LRFD probability data 4

[TimSchrader2]

Been awhile since I used LRFD. A design "based on statistical analysis of probability of failure". Upon rereading I have several simple questions that I do not see readily in the searches. Pls respond if you know or have time. Thanks

1) It seems that this design should not be used for machines or many mechanical designs as these probabilities are not documented or easy to find. So assumptions are made, which makes no sense to me. Is this design is mostly for buildings? Some say if the load factors are properly calibrated (Note this is done with ASD) then the results should be the same. Many disagree.

2) If the probabilty of a certain size load ocurance over a 10 year period is ???? percent, then the allowable stress (PER ASD) can be exceeded by the determined values. True? Some say you should not compare the two designs, but when all is said and done the safety factor/Ultimate resistance load of the members or structure is either less or more.

Some say when the Live load < 3 times the DL dead load, that ASD is too or should i say more conservative. Which is my case by a large margin. And only if the operator makes a big mistake.

I have read or at least scanned several articles and can not find these simple answers. I still have my Blue AISC Steel Manual (7th ed).

Thanks

 

[Blackstar123]

Before I give my opinion, just to say, i don't know a single thing about mechanical designs. As far as I'm concerned, you guys design engines.

 It seems that this design should not be used for machines or many mechanical designs as these probabilities are not documented or easy to find.

So if they don't allow you to design on lrfd, the most obvious reason which i could think of is lrfd allows your structure to deform past the yeild point (excessive deformation), which if allowed in mechanical designs could create problems.

Some say if the load factors are properly calibrated (Note this is done with ASD) then the results should be the same. Many disagree.

Nope, don't think so. I'm definitely from the disagreeing group. Calibrated how? As in decrease Asd safey factors to come close to lrfd capacities?

Some say when the Live load < 3 times the DL dead load, that ASD is too or should i say more conservative.

If someone asks me, Asd is more conservative period.

In most simple words, in ASD, strains will never be more than yield strains. If you are limiting your design to stay within yeild strains, it can never be less conservative than lrfd where you allow your strains to be much much higher than yeild strains.
We use lrfd in modern design in civil engineering because at failure we assume deformation of the structure will be in the plastic zone of stress strain curve.

 

[Blackstar123]

What forces do you design for by the way?

A different argument can be made for pure tension loads or pure compression loads.

 

[dik]

LRFD works for mechanical stuff, too... but your designs have a lot of serviceability criteria not addressed by LRFD. LRFD is likely so far down the list that it likely doesn't come into play.

Dik

 

[r13]

I believe the OP is asking the design method for equipment supporting structures. I think, depends on one's preference and constrain of the code/standard, either ASD or LRFD will work, and with the exception of special seismic considerations, the results will be equally conservative (difference small).

Blackstar123,
I am an ASD guy, always curios about the claim that LRFD utilizes strain beyond yield. Can you kindly provide a code provision, equation, or example that explicit address the utilization for cases other than special seismic considerations. Thanks in advance.

 

[Blackstar123]

Retired13,
For flexure capacity, use of Zx in equation is more than evident that code utilizes strain beyond yeild.

But in my opinion, beauty of lrfd is lost in case of pure tension or pure compression because, most of us likes to think of capacity in terms of maximum load, where as, capacity should be seen in terms of ductility instead. That is, failure will occur how long after the first yield.

So if you want to see plastic zone in
Pn = Fy*A, I'm afraid you cannot.

But you can find it if you look at the following equation this way.
ASD: Capacity Pn = Fy*A , strain = yeild strain
LRFD : Capacity Pn = Fy*A , strain >> yeild strain.

This is the basic concept behind the plastic and non linear analysis.

 

[r13]

Blackstar123,

Thanks for the reply. But you may need to elaborate little further, as I couldn't see the difference in the expressions below:

But you can find it if you look at the following equation this way.
ASD: Capacity Pn = Fy*A , strain = yeild strain
LRFD : Capacity Pn = Fy*A , strain >> yeild strain.

If both Fy and A are constant, then... (where is strain in the play?)

Please have a look and join the discussion on this thread Link.

