Who "owns" the factor of safety? 3

[jheidt2543]

I've thought about this question a few times over the years and I'm wondering what you fellows think about it.

Many years ago, during the construction of a local precast concrete warehouse, a dispute arose, while I was not involved in it (really I wasn't - honest!) here is the gist of it: The precast beams deflected during the placing of the double tee roof plank. The dead load on a building like this is much bigger than the live load. The job was stopped and a testing engineering firm was brought in. Their analysis maintained that the beams were under designed. The Precast company engineer maintained that they cast the beams per plans and even if the beams were under designed, they built them according to the plans.

The Engineer of Record actually did design all the precast (unusual today, but this was 40 years ago - I've been thinking about this a long time) and maintained that although there was "excessive" deflection, the beams were safe and no changes needed. His argument was:
1. The loading was below the beam ultimate strength.
2. The as-cast concrete strength of 6,500 psi was much higher than the
5,000 psi strength used in his design.

I've other projects that raise this same question, although the details may be different. SO, who "owns" the factor of safety. Does the design get to count it to protect his "mistake", is the owner of the building "cheated" because he doesn't get the "benefit" of the added strength, even though the building might be perfectly usable?

 

[rb1957]

IMO, the original EOR erred because whilst the structure may be strong enough, it has failed to meet the deflection criteria (required by code ?).

IMO, factor of safety is a tiny part of the calculations. anywhere you can make an assumption, conservative or unconservative, sometimes depending on your point of view. somebody involved in the design may prefer to make conservative assumptions to save doing extensive detailled analysis. some might say "the project can save money if you did a more detailled analysis and were less conservative". the analyst might reply either ...
1) i've fulfilled the terms of my contract (eff off), or
2) if the project will pay for the more detailled analysis and give me schedule to do it and ..., then sure (it's all pensionable time)

remains me of the "Jack the architect" sketch ...

Quando Omni Flunkus Moritati

 

[JAE]

I'm not sure I understand the root of your question.

Factor of Safety refers to the strength of a member. In the case you cited, it sounds like the members were safe and met the factor of safety generally required of a member.

The deflection of a member is a serviceability issue, not a strength issue, so "factor of safety" doesn't really apply.

I guess you could argue that there is an inherent "factor of safety" in the generally accepted limits on deflection of members, but that term is not usually applied to deflection/serviceability.

The EOR has the ultimate responsibility to oversee the project in terms of life safety. If a beam deflects too much, there would normally not be a life safety issue unless ponding or some other associated effect came into play.

The excessive deflection shows that the EOR perhaps failed in designing the structure to meet the owner's needs (i.e. stiff for functional and appearance sakes).

But that is a failure of perhaps meeting the owner's requirements and expectations...not a factor of safety issue.

 

[haynewp]

If the engineer concluded that the beam still worked but without the required factor of safety, then it didn't meet Code for new construction. For such a building under construction, it wouldn't surprise me if a building official shut it down for non-compliance.

 

[JedClampett]

If it's extra factor of safety, I'd say it belongs to whoever gets to it first. If it's the factor of safety prescribed by code, it belongs to no one. It can't be used.
As others have already said, if the design didn't meet the code factor of safety, it was inadequately designed. The owner or engineer should request a code variance.

 

[jheidt2543]

Good comments all and I agree in principle. I always had the feeling that the EOR was just playing with numbers, that is using the higher cylinder break strength as one justification to say the beams where OK. As I recall, this project never went to court, but a settlement was reached, with very little being done.

Another example might be considered, say when analyzing an existing older structure that met the code when built, but now falls under a newer code due to a change in use. If the existing member is just shy of meeting the newer criteria, say something less than 10% over stressed, would you consider leaving it as is knowing:
1. There is safety factor of roughly 33% for a steel beam.
2. We don't "know" the real loading usually within 5% or 10%.
Is the Owner done a disservice when not reinforcing a member like this because it could add big dollars to a project? Do you inform the Owner and let him make the call whether to spend the money to reinforce or not?

