Skewed Beam Moment Connection 5

[iponom]

Hi to all!

I made a search for this topic and could find very little relevant examples/information.

I have a condition where two beams are framing into column flanges. The beams are skewed and carry both gravity and lateral moments. See attached (with flange plates drawn).

The shear loading, V, is carried by the shear plates at beam webs...

1. Any thoughts on how to connect the pieces other than what I have shown?
2. How to avoid torsion in the column? (Used DG9 to check for stresses - NG)
3. The best possible placement of welds?

column W12X96; beams are 24 in deep, Force in Flanges = 133k

 

[hokie66]

Sorry you feel that way. You have been given worthwhile advice, and we would hate to think that has been a waste of our time. If that is your attitude, don't bother asking for advice in the future.

 

[nutte]

Next, if you are a structural engineer than you better have a say in the orientation and arrangement of the structural components. You need to convince the architect/owner that this current arrangement is not in their best interest by presenting an aesthetically pleasing arrangement which is constructable and, more importantly, structurally sound.

hmm. thanks. i'll just figure it out on my own... i knew this would be a waste of time... damn...

I can understand iponom's frustration. He should have latitude to change the framing? Larsacious must not be familiar with the relationship between structural engineer of record (EOR), contractor, fabricator, and connection designer. Oftentimes the connection designer is required to design a connection that one knows is the result of something other than the best framing configuration. Depending on the attitude of the EOR, the engineer who came up with this framing in the first place, he may or may not be receptive to suggestions. As for revising the framing and changing member sizes, avoiding this convoluted joint entirely, that point may have passed. Steel could already be ordered, foundations being constructed as we speak. The attitude of the EOR might be to simply throw his hands up and say "it's on the drawings, you bid it, so you figure out the connection." So, while many of the suggestions have been good ones, they don't apply to iponom's dilemma.

Back to iponom's original post, as the connection deigner, you are not responsible for any torsion on the column (whether it's there or not). This is the EOR's job. He sized the members. Your job is to get the load from the connected members into the column. To that end, you need to know the unbalanced moment that is going into the column. The EOR may or may not be willing to provide this information to you. You could assume a worst case scenario and try to get that moment into the column.

Overall, your concept looks good to me. Follow the flange force and make sure you have a sound load path.

 

[iponom]

thank you nutte, you understand my struggles like no other!

 

[csd72]

Agreed, some people seem to have more opinion than experience.

Back to the original question regarding the torsion:

You can look at the components of the flange forces in 2 ways.
1. you could take the flange force all the way to the centre line of the beam and then take the components there resulting in a couple around each axis but no torsion (as long as they are in line with the centre)
2. you could take the flange force to the nearest flange thus resulting in two components which are offset from the centre lines. Looking at this it would seem to result in torsion but if you add the torsion up from the two components you will find that they cancel each other out.

The result is the same either way.

By the way, I would also put a stiffener on the opposite side of the column also otherwise you will get a non uniform load on your welds.

 

[hokie66]

nutte, csd72,
We were merely trying to show iponom that the moments for which he is trying to design a connection will dramatically overstress the column. Only late in the thread did he let us know he is not the building engineer, but just trying to design connections. This says more about the system than about iponom...connections should be designed by an engineer who understands the whole structure.

 

[slickdeals]

AMEN. Completely agree with you mate.

We are Virginia Tech
Go HOKIES

 

[iponom]

I am a beginning structural engineer - and I love connection design - and I have designed many before. I think connection design is more fascinating than structural design as a whole. However the structural drawings rarely provide the insight the design engineer has about the structure. It takes experience to interpret structural drawings to their fullest. Some people look at the structural drawing and only see 10% of the whole, while others see much more than that.


hookie66,

the remark I made about wasting time - yea you should try to do less of that.. - because I hate coming back to this thread and see so much useless posts by you... You never answered a single question that I made... By the way the column will not be overstressed and will be fine - you did your checks wrong or something..

 

[iponom]

hokie66,

the column will be fine because there is a very small gravity component to that moment.. so in the end the resulting moment about the weak axis is much smaller than the Moment capacity.

I still need to understand why there is no torsion and I will work on that...

And all really wanted is to see if anyone had a sketch or an idea about connecting elements themselves without changing the framing.

csd72,

I like how you described the second option. I guess this is where I have a problem understanding it. How would you design the weld at that point? I designed the weld at the flange for the eccentricity from one of the components but not the other. Because only one of the components is eccentric to the weld. However the component that is not eccentric to the weld is eccentric to the column - and thus I say torsion! What am I missing/not understanding? And don't tell me to go back to school.


thanks to all who posted.