 

[dik]

Retired13: "Can you kindly provide a code provision, equation, or example that explicit address the utilization for cases other than"

A simple example would be a fixed end uniformly loaded beam. The elastic solution would be wl^2/12 using the elastic section of Sx and the plastic design moment would be wl^2/16 using the plastic section of Zx. In addition alternating LL is not usually necessary. You have other serviceability issues that have to be addressed, but these are normally not an issue. You can use longer members with fewer members and connections and generally lighter steel sections. It's a plus Plus situation for savings. I've been using plastic design for nearly 50 years. For warehouses it cannot be beat.


Dik

 

[dik]

Retired13: "I couldn't see the difference in the expressions below" There isn't much difference with tension and compression elements... usually, not much of a chance for re-distribution if failure occurs...


Dik

 

[Blackstar123]

Retired13,
That is the point. There is no difference as far as maximum load is concerned.

 

[r13]

Thanks both again. I am getting it now.

Dik, great example to refresh my head.

 

[dik]

retired13: "Dik, great example to refresh my head."

If you have regular multi-span structure, like a warehouse, there is considerable economics in particular when the snow loading is relatively high and you have to consider alternate loading.


Dik

 

[r13]

Yes. For continuous structure, we've justified additional load capacity through moment redistribution where ever possible.

 

[TimSchrader2]

Hello
Thanks for the complete replies.

I am designing a double Telescopic conveyor that mounts on a Mack truck and extends out to about 98FT. Other companies have done similiar things, like Putzmietzer and Schwing Loop belt. You may be more familiar with the articulated pipe trucks that pump concrete from the Concrete delivery trucks to the pads or roofs.

So most of the forces involved are with the concrete and mostly dead wieghts of the boom Truss sections. About 27lbs/ft. Most of the load is dead load. And i see that LRFD allows less structure in cases where the dead load is some value greater then the live Load. When all is said and done that seems to be why LRFD is mostly used. Although a 'very few" say it should give the same size members. Even more so if the probability of ocurance is low, another factor in the equations. I do not believe any reference is going to give me these probability values to use in the LRFD equations, so i am hesitant to invest the companies time in working on these formulas if its all based on some assumptions of occurance. I have a load case that can only occur if the operator makes a mistake and allows the discharge chute to totally fill with concrete, which will back up the concrete giving about 30% more stress in the Tele Boom. This is pretty hard to do but still possible, based on reports from operators. My only real source of probability data.

The other compounding factor is that i am using ultra high strength 100KSI tubing for the Truss cords. The AISC code does not specifically address these values and this tubing is not sold in any quanities in the US. The car companies do use some of the thinner sheets that thy bend. A European company called SSAB does sell some of the tubing here. It would be good to have a code backing my allowable stress or max capacity resistance of a LRFD design. LRFD design methods should give justifications for lighter sections when subjected to a low occurance load. The examples i have seen show this. Some say when this is done the values where not calibrated with ASD correctly. I have not performed this calibration by ASD, so not sure exactly how this is done. Some say you can not compare or should not use both the ASD and LRFD methods. But the latest AISC gives both and many do. i have designed my share of cranes but nothing like this

Thank again for any input

 

[dik]

You may want to catch the link:

https://www.youtube.com/watch?v=MYbpoSR6UnE

Dik

 

[r13]

My two cents:

i am using ultra high strength 100KSI tubing for the Truss cords. The AISC code does not specifically address these values

The supplier should provide structural characteristics of the product, also you may try searching ASTM if it is ever sold in the US.

Some say you can not compare or should not use both the ASD and LRFD methods.

You surely can compare, but do not mix design methodologies, as these are two different schools, there is no 1:1 relationship.

Here is a good reading material comparing ASD & LRFD. Link

 

[human909]

I have a load case that can only occur if the operator makes a mistake and allows the discharge chute to totally fill with concrete, which will back up the concrete giving about 30% more stress in the Tele Boom. This is pretty hard to do but still possible, based on reports from operators. My only real source of probability data.

That is not a load I would use probability data for. That is a realistic scenario that could occur and as such you should include that load in your appropriate load combinations.

