Eng. practice of allowing 5% overstress 33

[radair]

It seems to be standard structural engineering practice to allow up to 5 percent overstress in structural design. It's been this way since I graduated college in 1980 and I've seen the practice commonly used in the tower analysis field for the last 15 years.

I've been asked my opinion by a government agency as to why this is a safe and acceptable engineering practice, including citing any relevant structural codes. They are not questioning my work but are asking me for a signed & sealed letter of opinion. It seems to me that this would be a better question for their state engineering board of licensure.

Can any of you help?

 

[miecz]

civilperson-

ConnDOT and some municipalities.

 

[radair]

Teguchi - I have reviewed calculations by others on a number of occasions with regard to the 5% rule in IBC (3402.2 in IBC 2000). A relatively small load (roof-mounted antennas and radio equipment) was being added to roof framing members. The design engineer showed that the increase in load over Code-required live loads was less than 5%, indicating he did not have to run a full analysis of the building framing system.

Should the increase exceed 5%, calculations would be required showing the structural members meet Code requirements for new structures.

As I understand it, the purpose of this provision is to allow minor loads to be added to existing structures which may have been built to previous revisions of the building code (i.e. less stringent versions). Without this provision, older buildings would have to be retrofitted to current standards.

Again, this rule would allow 5% overstress only if members were designed to 100% of their capacity, however I don't believe that is the intent. Your view may differ.

 

[hokie66]

Nothing we design can be built perfectly. That is why we specify tolerances. So why is it so hard for some to accept tolerances in design decisions?

I don't agree with starting out with the intention of designing every component with 5% overload, as JStephen said his former employer did with footing bearing pressures, but what is wrong with an acceptance criteria with some tolerance when an experienced engineer judges that the component in question is satisfactory?

 

[Teguci]

We are not kings in our domain, we are servants to the public good.

As engineers we must work very hard to not be arrogant and to work within the limits that the public have set us. If we have to sharpen our pencils to get the loads under 100% allowable, then we need to do that simply because we have agreed to.

An experienced engineer has the option of finding a different load path to carry the extra load. He does not have the option of changing the agreed upon building code criteria.

Remember Babel

 

[miecz]

hokie66-

Acceptable fabrication and construction tolerances are specified in the contract documents. Industry Standard tolerances are written into published documents, such as ACI 117. Soon as a code committee writes a standard design tolerance, I'll follow it.

 

[bridgebuster]

Here's why we can usually live with a 5% overstress; as the saying goes "one picture's worth a thousand words", or as my late squad leader, Mr. Timoshenko would've said "the bridge is still standing".

 

[csd72]

Thats insane.

Its PT though so it still has a load path of sorts.

 

[frv]

bridgebuster-

Star for that one!

Indeed, a picture is worth 1000 words.

 

[bridgebuster]

Thanks frv

 

[Dinosaur]

Everyone has his own comfort zone. I design new to within 5% of allowable. I check other engineers work and wouldn't raise any issue if his numbers were less than 110% of my own. We do often delude ourselves into believing we know more than we do; however, in a courtroom situation, things can get awkward. I frequently reflect on what I would think on the witness stand when the opposing counsel asks, "Did you design this in accordance with the code?"

 

[MechEng2005]

I am not a structural engineer and have no experience working to building codes.

However, I was somewhat alarmed when I read the argument that using conservative loading estimates allows for exceeding limits. I see the acceptable stress as just that, a limit. If you are going to exceed that, you should not say that it was the limit. You need to set a limit and justify it, not set the limit and then say, "Well, we are only a little past acceptable, so close enough." It is acceptable or it is not. No gray area. Additionally, if your saying that conservative loading estimates allow exceeding the limit, then why use the conservative estimate? Why not refine your loading calculation and reduce the design load?

If I was having a second story balcony put on my house, I'd much rather hear, "Originally we were going to rate it so it could hold a heavy-duty truck, but based on design we can only rate it for a mid-size car." than having the engineer tell me it is rated for a heavy-duty truck, but is designed at over-stressed values.

I guess I just feel that a limit is a limit. If you don't like the limit, justify changing it (a little tricky if it is specified by code). If you don't agree with the calculations that caused the limit to be exceeded, then change the calculations and justify them. I completely disagree with calculating values and then disregarding them or diminishing their significance.

I mean no offence to anybody. These are just my opinions. As I said, I am not a structural engineer. I am sure many who have designed for 5% overstess are much more qualified than I. I'm also sure that if I were a prosecuting attorney that many of you would put me to shame in court when you justified your calculations. Just wanted to add my two cents...

-- MechEng2005

 

[frv]

MechEng2005-

The issue is that the difference isn't between a full-sized SUV ad a mid-size sedan.

It's that we designed the balcony for a full sized SUV, we actually have a mid-sized car and someone asks, is it OK if we fill it with gas?

 

[JAE]

As I and others have mentioned above, the 5% overstress is usually, and should be, confined only to checking existing conditions. There are many times where an existing structure is defined by on-site research or original design drawings and an owner wishes to alter something (add a roof unit, cut an opening, etc.). This is where the structural engineer is called upon to understand the framing, the materials and the applicable loads.

In all of these, there is an element of uncertainty and usually for we structural engineers, uncertainty tends to send us to higher levels of conservatism rather than lower.

