Structural Analysis

[EricaB]

My company is looking at an existing steel structure to determine if it is capable of handling increased loading due to the heavier equipment being installed on the operating deck.

The steel structure was originally designed and constructed in the 70s using A36 steel, ASD design to the 7th edition steel manual. We have CMTRs for all of the steel used in the building.

We have run into a situation where we need to either reinforce certain beams under the new equipment per the analysis using fy=36ksi or sharpen the pencil in the analysis. When reviewing the CMTRs for the steel, the steel averaged > fy=42ksi. When using the CMTR reported yield strength of the steel we limited the number of modifications to the existing steel.

Any guidance on whether this is a code accepted practice? Do we need to do some form of destructive testing to verify the yield strength of the steel?

Thanks.

 

[jdgengineer]

I'm Not familiar with the term CMTR but I'm assuming it means steel mill certificates? If that is the case I'd say I don't believe you would need to do destructive testing but it obviously couldn't hurt if there is budget fr it. I wouldn't use the average for your steel strength. I'd think you would at least want to use the average minus a standard deviation, but I'm sure there is a procedure out there to calibrate the Fy you have from your as-built conditons to the percentile that AISC uses (for instance I believe ACI assumes the f'c that is used in their equations correspond to the 10th percentile of tested cylinders)

 

[EricaB]

CMTR = certified mill test report

We want to use the reported minimum values less a small percentage for uncertainty.

Fortunately, budget is not a problem for this project... we could always take samples from the steel to prove the steel yields at a higher value than 36 ksi.

 

[Ron]

Other than a deviation from the standard of care, there is nothing that will prevent you from doing as you noted. I would take samples of the steel rather than rely on the mill certificates.

 

[Ron]

Also...since you seem to have a generous budget, why not just reinforce the members and be done with it.

 

[ToadJones]

I have always been afraid of this....I have heard that even testing coupons can be difficult since the steel can test differently in different locations.
For a nuke plant, I say reinforce.

 

[dcarr82775]

The Mill Cert tests are run pretty fast and tend to be a little higher then you get when the test is run slower.

Part of the safety factor is that the steel will be stronger than Fy. For temporary structures I have used the tested Fy, but not for permanent.

 

[steellion]

I would limit Fy to 36 ksi in the analysis regardless of what the tests or mill certs say. A36 steel SHOULD always have Fy>36 ksi; using 36 ksi is part of the safety factor as dcarr mentioned. My recommendation is to reinforce the beams and be done with it.

 

[MiketheEngineer]

Reinforce it.

 

[JoshPlumSE]

I don't think there is anything wrong with this approach. We've known for years that most A36 material has a much higher yield strength.

Running some new material tests so that you can take advantage of that extra strength seems like the most cost effective way to proceed.

You obviously don't want to cut any corners for nuclear work. But, this shouldn't be viewed as cutting corners. More like reducing the uncertainty from the original design....

 

[DWHA]

Be carefull. I see you handle says you are Nuclear. And I know that in the US, the NRC does not allow the use of CMTR's for justificaion of increased loading on a steel structure.

They do however allow the use of concrete cylinder test results. But you have to be very particual with this and gather all of the cylinder results for the area in question, then use the minimum. Atleast that is how I have seen it done.

 

[JAE]

One thing to note here is that you just don't want to take your mill tests, or your coupon tests, and use the minimum value.

It may be that there is actually a statistical way to determine an applicable Fy which would be sensitive to the variability of the data.

I don't have a clue where you'd go to look for that statistical treatment of the scatter (i.e. to get a 90% confidence value based on the variability) - so to me - using 36 ksi would be my first choice anyway - or if I absolutely had to - use some value below 36 and the lowest test/coupon value. So if your lowest test/coupon was, say, 41.5 ksi, I'd perhaps use something like 38 or 39 ksi for Fy.

 

[azcats]

This thread:

http://www.eng-tips.com/viewthread.cfm?qid=288762

and especially JAE's first post might also be of interest here.

 

[AUCE98]

Appendix 5 of the current AISC steel manual, page 16.1-417, section 2 has some recommendations for yield strength for insitu material. Fys=R(Fy-4), where Fys is the static yield stress, R is a reduction factor either 1.0 or 0.95 depending on where it was taken, flange or web, and Fy is the reported yield stress. I would take the lowest coupon test and apply it to this equation to determine you design stress.

 

[JAE]

azcats, wow....two years ago and I'd forgotten that I'd posted that. I thought there was something in the back of my mind about this topic. Thanks for digging that up.

 

[AUCE98]

azcats,

Sorry for stepping on your post, I didnt read it before I posted.

 

[ToadJones]

I guess the other obvious thing to keep in mind is stresses aren't the only part of design.

I'd also tend to think that if you were designing to today's AISC code vs 1970's, there will be some inherent reserve capacity.

 

[EricaB]

DWHA, you say that the NRC does not allow the use of the CMTRs for the increased yield strength. Where does this come from?

This is actually the issue I'm running into because the NRC tells us that it is not in compliance with the AISC code if we use the CMTR for yield strength. I disagree with their approach, but then again this is a nuclear world where things are very different. Thank goodness this is just our turbine building that does not serve any safety functions.

We are trying to limit the number of steel modifications because of time. Even though we have a healthy budget, we simply don't have to the time to be doing the reinforcement on the building. Obviously we will if we have to, but that causes serious construction delays.

 

[EricaB]

And since the 7th edition was the code of record, we are reanalyzing everything using the 7th edition instead of the current year code.

Thanks to everyone for your responses on this post! I appreciate all of your expertise.

 

[ToadJones]

"And since the 7th edition was the code of record, we are reanalyzing everything using the 7th edition instead of the current year code."

Is this the proper thing to do?