 

[r13]

Can there be a safety trigger that will sense and automatically release the overload material? Besides that, I think for certainty of the load is high (full backup), and the extreme rare nature of the event, you could be benefited from a probability based design approach. Please spend an extra hour to watch through the linked video, which is a great help to your case. Link

 

[human909]

Can there be a safety trigger that will sense and automatically release the overload material?

Naturally there can always be protections, but those protections also need redundancy and appropriate maintenance to ensure the probability of failure is negligible. Aircraft and other life critical cutting edge applications use this approach.

Based on the described likelihood it would be negligent to reduce the design load of a full chute by any probability factor. This is a binary event and one that sound like it is common enough that operators know about it.

So absolutely include the chute load in the ULS. However combinations of loads is where you might gain something back. Eg if there are two independent low probability loads then their combination becomes an extreme low probability event and its weight would be negligible and commonly rounded to zero. (EG many codes do this with ULS loads of winds and earthquakes)

Likewise you want your teleboom to remain within serviceable limits with normal operating loads. But a full chute isn't a normal operating load so you don't include that in the serviceability criteria.

1) It seems that this design should not be used for machines or many mechanical designs as these probabilities are not documented or easy to find. So assumptions are made, which makes no sense to me. Is this design is mostly for buildings? Some say if the load factors are properly calibrated (Note this is done with ASD) then the results should be the same. Many disagree.

Just because the probabilities are hard to find doesn't mean ULS should be avoided. In fact this is exactly why we use ULS rather than ASD.

ASD and any "safety factor design" is generally a black box of unknown probabilities. Safety factors are often easier to deal with and certainly easier to understand, but they have little logical justification without looking at the analytical and/or empirical probabilities that lead to them.

 

[TimSchrader2]

Hello

Thanks for all the well thought out replies. I Have watched or scanned the links suggested. The NASCC seems to be the most specific, offering code references that back up and load or resistance factors that may be used with data from the ASCE. Indicating any real codes are still mostly geared to Buildings or maybe elevators, which is probably the closest factors to machines in general, although most probaly too conservative. The Tanya examples are good. It still seems that without a code to call out the "factors" to use that there is more 'assumption' involved with LFRD then ASD. The AISC calls out .66Fy for compact and .6Fy for non compact sections. And less SF or reduced allowables per appendix C, for members not meeting non-compact section thickness ratio criterea. When all is said and done the resistance and Load factors decided on, seems to be a way to allow some load case combinations to exceed .6Fy, if non compact section critea or APPendix C is met.

The video form the NASCC call out such Non-definitve values as "demand capacity ratio" or "see the spec writers". The speaker admits that there is uncertainity involved.

"That is not a load I would use probability data for. That is a realistic scenario that could occur and as such you should include that load in your appropriate load combinations."

So the Chute load would need to be included under the 'servicability' factors and no other formula of LRFD. If I read this correctlly. So the STD deviation formulas are not applicable. Which would be fine.

an "there be a safety trigger that will sense and automatically release the overload material? Besides that, I think for certainty of the load is high (full backup), and the extreme rare nature of the event, you could be benefited from a probability based design approach. Please spend an extra hour to watch through the linked video, wh"

Yes there will be a pressure relief on the boom cylinders that should limit applied stresses, if the valve acts fast enough. As the chute load can be about 80Ft away at the end of the boom. i have done a fair amount of research and phone calls on ultra high strength steel. There is another thread on this where somone mentions some work being done in this area. Europe is way ahead of us USA here. But i do not think that the European EIN codes address ultra high strength steel, as the MFG admits he does not know. Maybe somone in Germany does know. I Still have some emails to send on this. In any case the Ult to yield strength ratio is even less than aluminum (like 710/700 MPa), so using .6Fy may be considered a stretch for a consistant Service load. As the Alum association code requires less than .6Fy if ultimate load of the alum used is too close to the yield. Similiar Logic should apply to Ultra High strength steel

I am quite familiar with the ASD and AISC manual but no so much LRFD. As my AISC book is the 7th edition just before LRFD. Indicating how long i have been around. Just downloaded the AISC manual 13th edition for 12$.

So Thanks again