel the need to break it down to this level. We design for a normal wind load, but in reality it will be skewed. If it fails from that difference then more than likely you didn't have any hope of it working anyway. As such we take the total wind load based on the Cf of 2.6 - 3.2 (or whatever it calculates to be) and divide it equally to the windward and leeward frame based on the projected area of the member. From an exact standpoint this is not correct but we think it's good enough. I am not one that gets bogged down in decimals in engineering. We don't know that loads that well.

I can not tell you that tension only bracing will not work. The response of the structure is real irregardless what the input assumptions are. If your tension bracing is tied to a joint the stiffness of the joint is affected by that tension only bracing. The joint does not care about assumptions an engineer is making that may or may not matter. If a chord deforms between two panel points it will drag any and everything attached to those joints with it. When those members are moved they will also resist that movement and stress as such. To me it is not a "gray area". The code tells you what is required to adequately brace a chord. Look at the joint displacements and see what the loads are in all the members based on the joint stiffness, not your assumptions. It may very well be that by code standards the truss will not work, but in reality it does. This could simply be the difference in a code design and a working design. Think about the I-35W bridge failure. It stood for 50 years and was under designed from day one. I have reviewed failures that had components that were overstressed three times yield under dead load only yet the machinery operated for three years. Steel has a funny way of trying not to fail meaning that it will find other ways to carry the load if at all possible. One other thing is that you have a box truss. It does not respond the same as a roof truss or a floor truss, etc. Remember it does not have a diaphragm. This is a three dimensional structure and the load path will be as such.

My advice is that you do as much research as need be to prove to yourself why a structure is currently working and if it has a weakness or not. An engineers knowledge is their best tool. If you approach your client and tell them the structure is not adequate the first thing they are going to say is "well then why is it working today as is?" Saying because the code said it will fail will not give you much credibility. If you look at the real operational loads and try to determine the limits of the structure as currently designed you can discuss your thoughts about that to them. More than likely they will then respect your opinion and move forward with a solution. When you discuss design flaws you have to discuss input loadings as well. Just because a platform will fail with a 100 psf load on it does not necessarly mean it will be collapsing at any moment. That platform may never see the "design" loading. Remember they only see the loads that are applied in real life.

 

[ARS97]

I think I see what you're saying about the wind load. In a nutshell, we would roughly use the same overall wind load.....I would distribute it entirely to the windward side, while you would divide between the windward & leeward sides.

I'm certainly aware of the difference between being code-compliant and and labeling something unsafe. Probability of design loads, inherent safety factors built into the codes, minimum material strengths versus actual material strengths, etc.

In my particular case, it's unlikely that the customer will question what remedial measures I end up recommending, so I'm not so much concerned about proving my case.......I'm just looking to open up a helpful discussion on the topic. I appreciate your contributions so far.....