 

[BAretired]

iponom,

Let's try to keep the discussion on a professional level.

I have been trying to understand your argument and I believe you do have torsion, but only over the height of the beam, i.e. for 24".

I think you should box the column for the height of the beam by adding plates from flange to flange.

BA

 

[iponom]

BAretired,

Thank you for a helpful tip - I thought about boxing the column but I think that its too much work for the shop... and we just got a word from EOR not to worry about torsion. So far I am the only one who is worried about it - and I think mostly because I really want to justify it to myself.

Thanks for reminding me that we are trying to keep it professional here... Lately I have been talking to people and it is so intense, especially when there is a conflict in the schedule. Instead of accepting the fault - people act cocky and try to blame each other...

I think that I was wrong about the way I responded to hokie66 and I apologize. I should have been more clear in what I wanted to find out.

 

[BAretired]

The attached sketch indicates the Fy forces on top and bottom flanges assuming 100% lateral. It also shows the proposed plates in green.

BA

 

[BAretired]

iponom,

I believe the EOR is wrong to dismiss the matter out of hand. He can not develop sway forces unless the column can accept torsion over the height of the beam.

BA

 

[iponom]

BAretired,

I disagree about the 24", the torsional effect I am try to describe would result from the forces on bottom flanges of the beams framing into the column and would act over the length of the column.

So the top of the column is free (cap plate), the bottom is pinned/fixed, and the torsional forces occur 24" below the top.

here is another sketch hope that helps.

 

[csd72]

BAretired,

I strongly disagree,refer to my last post. Look at first principles force components or use method of sections or free body diagrams.

iponom,

I have found Hokie66 to be very insightfull and helpful on most occasions so I am surprised that he has disapointed you. We also use a considerable amount of our own time responding to these posts so a little appreciation would be nice even if you do not find the posts useful.

Regarding how to analyse the connection, it is normal to assume uniform stress across the width of the flange plates and so the welds to the web and on each side of the flange should be capable of transfering this.

 

[larsacious]

iponom,

Unfortunately with these posts you can not ascertain an individuals background. I apologize if you took offense regarding my post but it was not clear what your role/limitations were for this design issue. Knowing that your role is the steel detailer/connection designer I would still recommend the advice I gave you on 8/9 with respect to framing the beams over the column. Perhaps you can submit a modified detail to the EOR/Architect for their approval?

 

[BAretired]

Attached is a vector drawing of the moments acting at the top of the corner column. The column moment can be resolved into two components perpendicular to each of the beams framing in.

Assuming no gravity load moment, the column carries only the sway moment.

There is no torsion in the column below the beam but there is torsion within the beam.

BA

 

[BAretired]

The last sentence should read:

"There is no torsion in the column below the beam but there is torsion in the column within the beam depth, i.e. for a length of two feet".

The column needs to be boxed for a length of two feet as shown in an earlier post.

BA

 

[csd72]

That force vector gives a moment around the major axis of the column not torsion.

A torsion by definition is a moment along the major axis of a member, in the case of a column this would be into the screen which is obviously not possible if all moments are in a perpendicular plain.

 

[nutte]

Iponom, you say you're a beginning engineer. Hopefully you have a supervisor at your office that can assist you with this connection and guide you down the right path.

I do not agree that the column should be boxed in for the length of the connection. I would resolve each moment into a flange force, then look at each flange plate as a gusset. The upper flange plate/gusset has two forces coming into it. These forces align with the column centroid. These forces don't balance, but that's not a concern. We know they result from the worst case out of likely hundreds of load combinations. Don't trim the gusset as it's shown, down close to the bottom of the column ("bottom" being as viewed in the sketch). Run the gusset from the left beam to the right, without skinnying it up.

Take your two forces at the upper gusset. Assume one is into the columns and one is out of the column. Add the x-component of one force to the x-component of the other (where "x" is parallel to the column web). Size the weld at the gusset to web for this force. You have two welds acting concentrically to resist this force.

Now assume both flange forces go into the column. Add the y-component of one to the y-component of the other (where "y" is parallel to the column flanges). Size the welds at the flanges for this force. You have 4 welds that are bf long, and 4 welds that are just under bf/2 long.

Size the gusset to beam connection for the flange force and be done with it. Add a stiffener on the back of the column for good measure. Repeat at bottom flange plate/gusset.

 

[rowingengineer]

iponom,
Hokie66 posts in this thread are in my opinion spot on the money and it would put you in good stead if you were to read what he is writing and take it under your belt. In my experience hokie66 posts are always of a high quality (aka pure gold), I struggle to think of many times I have disagreed with them.


An expert is a man who has made all the mistakes which can be made in a very narrow field