how much above "capacity" you feel comfortable going over? 2

[AskTooMuch]

I have an existing steel beam analyzed using LRFD for additional dead load. I used Risa3d and unity ratio is 1.2 for bending check. Say I already sharpened the pencil as far loading is concerned.

 

[MIStructE_IRE]

6-7% is my limit if I’m stuck.

20% is way too much for me, particularly in bending. If its 20% over in bending alone, then deflection must be through the floor?

Is there any option to strengthen?

 

[Ron247]

Confirm the following that are unrelated to load:

Length-we tend to rough measure things. Be more accurate; longer beams give higher moments AND more total load

Unbraced length; we tend to use one value for a beam when in fact there are times the maximum unbraced length does not occur at the max moment. Most softwares tend to allow one value for unbraced length.

I'd say anything over 5% is pushing it especially if the beam is not in great condition. 20% is way too far over. In the old ASD days, beams were about 67% stronger than the ASD load indicated. So taking 20% out of the 67% is a big deal.

 

[skeletron]

Everybody above is pointing out some really good considerations. Length is one of the best ways to "sharpen your pencil" for bending since it's the squared term. And I agree with MIStructE: what's the deflection on the beam if bending is 20% over?

My comfort zone is <10% and it really depends what I'm analyzing. I'll usually put a note in the calc explaining why the overstress is okay. Or I'll include a secondary calc with relaxed numbers.

 

[AskTooMuch]

I'm not worried about deflection because additional load is temporary.
I put it in DL for lesser factor in load combination because I have the exact weight of the equipment.
The equipment will be on the floor only once every few years for maintenance.
I already used beam appropriate LTB length.

 

[BridgeSmith]

If you're saying that you're 20% over using the actual weight of the equipment, it'd be a no-go for me. That's beyond the margin between minimum strength and expected strength for most structural materials. Without some unaccounted for load distribution or redundancy in the system, chances of it coming crashing down are probably better than 50-50.

Also, what happens a few years down the road, when they upgrade to a new piece of equipment that weighs more and don't check with you first?

 

[Lomarandil]

I'd need to have a LOT in my back pocket (e.g. areas where I know I can still sharpen the pencil) before I moved on from a 20% overstress. Even then, I probably would at least do enough side calcs to get down to 105% or so.

Which flexure mode is controlling? if LTB, I'm assuming you've accounted for Cb already.

Remember, in the client's view, material is cheap. Bad design, time, and miscommunication are expensive.

----
just call me Lo.

 

[AskTooMuch]

Just to be clear, I didn't say I'm ok with 20% over. I'm just getting information about your opinion when it's over 1 in unity ratio on analysis.

 

[EngineerEIT]

The below video should be enlightening and give you the necessary background information to let you make this determination using your engineering judgement.

Link

 

[WARose]

If it's a once-in-a-lifetime load.....maybe around 5%.

 

[AskTooMuch]

Link by EIT (at 37 mins)says it's probabaly "ok" to go to 1.2 DCR/unity ratio as long as you don't do it often.
I'm not saying I'm ok with it. Just repeating what AISC speaker said.

 

[Tomfh]

It’s safe to break the speed limit a lot of the time too. But I’m not sure the last time “I felt it was safe” held up in court as a defence against a speeding ticket.

Is there something in the American codes that says a designer can cross the line?

 

[AskTooMuch]

Tomfh, why don't you watch the link so you understand better. It's from AISC webinar and not some random engineer.

 

[Tomfh]

Tomfh, why don't you watch the link so you understand better.

Yeah I've watched it, and I'm familiar with the probability theory behind it.

It's scientifically completely valid, but as the AISC video points out, if something goes wrong (for whatever reason) then your decision to go over 1.0 can be used against you. It can be used as weapon to attack you by courts, lawyers, opposing experts, the media, etc. That's why I'm not comfortable with it.