 

[fel3]

IMHO, Factors of Safety are "owned" by the public that our designs are supposed to protect and are enforced by the Agency With Jurisdiction, who thus own the FOS by proxy. If we as designers add additional FOS (and we all do, either deliberately or simply by rounding up), then we own the additional FOS and our owners (who have paid for the additional FOS) have an ownership interest in them.

Several times over the years, a contractor screw-up has eaten into one of my design's FOS. If it's only my additional FOS, then the construction is OK, but the owner may deserve a rebate. If the screw-up is big enough that it also eats into the code-mandated FOS, then he gets to fix it.

==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

 

[BAretired]

The Engineer of Record actually did design all the precast (unusual today, but this was 40 years ago - I've been thinking about this a long time) and maintained that although there was "excessive" deflection, the beams were safe and no changes needed. His argument was:
1. The loading was below the beam ultimate strength.
2. The as-cast concrete strength of 6,500 psi was much higher than the
5,000 psi strength used in his design.

If those were his only arguments, they are not valid. Ultimate strength refers to failure strength. It is not enough for loads to be "below ultimate strength". It must be lower by a specified factor of safety.

Excess compressive strength in the concrete does not increase bending strength significantly.

BA

 

[haynewp]

For the next question; there are sections in the IBC and IEBC that discuss new loads on existing members and when upgrading is needed based on % increases in stress.

 

[JStephen]

You can make a lot of good arguments about why something ought to be okay, but if it fails, they don't sound very convincing any more.

There's a big difference between knowing something hasn't failed and knowing it won't fail. That factory building in Bangladesh that collapsed is a good example of the difference.

 

[haynewp]

Using concrete for example, I believe there are ways to remove the material safety factors (like phi=0.9) by one of the ACI documents for existing members. This includes testing a certain amount of the existing concrete per ACI. But I wouldn't remove the load safety factors (1.2D+1.6L etc) to justify an existing member works when required to meet the load criteria for a new building.

 

[miecz]

The deflection issue aside, it sounds like the EOR designed the beams to a specified strength of 5,000 psi and the testing agency said it didn't meet code based on the specified strength. But the testing agency reported that the tested strength was 6,500 psi. So the EOR checked the beams based on the 6,500 psi strength and found that they were adequate. I think your question is "is it legitimate for the EOR to use the tested strength, instead of the specified strength?" If that's the question, I would say yes.

As far as the deflection issue goes, there is some disagreement in the industry as to how to calculate deflection of precast beams. I think it would be difficult to prove that a beam doesn't meet "code".

 

[frv]

I've seen that done before, and I think it's a mistake. You can't take the concrete cylinder test results and use that value directly. If you want to base your design based only on the tests performed on the samples of the in-situ concrete you'd have to use the provisions of 5.3.2.2 (assuming fewer than 15 tests) to determine what f'c you can use in design. You certainly can't use the full strength of the test as a basis for design. Alternatively, you could use the plant's historical records of the mix design to back-out the maximum f'c you can use based on the average of 15 or more tests, per 5.3.2.1.

In any case, as BARetired pointed out, the increase in flexural strength due to increased concrete strength is trivial.

 

[oldrunner]

I noticed that no one mentioned whether the reinforcing steel was adequate.

 

[haynewp]

There is guidance in ACI 318 Chapter 20, ACI 228.1, and ACI 437 on sampling requirements and the permitted phi factor increases. But I can't seem to find how to arrive at a usable design strength for an existing member from the in-situ test results.

 

[miecz]

For bridge work,Article 6.5.2.1 of the Manual for Condition Evaluation and LRFR of Highway Bridges reads:

Where mechanical properties have been established by testing, the nominal value for strength is typically taken as the mean of the test values minus 1.65 standard deviations to provide a 95% confidence limit. Average test values should not be used for evaluation.

 

[darthsoilsguy2]

to add another layer of onion skin to that old problem.

unless all the tees were pulled, that beam stayed in service during the "testing". I bet it was tested with a swiss hammer or windsor probes. it is hard enough to find a good core of concrete in a precast beam, but going coring on one that is bowing and suspected of underdesign would be suicidal without shoring, and just plain legally stupid with shoring. WP and SH are nice tools for general investigation, consistency, and finding the "bad zones". But don't use any numbers alone to accept, and never to bump-up the design.