However, we are also professional who understand structural behavior and how dead loads and code prescribed live loads are developed. We aren't (and shouldn't be) blind slaves to code provisions.

For new construction - forget about any 5% overstress. It usually isn't justified or needed.

For existing circumstances there is nothing wrong with initially and conservatively estimating loads, finding out it isn't going to work, refining the assumptions to a tighter (more time consuming) estimate, and re-calculating.

I have zero qualms about facing a lawyer and justifying my calculation where I write "Unity = 1.03 - call OK". There is nothing unsafe about this since I know the history of code provisions and understand that even the base forumlae for strength has uncertainties greater than 5%.

If you read anything about reliability concepts you know this. Concrete beam shear capacities, for example ([Φ]Vc = [Φ]2(sqrt(f'c)bd is set off of a very scattered set of lab results with highs and lows beyond the 5% either way.

MechEng2004 - I understand you saying that the idea of first calculating something and then "disregarding it" is wrong. But I'd counter that many initial calculations are purposely conservative to help speed up the design. After struggling with a result, it is common to go back and refine the numbers to tighter estimates to get something to work.

 

[hokie66]

JAE,

We will have to agree to disagree on this one. I fail to see the distinction between applying your "Unity=1.03-call OK" to new or existing structures. But I will give you a star for arguing both sides.

 

[JAE]

Did I argue both sides? I didn't think I did but hammered out the response above pretty quickly. Sorry if I did.

The distinction between new and existing is that with new, we aren't usually compelled to "stretch" our design a bit since we have command of most design outcomes - simply size the member for what you need.

For existing conditions, we are given a fixed structure with presumed applied loads. Under this situation, we sometimes get faced with a structure that doesn't work under our initial assumptions.

My argument then is that when faced with this constraint, we should keep in mind that we do have some wiggle room in that:
1. The theoretical capacity of a structural element is based upon formulae with variable accuracy.
2. Our assumed dead loads are sometimes a bit on the conservative side.
3. The live loads are also global code-mandated loads that in reality do not normally get up to the supposed amounts (i.e. office live loads are typically in reality about 10 to 20 psf vs. the 50 psf mandated).
4. The safety factors were not developed as some kind of hard limit. For example, a 1.6 factor on loads does not presume to suggest that at a unity of 1.61 means it will collapse. They are based on reliabilty methods that tried to match load factors and resistance factors combined with the current ASD allowable stress levels of safety.

Finally - look at the relative designs between using ASD and LRFD and we see that doing the design under each of these results in different member sizes. Sometimes these differences are more than 5%.

 

[hokie66]

Even if you didn't intend to argue both sides, I think some of your arguments are just as applicable to new structures as to old. And I think it is spurious to argue that old designs are more conservative than new ones because the codes have changed. Some changes have been made to correct unconservative clauses.

 

[oneintheeye]

i would look at it this way if are you ever going to get the absolute code maximum live load over an entire floor area? No. If continuous structure are the adjacent rooms of a building going to be absolutely crammed so full as to hae the maximum load per unit area on them, while the adjacent span has no live load or vice versa. No. Are the allowable stresses conservative to allow for statistical variances in material. Yes. If one beam is 5% over is it going to fail. No. What is worst that can happen? It will deflect and the load will transfer throughout the structure. So if you had a series of beams and one was 5% over, would you up all the beams in the floor? No. Is this unsafe. I believe not.

 

[cap4000]

I think JAE has it exactly right in terms of a diminished level of safety. In ASD typically noncompact beam W shapes equals 1.67 for the Max Moment to the Allowable Moment. And for compact W shapes it equals 1.73. I doubt that an isolated 5% overstress over a small discrete area would ever cause a failure and therefore is a judgement call for the engineer and not the lawyer.

 

[miecz]

I think some of the confusion of this discussion is that it has tried to address too many scenarios: existing construction, new construction, and existing telecommunication construction. I must admit, that I overlooked the fact that the op was addressing existing telecommunication towers , as that didn't appear in the original post. All of my comments have been directed toward new construction, so I apologize for muddying up the discussion.

Transmission towers (I'm thinking cellphone towers, here) are unique structures;

1. The antennas are frequently changed.

2. There is a Design Standard that specifically addresses these (TIA/EIA-222).

In my mind, when you have a Specific Design Standard for a type of construction, that standard overrides the generic Building Code. We wouldn't want builders of Nuclear Generating stations to be following IBC. Article 15 and Appendix F of TIA/EIA-222 covers criteria for the analysis of existing structures. There is nothing in Appendix F that allows (or disallows) an overstress. Because antennas are changed so frequently on these towers, I believe an overstress allowance would be mentioned in Appendix F, if it were applicable. I believe it is standard in the tower industry to recheck a tower each time an antenna group is added, with no allowable overstress. After all, many separate additions could add up.

 

[MechEng2005]

frv -
In the example, if the load of the car and fuel were less than the SUV that it was designed for, then this wouldn't cause over-stress. If the structure was designed to hold exactly the weight of the car and somebody asked, "can we put gas in it?" in my opinion the answer is no.

JAE -
I understand the using conservative estimates for a quick calculation. However, I feel that if it is worth enough to not have to increase the size of a beam or somehow create a more heavy-duty design, then the extra time spent on a calculation to re-evalaluate the calculations is the cost of that. If it is a minor change in load, it shouldn't be hard to "tweak" the calculation to get the result you desire.