 

[dhengr]

Me thinks you ‘ask too much’ and listen too little. If you get another 10 posts saying about the same thing…, that you’re skatin on kinda thin ice, will you start thinking a little about another tack on the problem? But, you use your own (engineering?) judgement, not Risa. Since you asked if 1.2 was o.k., I suppose everyone thought that was what you were asking about, would 1.18 be o.k? It should probably be o.k. as long as it doesn’t cause any problems, any failures or cause you any trouble. Maintenance equip. which must be moved in and moved out, and operate too, is probably not normally DL, so that may not be a good way to sharpen your pencil, with a low load factor. And, should we assume that moving the equip. doesn’t load any beams that Risa3d hasn’t looked at? What does your boss think of this kind of overload?

 

[pikku14]

My .02: It depends on the situation. How likely is the load/combination of loads to occur? How refined has the analysis been? How are the construction practices/oversight?
And also: how will it fail? Is it a ductile failure that will give warning, and the system has capacity to handle the re-distribution of loads? or is it going to be a sudden collapse and/or have no possibility of re-distributing the load?

In this instance where it seems like a primary beam and dead loads, I wouldn't be pushing it over it's predicted capacity

 

[ThomasH]

A few thoughts on the subject.
A utilization of 1.2 for a bending check is high but does not necessarily mean failure. It would however make me take a second look at the analysis.

As I understand it, the loading is know with confidence. That is a good start.
Second question is the geometry, span width, simply supported or more spans?
Third question, material im the beam. Since it is an existing beam, is the material known with certainty?

Then I would make an analysis using the loading with no load factors. Does the result keep me below yield stress? If it doesn't do that for simple bending with elastic section properties, this won't work.
The same analysis with plastic properties, what utilization does that give? My assumption is that both of these will be fine since you have 1.2 and that can be the result of a load factor.

You say that you have the exact weight of the equipment. How is the equipment installed? Could there be any dynamic amplification in this?

The factior 1.2 can indicate different problems depending of the exact context. What exact load factors have you used? What is your safety level?

As I understand the question, you don't meet the code, you wonder if it is dangerous? Probably not, depending on the exact circumstanses. But even if it isn't dangerous (it will probably not fail), why should you take the risk/responsibility?

Speaking in general terms:

I have done this type of analysis using non-linear FEM-analysis. I would model the critical beam with shell elements to get as accurate stress distribution as possible within reason. Then I would start with nominal load and increase the loading to failure.

When I know how it fails, when it fails and how sudden the failure is, then I make my conclusion/recommendation. But I would also explain to the the client that the structure does not meet the requirements in the code. The latter part can be a pedagogic issue if you don't have a professional client.

Thomas

 

[Ron247]

The equipment will be on the floor only once every few years for maintenance.

If the equipment is in fact very temporary, to me, you have a little more engineering judgement leeway. What is the odds you will have max FLL over the area while doing the maintenance item? On industrial, the Client has control over some loadings during some maintenance operations just like Lock Out and Tag Out. They may cordon the area off to prevent forklift traffic or remove all loaded pallets before doing some operation. This same type of control may not exist in Commercial.

 

[WinelandV]

Not a concrete rule at my office, but we do try to get everything analyzed in RISA to under a UR of 1 - because if anything does go pear-shaped it's going to be real difficult to explain away any red (RISA's color code for UR >= 1.0) members in the output.

In your case, I'd probably start looking at ways to reduce the unbraced length - that is, engineering new steel to reduce the LTB. Failing that, I'd suggest a shoring solution for the times that this equipment is on the beam.

What is this equipment being used for? Remember, the goal of the people using said equipment won't be to make sure that they don't hurt the structure, but to get their job done as fast as possible. You may wish to force some fat into the design to account for uncertainties in the procedure.

 

[JLNJ]

30 years ago, the old guys would allow 10% over on existing conditions using ASD.

Now, we have to think about telling your story to the judge what your justification was for your methodology should anything happen. I suppose the real question is, "What is the standard of care?"